openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/arch/sparc/kernel/traps_64.c

2896 lines
82 KiB
C
Raw Normal View History

[SPARC64]: __inline__ --> inline Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-27 15:13:04 +08:00
/* arch/sparc64/kernel/traps.c
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
* Copyright (C) 1995,1997,2008,2009,2012 David S. Miller (davem@davemloft.net)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
* Copyright (C) 1997,1999,2000 Jakub Jelinek (jakub@redhat.com)
*/
/*
* I like traps on v9, :))))
*/
sparc: migrate exception table users off module.h and onto extable.h These files were only including module.h for exception table related functions. We've now separated that content out into its own file "extable.h" so now move over to that and avoid all the extra header content in module.h that we don't really need to compile these files. Cc: "David S. Miller" <davem@davemloft.net> Cc: sparclinux@vger.kernel.org Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-09-20 05:36:29 +08:00
#include <linux/extable.h>
sched/headers: Move task->mm handling methods to <linux/sched/mm.h> Move the following task->mm helper APIs into a new header file, <linux/sched/mm.h>, to further reduce the size and complexity of <linux/sched.h>. Here are how the APIs are used in various kernel files: # mm_alloc(): arch/arm/mach-rpc/ecard.c fs/exec.c include/linux/sched/mm.h kernel/fork.c # __mmdrop(): arch/arc/include/asm/mmu_context.h include/linux/sched/mm.h kernel/fork.c # mmdrop(): arch/arm/mach-rpc/ecard.c arch/m68k/sun3/mmu_emu.c arch/x86/mm/tlb.c drivers/gpu/drm/amd/amdkfd/kfd_process.c drivers/gpu/drm/i915/i915_gem_userptr.c drivers/infiniband/hw/hfi1/file_ops.c drivers/vfio/vfio_iommu_spapr_tce.c fs/exec.c fs/proc/base.c fs/proc/task_mmu.c fs/proc/task_nommu.c fs/userfaultfd.c include/linux/mmu_notifier.h include/linux/sched/mm.h kernel/fork.c kernel/futex.c kernel/sched/core.c mm/khugepaged.c mm/ksm.c mm/mmu_context.c mm/mmu_notifier.c mm/oom_kill.c virt/kvm/kvm_main.c # mmdrop_async_fn(): include/linux/sched/mm.h # mmdrop_async(): include/linux/sched/mm.h kernel/fork.c # mmget_not_zero(): fs/userfaultfd.c include/linux/sched/mm.h mm/oom_kill.c # mmput(): arch/arc/include/asm/mmu_context.h arch/arc/kernel/troubleshoot.c arch/frv/mm/mmu-context.c arch/powerpc/platforms/cell/spufs/context.c arch/sparc/include/asm/mmu_context_32.h drivers/android/binder.c drivers/gpu/drm/etnaviv/etnaviv_gem.c drivers/gpu/drm/i915/i915_gem_userptr.c drivers/infiniband/core/umem.c drivers/infiniband/core/umem_odp.c drivers/infiniband/core/uverbs_main.c drivers/infiniband/hw/mlx4/main.c drivers/infiniband/hw/mlx5/main.c drivers/infiniband/hw/usnic/usnic_uiom.c drivers/iommu/amd_iommu_v2.c drivers/iommu/intel-svm.c drivers/lguest/lguest_user.c drivers/misc/cxl/fault.c drivers/misc/mic/scif/scif_rma.c drivers/oprofile/buffer_sync.c drivers/vfio/vfio_iommu_type1.c drivers/vhost/vhost.c drivers/xen/gntdev.c fs/exec.c fs/proc/array.c fs/proc/base.c fs/proc/task_mmu.c fs/proc/task_nommu.c fs/userfaultfd.c include/linux/sched/mm.h kernel/cpuset.c kernel/events/core.c kernel/events/uprobes.c kernel/exit.c kernel/fork.c kernel/ptrace.c kernel/sys.c kernel/trace/trace_output.c kernel/tsacct.c mm/memcontrol.c mm/memory.c mm/mempolicy.c mm/migrate.c mm/mmu_notifier.c mm/nommu.c mm/oom_kill.c mm/process_vm_access.c mm/rmap.c mm/swapfile.c mm/util.c virt/kvm/async_pf.c # mmput_async(): include/linux/sched/mm.h kernel/fork.c mm/oom_kill.c # get_task_mm(): arch/arc/kernel/troubleshoot.c arch/powerpc/platforms/cell/spufs/context.c drivers/android/binder.c drivers/gpu/drm/etnaviv/etnaviv_gem.c drivers/infiniband/core/umem.c drivers/infiniband/core/umem_odp.c drivers/infiniband/hw/mlx4/main.c drivers/infiniband/hw/mlx5/main.c drivers/infiniband/hw/usnic/usnic_uiom.c drivers/iommu/amd_iommu_v2.c drivers/iommu/intel-svm.c drivers/lguest/lguest_user.c drivers/misc/cxl/fault.c drivers/misc/mic/scif/scif_rma.c drivers/oprofile/buffer_sync.c drivers/vfio/vfio_iommu_type1.c drivers/vhost/vhost.c drivers/xen/gntdev.c fs/proc/array.c fs/proc/base.c fs/proc/task_mmu.c include/linux/sched/mm.h kernel/cpuset.c kernel/events/core.c kernel/exit.c kernel/fork.c kernel/ptrace.c kernel/sys.c kernel/trace/trace_output.c kernel/tsacct.c mm/memcontrol.c mm/memory.c mm/mempolicy.c mm/migrate.c mm/mmu_notifier.c mm/nommu.c mm/util.c # mm_access(): fs/proc/base.c include/linux/sched/mm.h kernel/fork.c mm/process_vm_access.c # mm_release(): arch/arc/include/asm/mmu_context.h fs/exec.c include/linux/sched/mm.h include/uapi/linux/sched.h kernel/exit.c kernel/fork.c Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-02-02 02:08:20 +08:00
#include <linux/sched/mm.h>
sched/headers: Prepare for new header dependencies before moving code to <linux/sched/debug.h> We are going to split <linux/sched/debug.h> out of <linux/sched.h>, which will have to be picked up from other headers and a couple of .c files. Create a trivial placeholder <linux/sched/debug.h> file that just maps to <linux/sched.h> to make this patch obviously correct and bisectable. Include the new header in the files that are going to need it. Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-02-09 01:51:35 +08:00
#include <linux/sched/debug.h>
sparc64: Fix SMP bootup with CONFIG_STACK_DEBUG or ftrace. Based upon a report by Meelis Roos. Any function call can try to access the current thread register via the _mcount hooks when the kernel is built with -pg (via ftrace or STACK_DEBUG). That can't be setup properly very early on during the bootup of other cpus for sun4u and some early sun4v systems. So add notrace markers to these specific functions, so that _mcount doesn't get invoked too early. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-09-17 02:44:00 +08:00
#include <linux/linkage.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <linux/kernel.h>
#include <linux/signal.h>
#include <linux/smp.h>
#include <linux/mm.h>
#include <linux/init.h>
move die notifier handling to common code This patch moves the die notifier handling to common code. Previous various architectures had exactly the same code for it. Note that the new code is compiled unconditionally, this should be understood as an appel to the other architecture maintainer to implement support for it aswell (aka sprinkling a notify_die or two in the proper place) arm had a notifiy_die that did something totally different, I renamed it to arm_notify_die as part of the patch and made it static to the file it's declared and used at. avr32 used to pass slightly less information through this interface and I brought it into line with the other architectures. [akpm@linux-foundation.org: build fix] [akpm@linux-foundation.org: fix vmalloc_sync_all bustage] [bryan.wu@analog.com: fix vmalloc_sync_all in nommu] Signed-off-by: Christoph Hellwig <hch@lst.de> Cc: <linux-arch@vger.kernel.org> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by: Bryan Wu <bryan.wu@analog.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-05-08 15:27:03 +08:00
#include <linux/kdebug.h>
sparc64: Fix stack dumping and tracing when function graph is enabled. Like x86, when the function graph tracer is enabled, emit the ftrace stub as well as the program counter it will be transformed back into. We duplicate a lot of similar stack walking logic in 3 or 4 spots, so eventually we should consolidate things like x86 does. Thanks to Frederic Weisbecker for pointing this out. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-21 18:08:11 +08:00
#include <linux/ftrace.h>
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
#include <linux/reboot.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/gfp.h>
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
#include <linux/context_tracking.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
Remove hardcoding of hard_smp_processor_id on UP systems With the advent of kdump, the assumption that the boot CPU when booting an UP kernel is always the CPU with a particular hardware ID (often 0) (usually referred to as BSP on some architectures) is not valid anymore. The reason being that the dump capture kernel boots on the crashed CPU (the CPU that invoked crash_kexec), which may be or may not be that particular CPU. Move definition of hard_smp_processor_id for the UP case to architecture-specific code ("asm/smp.h") where it belongs, so that each architecture can provide its own implementation. Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp> Cc: "Luck, Tony" <tony.luck@intel.com> Acked-by: Andi Kleen <ak@suse.de> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Vivek Goyal <vgoyal@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-05-09 17:33:25 +08:00
#include <asm/smp.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <asm/delay.h>
#include <asm/ptrace.h>
#include <asm/oplib.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/unistd.h>
Replace <asm/uaccess.h> with <linux/uaccess.h> globally This was entirely automated, using the script by Al: PATT='^[[:blank:]]*#[[:blank:]]*include[[:blank:]]*<asm/uaccess.h>' sed -i -e "s!$PATT!#include <linux/uaccess.h>!" \ $(git grep -l "$PATT"|grep -v ^include/linux/uaccess.h) to do the replacement at the end of the merge window. Requested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-12-25 03:46:01 +08:00
#include <linux/uaccess.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <asm/fpumacro.h>
#include <asm/lsu.h>
#include <asm/dcu.h>
#include <asm/estate.h>
#include <asm/chafsr.h>
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
#include <asm/sfafsr.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <asm/psrcompat.h>
#include <asm/processor.h>
#include <asm/timer.h>
[SPARC64]: Refine code sequences to get the cpu id. On uniprocessor, it's always zero for optimize that. On SMP, the jmpl to the stub kills the return address stack in the cpu branch prediction logic, so expand the code sequence inline and use a code patching section to fix things up. This also always better and explicit register selection, which will be taken advantage of in a future changeset. The hard_smp_processor_id() function is big, so do not inline it. Fix up tests for Jalapeno to also test for Serrano chips too. These tests want "jbus Ultra-IIIi" cases to match, so that is what we should test for. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:27:19 +08:00
#include <asm/head.h>
[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 14:34:02 +08:00
#include <asm/prom.h>
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
#include <asm/memctrl.h>
Disintegrate asm/system.h for Sparc Disintegrate asm/system.h for Sparc. Signed-off-by: David Howells <dhowells@redhat.com> cc: sparclinux@vger.kernel.org
2012-03-29 01:30:03 +08:00
#include <asm/cacheflush.h>
sparc64: fix sparse warnings in sys_sparc_64.c + unaligned_64.c Fix following warnings: kernel/sys_sparc_64.c:643:17: warning: symbol 'sys_kern_features' was not declared. Should it be static? kernel/unaligned_64.c:297:17: warning: symbol 'kernel_unaligned_trap' was not declared. Should it be static? kernel/unaligned_64.c:387:5: warning: symbol 'handle_popc' was not declared. Should it be static? kernel/unaligned_64.c:428:5: warning: symbol 'handle_ldf_stq' was not declared. Should it be static? kernel/unaligned_64.c:553:6: warning: symbol 'handle_ld_nf' was not declared. Should it be static? kernel/unaligned_64.c:579:6: warning: symbol 'handle_lddfmna' was not declared. Should it be static? kernel/unaligned_64.c:643:6: warning: symbol 'handle_stdfmna' was not declared. Should it be static? Functions that are only used in kernel/ - add prototypes in kernel.h Functions used outside kernel/ - add prototype in asm/setup.h Removed local prototypes One of the local prototypes had wrong signature (return void - not int). Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-17 05:25:54 +08:00
#include <asm/setup.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Fix sparse errors in arch/sparc64/kernel/traps.c Add 'UL' markers to DCU_* macros. Declare C functions called from assembler in entry.h Declare C functions called from within the sparc64 arch code in include/asm-sparc64/*.h headers as appropriate. Remove unused routines in traps.c Signed-off-by: David S. Miller <davem@davemloft.net>
2008-03-26 15:19:43 +08:00
#include "entry.h"
sparc64: fix sparse warnings in sys_sparc_64.c + unaligned_64.c Fix following warnings: kernel/sys_sparc_64.c:643:17: warning: symbol 'sys_kern_features' was not declared. Should it be static? kernel/unaligned_64.c:297:17: warning: symbol 'kernel_unaligned_trap' was not declared. Should it be static? kernel/unaligned_64.c:387:5: warning: symbol 'handle_popc' was not declared. Should it be static? kernel/unaligned_64.c:428:5: warning: symbol 'handle_ldf_stq' was not declared. Should it be static? kernel/unaligned_64.c:553:6: warning: symbol 'handle_ld_nf' was not declared. Should it be static? kernel/unaligned_64.c:579:6: warning: symbol 'handle_lddfmna' was not declared. Should it be static? kernel/unaligned_64.c:643:6: warning: symbol 'handle_stdfmna' was not declared. Should it be static? Functions that are only used in kernel/ - add prototypes in kernel.h Functions used outside kernel/ - add prototype in asm/setup.h Removed local prototypes One of the local prototypes had wrong signature (return void - not int). Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-17 05:25:54 +08:00
#include "kernel.h"
sparc64: Implement IRQ stacks. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-13 09:33:56 +08:00
#include "kstack.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* When an irrecoverable trap occurs at tl > 0, the trap entry
* code logs the trap state registers at every level in the trap
* stack. It is found at (pt_regs + sizeof(pt_regs)) and the layout
* is as follows:
*/
struct tl1_traplog {
struct {
unsigned long tstate;
unsigned long tpc;
unsigned long tnpc;
unsigned long tt;
} trapstack[4];
unsigned long tl;
};
static void dump_tl1_traplog(struct tl1_traplog *p)
{
[SPARC64]: Fix tl1 trap state capture/dump on SUN4V. No trap levels above 2 in privileged mode on SUN4V. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-16 17:41:41 +08:00
int i, limit;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG "TRAPLOG: Error at trap level 0x%lx, "
"dumping track stack.\n", p->tl);
[SPARC64]: Fix tl1 trap state capture/dump on SUN4V. No trap levels above 2 in privileged mode on SUN4V. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-16 17:41:41 +08:00
limit = (tlb_type == hypervisor) ? 2 : 4;
[SPARC64]: Fix typo in dump_tl1_traplog() Actually make use of the 'limit' we compute. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-20 16:54:09 +08:00
for (i = 0; i < limit; i++) {
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
"TRAPLOG: Trap level %d TSTATE[%016lx] TPC[%016lx] "
"TNPC[%016lx] TT[%lx]\n",
i + 1,
p->trapstack[i].tstate, p->trapstack[i].tpc,
p->trapstack[i].tnpc, p->trapstack[i].tt);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk("TRAPLOG: TPC<%pS>\n", (void *) p->trapstack[i].tpc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
void bad_trap(struct pt_regs *regs, long lvl)
{
sparc64: make string buffers large enough My static checker complains that if "lvl" is ULONG_MAX (this is 64 bit) then some of the strings will overflow. I don't know if that's possible but it seems simple enough to make the buffers slightly larger. Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-25 19:03:55 +08:00
char buffer[36];
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "bad trap", regs,
0, lvl, SIGTRAP) == NOTIFY_STOP)
return;
if (lvl < 0x100) {
sprintf(buffer, "Bad hw trap %lx at tl0\n", lvl);
die_if_kernel(buffer, regs);
}
lvl -= 0x100;
if (regs->tstate & TSTATE_PRIV) {
sprintf(buffer, "Kernel bad sw trap %lx", lvl);
die_if_kernel(buffer, regs);
}
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGILL;
info.si_errno = 0;
info.si_code = ILL_ILLTRP;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = lvl;
force_sig_info(SIGILL, &info, current);
}
void bad_trap_tl1(struct pt_regs *regs, long lvl)
{
sparc64: make string buffers large enough My static checker complains that if "lvl" is ULONG_MAX (this is 64 bit) then some of the strings will overflow. I don't know if that's possible but it seems simple enough to make the buffers slightly larger. Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-25 19:03:55 +08:00
char buffer[36];
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (notify_die(DIE_TRAP_TL1, "bad trap tl1", regs,
0, lvl, SIGTRAP) == NOTIFY_STOP)
return;
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
sprintf (buffer, "Bad trap %lx at tl>0", lvl);
die_if_kernel (buffer, regs);
}
#ifdef CONFIG_DEBUG_BUGVERBOSE
void do_BUG(const char *file, int line)
{
bust_spinlocks(1);
printk("kernel BUG at %s:%d!\n", file, line);
}
sparc64: move EXPORT_SYMBOL to the symbols definition Move all applicable EXPORT_SYMBOL()s to the file where the respective symbol is defined. Removed all the includes that are no longer needed in sparc_ksyms_64.c Comment all remaining EXPORT_SYMBOL()s in sparc_ksyms_64.c Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Additions by Julian Calaby: * Moved EXPORT_SYMBOL()s for prom functions to their rightful places. * Made some minor cleanups to the includes and comments of sparc_ksyms_64.c * Updated and tidied commit message. * Rebased patch over sparc-2.6.git HEAD. * Ensured that all modified files have the correct includes. Signed-off-by: Julian Calaby <julian.calaby@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-09 08:58:20 +08:00
EXPORT_SYMBOL(do_BUG);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#endif
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
static DEFINE_SPINLOCK(dimm_handler_lock);
static dimm_printer_t dimm_handler;
static int sprintf_dimm(int synd_code, unsigned long paddr, char *buf, int buflen)
{
unsigned long flags;
int ret = -ENODEV;
spin_lock_irqsave(&dimm_handler_lock, flags);
if (dimm_handler) {
ret = dimm_handler(synd_code, paddr, buf, buflen);
} else if (tlb_type == spitfire) {
if (prom_getunumber(synd_code, paddr, buf, buflen) == -1)
ret = -EINVAL;
else
ret = 0;
} else
ret = -ENODEV;
spin_unlock_irqrestore(&dimm_handler_lock, flags);
return ret;
}
int register_dimm_printer(dimm_printer_t func)
{
unsigned long flags;
int ret = 0;
spin_lock_irqsave(&dimm_handler_lock, flags);
if (!dimm_handler)
dimm_handler = func;
else
ret = -EEXIST;
spin_unlock_irqrestore(&dimm_handler_lock, flags);
return ret;
}
sparc64: Allow chmc to be built as a module. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:17:29 +08:00
EXPORT_SYMBOL_GPL(register_dimm_printer);
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
void unregister_dimm_printer(dimm_printer_t func)
{
unsigned long flags;
spin_lock_irqsave(&dimm_handler_lock, flags);
if (dimm_handler == func)
dimm_handler = NULL;
spin_unlock_irqrestore(&dimm_handler_lock, flags);
}
sparc64: Allow chmc to be built as a module. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:17:29 +08:00
EXPORT_SYMBOL_GPL(unregister_dimm_printer);
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
void spitfire_insn_access_exception(struct pt_regs *regs, unsigned long sfsr, unsigned long sfar)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "instruction access exception", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV) {
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
printk("spitfire_insn_access_exception: SFSR[%016lx] "
"SFAR[%016lx], going.\n", sfsr, sfar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
die_if_kernel("Iax", regs);
}
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGSEGV;
info.si_errno = 0;
info.si_code = SEGV_MAPERR;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = 0;
force_sig_info(SIGSEGV, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
void spitfire_insn_access_exception_tl1(struct pt_regs *regs, unsigned long sfsr, unsigned long sfar)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (notify_die(DIE_TRAP_TL1, "instruction access exception tl1", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
return;
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
spitfire_insn_access_exception(regs, sfsr, sfar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
void sun4v_insn_access_exception(struct pt_regs *regs, unsigned long addr, unsigned long type_ctx)
{
unsigned short type = (type_ctx >> 16);
unsigned short ctx = (type_ctx & 0xffff);
siginfo_t info;
if (notify_die(DIE_TRAP, "instruction access exception", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
return;
if (regs->tstate & TSTATE_PRIV) {
printk("sun4v_insn_access_exception: ADDR[%016lx] "
"CTX[%04x] TYPE[%04x], going.\n",
addr, ctx, type);
die_if_kernel("Iax", regs);
}
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGSEGV;
info.si_errno = 0;
info.si_code = SEGV_MAPERR;
info.si_addr = (void __user *) addr;
info.si_trapno = 0;
force_sig_info(SIGSEGV, &info, current);
}
void sun4v_insn_access_exception_tl1(struct pt_regs *regs, unsigned long addr, unsigned long type_ctx)
{
if (notify_die(DIE_TRAP_TL1, "instruction access exception tl1", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
return;
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
sun4v_insn_access_exception(regs, addr, type_ctx);
}
sparc64: Handle additional cases of no fault loads Load instructions using ASI_PNF or other no-fault ASIs should not cause a SIGSEGV or SIGBUS. A garden variety unmapped address follows the TSB miss path, and when no valid mapping is found in the process page tables, the miss handler checks to see if the access was via a no-fault ASI. It then fixes up the target register with a zero, and skips the no-fault load instruction. But different paths are taken for data access exceptions and alignment traps, and these do not respect the no-fault ASI. We add checks in these paths for the no-fault ASI, and fix up the target register and TPC just like in the TSB miss case. Signed-off-by: Rob Gardner <rob.gardner@oracle.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-09-09 06:34:21 +08:00
bool is_no_fault_exception(struct pt_regs *regs)
{
unsigned char asi;
u32 insn;
if (get_user(insn, (u32 __user *)regs->tpc) == -EFAULT)
return false;
/*
* Must do a little instruction decoding here in order to
* decide on a course of action. The bits of interest are:
* insn[31:30] = op, where 3 indicates the load/store group
* insn[24:19] = op3, which identifies individual opcodes
* insn[13] indicates an immediate offset
* op3[4]=1 identifies alternate space instructions
* op3[5:4]=3 identifies floating point instructions
* op3[2]=1 identifies stores
* See "Opcode Maps" in the appendix of any Sparc V9
* architecture spec for full details.
*/
if ((insn & 0xc0800000) == 0xc0800000) { /* op=3, op3[4]=1 */
if (insn & 0x2000) /* immediate offset */
asi = (regs->tstate >> 24); /* saved %asi */
else
asi = (insn >> 5); /* immediate asi */
if ((asi & 0xf2) == ASI_PNF) {
if (insn & 0x1000000) { /* op3[5:4]=3 */
handle_ldf_stq(insn, regs);
return true;
} else if (insn & 0x200000) { /* op3[2], stores */
return false;
}
handle_ld_nf(insn, regs);
return true;
}
}
return false;
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
void spitfire_data_access_exception(struct pt_regs *regs, unsigned long sfsr, unsigned long sfar)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "data access exception", regs,
0, 0x30, SIGTRAP) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV) {
/* Test if this comes from uaccess places. */
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
const struct exception_table_entry *entry;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
entry = search_exception_tables(regs->tpc);
if (entry) {
/* Ouch, somebody is trying VM hole tricks on us... */
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#ifdef DEBUG_EXCEPTIONS
printk("Exception: PC<%016lx> faddr<UNKNOWN>\n", regs->tpc);
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
printk("EX_TABLE: insn<%016lx> fixup<%016lx>\n",
regs->tpc, entry->fixup);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#endif
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
regs->tpc = entry->fixup;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
regs->tnpc = regs->tpc + 4;
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
/* Shit... */
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
printk("spitfire_data_access_exception: SFSR[%016lx] "
"SFAR[%016lx], going.\n", sfsr, sfar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
die_if_kernel("Dax", regs);
}
sparc64: Handle additional cases of no fault loads Load instructions using ASI_PNF or other no-fault ASIs should not cause a SIGSEGV or SIGBUS. A garden variety unmapped address follows the TSB miss path, and when no valid mapping is found in the process page tables, the miss handler checks to see if the access was via a no-fault ASI. It then fixes up the target register with a zero, and skips the no-fault load instruction. But different paths are taken for data access exceptions and alignment traps, and these do not respect the no-fault ASI. We add checks in these paths for the no-fault ASI, and fix up the target register and TPC just like in the TSB miss case. Signed-off-by: Rob Gardner <rob.gardner@oracle.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-09-09 06:34:21 +08:00
if (is_no_fault_exception(regs))
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
info.si_signo = SIGSEGV;
info.si_errno = 0;
info.si_code = SEGV_MAPERR;
info.si_addr = (void __user *)sfar;
info.si_trapno = 0;
force_sig_info(SIGSEGV, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
void spitfire_data_access_exception_tl1(struct pt_regs *regs, unsigned long sfsr, unsigned long sfar)
[SPARC64]: Do not call winfix_dax blindly Verify we really are taking a data access exception trap, at TL1, from one of the window spill/fill handlers. Else call a new function, data_access_exception_tl1, to log the error. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:44:57 +08:00
{
if (notify_die(DIE_TRAP_TL1, "data access exception tl1", regs,
0, 0x30, SIGTRAP) == NOTIFY_STOP)
return;
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
spitfire_data_access_exception(regs, sfsr, sfar);
[SPARC64]: Do not call winfix_dax blindly Verify we really are taking a data access exception trap, at TL1, from one of the window spill/fill handlers. Else call a new function, data_access_exception_tl1, to log the error. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:44:57 +08:00
}
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
void sun4v_data_access_exception(struct pt_regs *regs, unsigned long addr, unsigned long type_ctx)
{
unsigned short type = (type_ctx >> 16);
unsigned short ctx = (type_ctx & 0xffff);
siginfo_t info;
if (notify_die(DIE_TRAP, "data access exception", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
return;
if (regs->tstate & TSTATE_PRIV) {
sparc64: Fix DAX handling via userspace access from kernel. If we do a userspace access from kernel mode, and get a data access exception, we need to check the exception table just like a normal fault does. The spitfire DAX handler was doing this, but such logic was missing from the sun4v DAX code. Reported-by: Dennis Gilmore <dgilmore@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-20 14:44:04 +08:00
/* Test if this comes from uaccess places. */
const struct exception_table_entry *entry;
entry = search_exception_tables(regs->tpc);
if (entry) {
/* Ouch, somebody is trying VM hole tricks on us... */
#ifdef DEBUG_EXCEPTIONS
printk("Exception: PC<%016lx> faddr<UNKNOWN>\n", regs->tpc);
printk("EX_TABLE: insn<%016lx> fixup<%016lx>\n",
regs->tpc, entry->fixup);
#endif
regs->tpc = entry->fixup;
regs->tnpc = regs->tpc + 4;
return;
}
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
printk("sun4v_data_access_exception: ADDR[%016lx] "
"CTX[%04x] TYPE[%04x], going.\n",
addr, ctx, type);
[SPARC64]: Typo in sun4v_data_access_exception log message. Should be "Dax" not "Iax". Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-20 17:50:09 +08:00
die_if_kernel("Dax", regs);
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
}
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
sparc64: Handle additional cases of no fault loads Load instructions using ASI_PNF or other no-fault ASIs should not cause a SIGSEGV or SIGBUS. A garden variety unmapped address follows the TSB miss path, and when no valid mapping is found in the process page tables, the miss handler checks to see if the access was via a no-fault ASI. It then fixes up the target register with a zero, and skips the no-fault load instruction. But different paths are taken for data access exceptions and alignment traps, and these do not respect the no-fault ASI. We add checks in these paths for the no-fault ASI, and fix up the target register and TPC just like in the TSB miss case. Signed-off-by: Rob Gardner <rob.gardner@oracle.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-09-09 06:34:21 +08:00
if (is_no_fault_exception(regs))
return;
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
info.si_signo = SIGSEGV;
info.si_errno = 0;
info.si_code = SEGV_MAPERR;
info.si_addr = (void __user *) addr;
info.si_trapno = 0;
force_sig_info(SIGSEGV, &info, current);
}
void sun4v_data_access_exception_tl1(struct pt_regs *regs, unsigned long addr, unsigned long type_ctx)
{
if (notify_die(DIE_TRAP_TL1, "data access exception tl1", regs,
0, 0x8, SIGTRAP) == NOTIFY_STOP)
return;
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
sun4v_data_access_exception(regs, addr, type_ctx);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#ifdef CONFIG_PCI
sparc64: Fix sparse warnings in pci.c 1) Declare pci_poke_* in pci_impl.h 2) of_create_pci_dev() should be static 3) ->setup_msi_irq() wants an unsigned int pointer not a plain int one 4) void value expression return in arch_teardown_msi_irq() Signed-off-by: David S. Miller <davem@davemloft.net>
2008-09-12 14:57:40 +08:00
#include "pci_impl.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#endif
/* When access exceptions happen, we must do this. */
static void spitfire_clean_and_reenable_l1_caches(void)
{
unsigned long va;
if (tlb_type != spitfire)
BUG();
/* Clean 'em. */
for (va = 0; va < (PAGE_SIZE << 1); va += 32) {
spitfire_put_icache_tag(va, 0x0);
spitfire_put_dcache_tag(va, 0x0);
}
/* Re-enable in LSU. */
__asm__ __volatile__("flush %%g6\n\t"
"membar #Sync\n\t"
"stxa %0, [%%g0] %1\n\t"
"membar #Sync"
: /* no outputs */
: "r" (LSU_CONTROL_IC | LSU_CONTROL_DC |
LSU_CONTROL_IM | LSU_CONTROL_DM),
"i" (ASI_LSU_CONTROL)
: "memory");
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
static void spitfire_enable_estate_errors(void)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
"membar #Sync"
: /* no outputs */
: "r" (ESTATE_ERR_ALL),
"i" (ASI_ESTATE_ERROR_EN));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static char ecc_syndrome_table[] = {
0x4c, 0x40, 0x41, 0x48, 0x42, 0x48, 0x48, 0x49,
0x43, 0x48, 0x48, 0x49, 0x48, 0x49, 0x49, 0x4a,
0x44, 0x48, 0x48, 0x20, 0x48, 0x39, 0x4b, 0x48,
0x48, 0x25, 0x31, 0x48, 0x28, 0x48, 0x48, 0x2c,
0x45, 0x48, 0x48, 0x21, 0x48, 0x3d, 0x04, 0x48,
0x48, 0x4b, 0x35, 0x48, 0x2d, 0x48, 0x48, 0x29,
0x48, 0x00, 0x01, 0x48, 0x0a, 0x48, 0x48, 0x4b,
0x0f, 0x48, 0x48, 0x4b, 0x48, 0x49, 0x49, 0x48,
0x46, 0x48, 0x48, 0x2a, 0x48, 0x3b, 0x27, 0x48,
0x48, 0x4b, 0x33, 0x48, 0x22, 0x48, 0x48, 0x2e,
0x48, 0x19, 0x1d, 0x48, 0x1b, 0x4a, 0x48, 0x4b,
0x1f, 0x48, 0x4a, 0x4b, 0x48, 0x4b, 0x4b, 0x48,
0x48, 0x4b, 0x24, 0x48, 0x07, 0x48, 0x48, 0x36,
0x4b, 0x48, 0x48, 0x3e, 0x48, 0x30, 0x38, 0x48,
0x49, 0x48, 0x48, 0x4b, 0x48, 0x4b, 0x16, 0x48,
0x48, 0x12, 0x4b, 0x48, 0x49, 0x48, 0x48, 0x4b,
0x47, 0x48, 0x48, 0x2f, 0x48, 0x3f, 0x4b, 0x48,
0x48, 0x06, 0x37, 0x48, 0x23, 0x48, 0x48, 0x2b,
0x48, 0x05, 0x4b, 0x48, 0x4b, 0x48, 0x48, 0x32,
0x26, 0x48, 0x48, 0x3a, 0x48, 0x34, 0x3c, 0x48,
0x48, 0x11, 0x15, 0x48, 0x13, 0x4a, 0x48, 0x4b,
0x17, 0x48, 0x4a, 0x4b, 0x48, 0x4b, 0x4b, 0x48,
0x49, 0x48, 0x48, 0x4b, 0x48, 0x4b, 0x1e, 0x48,
0x48, 0x1a, 0x4b, 0x48, 0x49, 0x48, 0x48, 0x4b,
0x48, 0x08, 0x0d, 0x48, 0x02, 0x48, 0x48, 0x49,
0x03, 0x48, 0x48, 0x49, 0x48, 0x4b, 0x4b, 0x48,
0x49, 0x48, 0x48, 0x49, 0x48, 0x4b, 0x10, 0x48,
0x48, 0x14, 0x4b, 0x48, 0x4b, 0x48, 0x48, 0x4b,
0x49, 0x48, 0x48, 0x49, 0x48, 0x4b, 0x18, 0x48,
0x48, 0x1c, 0x4b, 0x48, 0x4b, 0x48, 0x48, 0x4b,
0x4a, 0x0c, 0x09, 0x48, 0x0e, 0x48, 0x48, 0x4b,
0x0b, 0x48, 0x48, 0x4b, 0x48, 0x4b, 0x4b, 0x4a
};
static char *syndrome_unknown = "<Unknown>";
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
static void spitfire_log_udb_syndrome(unsigned long afar, unsigned long udbh, unsigned long udbl, unsigned long bit)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
unsigned short scode;
char memmod_str[64], *p;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
if (udbl & bit) {
scode = ecc_syndrome_table[udbl & 0xff];
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
if (sprintf_dimm(scode, afar, memmod_str, sizeof(memmod_str)) < 0)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
p = syndrome_unknown;
else
p = memmod_str;
printk(KERN_WARNING "CPU[%d]: UDBL Syndrome[%x] "
"Memory Module \"%s\"\n",
smp_processor_id(), scode, p);
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
if (udbh & bit) {
scode = ecc_syndrome_table[udbh & 0xff];
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
if (sprintf_dimm(scode, afar, memmod_str, sizeof(memmod_str)) < 0)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
p = syndrome_unknown;
else
p = memmod_str;
printk(KERN_WARNING "CPU[%d]: UDBH Syndrome[%x] "
"Memory Module \"%s\"\n",
smp_processor_id(), scode, p);
}
[SPARC64]: Revamp Spitfire error trap handling. Current uncorrectable error handling was poor enough that the processor could just loop taking the same trap over and over again. Fix things up so that we at least get a log message and perhaps even some register state. In the process, much consolidation became possible, particularly with the correctable error handler. Prefix assembler and C function names with "spitfire" to indicate that these are for Ultra-I/II/IIi/IIe only. More work is needed to make these routines robust and featureful to the level of the Ultra-III error handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 03:45:11 +08:00
}
static void spitfire_cee_log(unsigned long afsr, unsigned long afar, unsigned long udbh, unsigned long udbl, int tl1, struct pt_regs *regs)
{
printk(KERN_WARNING "CPU[%d]: Correctable ECC Error "
"AFSR[%lx] AFAR[%016lx] UDBL[%lx] UDBH[%lx] TL>1[%d]\n",
smp_processor_id(), afsr, afar, udbl, udbh, tl1);
spitfire_log_udb_syndrome(afar, udbh, udbl, UDBE_CE);
/* We always log it, even if someone is listening for this
* trap.
*/
notify_die(DIE_TRAP, "Correctable ECC Error", regs,
0, TRAP_TYPE_CEE, SIGTRAP);
/* The Correctable ECC Error trap does not disable I/D caches. So
* we only have to restore the ESTATE Error Enable register.
*/
spitfire_enable_estate_errors();
}
static void spitfire_ue_log(unsigned long afsr, unsigned long afar, unsigned long udbh, unsigned long udbl, unsigned long tt, int tl1, struct pt_regs *regs)
{
siginfo_t info;
printk(KERN_WARNING "CPU[%d]: Uncorrectable Error AFSR[%lx] "
"AFAR[%lx] UDBL[%lx] UDBH[%ld] TT[%lx] TL>1[%d]\n",
smp_processor_id(), afsr, afar, udbl, udbh, tt, tl1);
/* XXX add more human friendly logging of the error status
* XXX as is implemented for cheetah
*/
spitfire_log_udb_syndrome(afar, udbh, udbl, UDBE_UE);
/* We always log it, even if someone is listening for this
* trap.
*/
notify_die(DIE_TRAP, "Uncorrectable Error", regs,
0, tt, SIGTRAP);
if (regs->tstate & TSTATE_PRIV) {
if (tl1)
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("UE", regs);
}
/* XXX need more intelligent processing here, such as is implemented
* XXX for cheetah errors, in fact if the E-cache still holds the
* XXX line with bad parity this will loop
*/
spitfire_clean_and_reenable_l1_caches();
spitfire_enable_estate_errors();
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGBUS;
info.si_errno = 0;
info.si_code = BUS_OBJERR;
info.si_addr = (void *)0;
info.si_trapno = 0;
force_sig_info(SIGBUS, &info, current);
}
void spitfire_access_error(struct pt_regs *regs, unsigned long status_encoded, unsigned long afar)
{
unsigned long afsr, tt, udbh, udbl;
int tl1;
afsr = (status_encoded & SFSTAT_AFSR_MASK) >> SFSTAT_AFSR_SHIFT;
tt = (status_encoded & SFSTAT_TRAP_TYPE) >> SFSTAT_TRAP_TYPE_SHIFT;
tl1 = (status_encoded & SFSTAT_TL_GT_ONE) ? 1 : 0;
udbl = (status_encoded & SFSTAT_UDBL_MASK) >> SFSTAT_UDBL_SHIFT;
udbh = (status_encoded & SFSTAT_UDBH_MASK) >> SFSTAT_UDBH_SHIFT;
#ifdef CONFIG_PCI
if (tt == TRAP_TYPE_DAE &&
pci_poke_in_progress && pci_poke_cpu == smp_processor_id()) {
spitfire_clean_and_reenable_l1_caches();
spitfire_enable_estate_errors();
pci_poke_faulted = 1;
regs->tnpc = regs->tpc + 4;
return;
}
#endif
if (afsr & SFAFSR_UE)
spitfire_ue_log(afsr, afar, udbh, udbl, tt, tl1, regs);
if (tt == TRAP_TYPE_CEE) {
/* Handle the case where we took a CEE trap, but ACK'd
* only the UE state in the UDB error registers.
*/
if (afsr & SFAFSR_UE) {
if (udbh & UDBE_CE) {
__asm__ __volatile__(
"stxa %0, [%1] %2\n\t"
"membar #Sync"
: /* no outputs */
: "r" (udbh & UDBE_CE),
"r" (0x0), "i" (ASI_UDB_ERROR_W));
}
if (udbl & UDBE_CE) {
__asm__ __volatile__(
"stxa %0, [%1] %2\n\t"
"membar #Sync"
: /* no outputs */
: "r" (udbl & UDBE_CE),
"r" (0x18), "i" (ASI_UDB_ERROR_W));
}
}
spitfire_cee_log(afsr, afar, udbh, udbl, tl1, regs);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Add boot option to force UltraSPARC-III P-Cache on. Older UltraSPARC-III chips have a P-Cache bug that makes us disable it by default at boot time. However, this does hurt performance substantially, particularly with memcpy(), and the bug is _incredibly_ obscure. I have never seen it triggered in practice, ever. So provide a "-P" boot option that forces the P-Cache on. It taints the kernel, so if it does trigger and cause some data corruption or OOPS, we will find out in the logs that this option was on when it happened. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-05-24 06:52:08 +08:00
int cheetah_pcache_forced_on;
void cheetah_enable_pcache(void)
{
unsigned long dcr;
printk("CHEETAH: Enabling P-Cache on cpu %d.\n",
smp_processor_id());
__asm__ __volatile__("ldxa [%%g0] %1, %0"
: "=r" (dcr)
: "i" (ASI_DCU_CONTROL_REG));
dcr |= (DCU_PE | DCU_HPE | DCU_SPE | DCU_SL);
__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
"membar #Sync"
: /* no outputs */
: "r" (dcr), "i" (ASI_DCU_CONTROL_REG));
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Cheetah error trap handling. */
static unsigned long ecache_flush_physbase;
static unsigned long ecache_flush_linesize;
static unsigned long ecache_flush_size;
/* This table is ordered in priority of errors and matches the
* AFAR overwrite policy as well.
*/
struct afsr_error_table {
unsigned long mask;
const char *name;
};
static const char CHAFSR_PERR_msg[] =
"System interface protocol error";
static const char CHAFSR_IERR_msg[] =
"Internal processor error";
static const char CHAFSR_ISAP_msg[] =
tree-wide: fix comment/printk typos "gadget", "through", "command", "maintain", "maintain", "controller", "address", "between", "initiali[zs]e", "instead", "function", "select", "already", "equal", "access", "management", "hierarchy", "registration", "interest", "relative", "memory", "offset", "already", Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2010-11-02 03:38:34 +08:00
"System request parity error on incoming address";
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static const char CHAFSR_UCU_msg[] =
"Uncorrectable E-cache ECC error for ifetch/data";
static const char CHAFSR_UCC_msg[] =
"SW Correctable E-cache ECC error for ifetch/data";
static const char CHAFSR_UE_msg[] =
"Uncorrectable system bus data ECC error for read";
static const char CHAFSR_EDU_msg[] =
"Uncorrectable E-cache ECC error for stmerge/blkld";
static const char CHAFSR_EMU_msg[] =
"Uncorrectable system bus MTAG error";
static const char CHAFSR_WDU_msg[] =
"Uncorrectable E-cache ECC error for writeback";
static const char CHAFSR_CPU_msg[] =
"Uncorrectable ECC error for copyout";
static const char CHAFSR_CE_msg[] =
"HW corrected system bus data ECC error for read";
static const char CHAFSR_EDC_msg[] =
"HW corrected E-cache ECC error for stmerge/blkld";
static const char CHAFSR_EMC_msg[] =
"HW corrected system bus MTAG ECC error";
static const char CHAFSR_WDC_msg[] =
"HW corrected E-cache ECC error for writeback";
static const char CHAFSR_CPC_msg[] =
"HW corrected ECC error for copyout";
static const char CHAFSR_TO_msg[] =
"Unmapped error from system bus";
static const char CHAFSR_BERR_msg[] =
"Bus error response from system bus";
static const char CHAFSR_IVC_msg[] =
"HW corrected system bus data ECC error for ivec read";
static const char CHAFSR_IVU_msg[] =
"Uncorrectable system bus data ECC error for ivec read";
static struct afsr_error_table __cheetah_error_table[] = {
{ CHAFSR_PERR, CHAFSR_PERR_msg },
{ CHAFSR_IERR, CHAFSR_IERR_msg },
{ CHAFSR_ISAP, CHAFSR_ISAP_msg },
{ CHAFSR_UCU, CHAFSR_UCU_msg },
{ CHAFSR_UCC, CHAFSR_UCC_msg },
{ CHAFSR_UE, CHAFSR_UE_msg },
{ CHAFSR_EDU, CHAFSR_EDU_msg },
{ CHAFSR_EMU, CHAFSR_EMU_msg },
{ CHAFSR_WDU, CHAFSR_WDU_msg },
{ CHAFSR_CPU, CHAFSR_CPU_msg },
{ CHAFSR_CE, CHAFSR_CE_msg },
{ CHAFSR_EDC, CHAFSR_EDC_msg },
{ CHAFSR_EMC, CHAFSR_EMC_msg },
{ CHAFSR_WDC, CHAFSR_WDC_msg },
{ CHAFSR_CPC, CHAFSR_CPC_msg },
{ CHAFSR_TO, CHAFSR_TO_msg },
{ CHAFSR_BERR, CHAFSR_BERR_msg },
/* These two do not update the AFAR. */
{ CHAFSR_IVC, CHAFSR_IVC_msg },
{ CHAFSR_IVU, CHAFSR_IVU_msg },
{ 0, NULL },
};
static const char CHPAFSR_DTO_msg[] =
"System bus unmapped error for prefetch/storequeue-read";
static const char CHPAFSR_DBERR_msg[] =
"System bus error for prefetch/storequeue-read";
static const char CHPAFSR_THCE_msg[] =
"Hardware corrected E-cache Tag ECC error";
static const char CHPAFSR_TSCE_msg[] =
"SW handled correctable E-cache Tag ECC error";
static const char CHPAFSR_TUE_msg[] =
"Uncorrectable E-cache Tag ECC error";
static const char CHPAFSR_DUE_msg[] =
"System bus uncorrectable data ECC error due to prefetch/store-fill";
static struct afsr_error_table __cheetah_plus_error_table[] = {
{ CHAFSR_PERR, CHAFSR_PERR_msg },
{ CHAFSR_IERR, CHAFSR_IERR_msg },
{ CHAFSR_ISAP, CHAFSR_ISAP_msg },
{ CHAFSR_UCU, CHAFSR_UCU_msg },
{ CHAFSR_UCC, CHAFSR_UCC_msg },
{ CHAFSR_UE, CHAFSR_UE_msg },
{ CHAFSR_EDU, CHAFSR_EDU_msg },
{ CHAFSR_EMU, CHAFSR_EMU_msg },
{ CHAFSR_WDU, CHAFSR_WDU_msg },
{ CHAFSR_CPU, CHAFSR_CPU_msg },
{ CHAFSR_CE, CHAFSR_CE_msg },
{ CHAFSR_EDC, CHAFSR_EDC_msg },
{ CHAFSR_EMC, CHAFSR_EMC_msg },
{ CHAFSR_WDC, CHAFSR_WDC_msg },
{ CHAFSR_CPC, CHAFSR_CPC_msg },
{ CHAFSR_TO, CHAFSR_TO_msg },
{ CHAFSR_BERR, CHAFSR_BERR_msg },
{ CHPAFSR_DTO, CHPAFSR_DTO_msg },
{ CHPAFSR_DBERR, CHPAFSR_DBERR_msg },
{ CHPAFSR_THCE, CHPAFSR_THCE_msg },
{ CHPAFSR_TSCE, CHPAFSR_TSCE_msg },
{ CHPAFSR_TUE, CHPAFSR_TUE_msg },
{ CHPAFSR_DUE, CHPAFSR_DUE_msg },
/* These two do not update the AFAR. */
{ CHAFSR_IVC, CHAFSR_IVC_msg },
{ CHAFSR_IVU, CHAFSR_IVU_msg },
{ 0, NULL },
};
static const char JPAFSR_JETO_msg[] =
"System interface protocol error, hw timeout caused";
static const char JPAFSR_SCE_msg[] =
"Parity error on system snoop results";
static const char JPAFSR_JEIC_msg[] =
"System interface protocol error, illegal command detected";
static const char JPAFSR_JEIT_msg[] =
"System interface protocol error, illegal ADTYPE detected";
static const char JPAFSR_OM_msg[] =
"Out of range memory error has occurred";
static const char JPAFSR_ETP_msg[] =
"Parity error on L2 cache tag SRAM";
static const char JPAFSR_UMS_msg[] =
"Error due to unsupported store";
static const char JPAFSR_RUE_msg[] =
"Uncorrectable ECC error from remote cache/memory";
static const char JPAFSR_RCE_msg[] =
"Correctable ECC error from remote cache/memory";
static const char JPAFSR_BP_msg[] =
"JBUS parity error on returned read data";
static const char JPAFSR_WBP_msg[] =
"JBUS parity error on data for writeback or block store";
static const char JPAFSR_FRC_msg[] =
"Foreign read to DRAM incurring correctable ECC error";
static const char JPAFSR_FRU_msg[] =
"Foreign read to DRAM incurring uncorrectable ECC error";
static struct afsr_error_table __jalapeno_error_table[] = {
{ JPAFSR_JETO, JPAFSR_JETO_msg },
{ JPAFSR_SCE, JPAFSR_SCE_msg },
{ JPAFSR_JEIC, JPAFSR_JEIC_msg },
{ JPAFSR_JEIT, JPAFSR_JEIT_msg },
{ CHAFSR_PERR, CHAFSR_PERR_msg },
{ CHAFSR_IERR, CHAFSR_IERR_msg },
{ CHAFSR_ISAP, CHAFSR_ISAP_msg },
{ CHAFSR_UCU, CHAFSR_UCU_msg },
{ CHAFSR_UCC, CHAFSR_UCC_msg },
{ CHAFSR_UE, CHAFSR_UE_msg },
{ CHAFSR_EDU, CHAFSR_EDU_msg },
{ JPAFSR_OM, JPAFSR_OM_msg },
{ CHAFSR_WDU, CHAFSR_WDU_msg },
{ CHAFSR_CPU, CHAFSR_CPU_msg },
{ CHAFSR_CE, CHAFSR_CE_msg },
{ CHAFSR_EDC, CHAFSR_EDC_msg },
{ JPAFSR_ETP, JPAFSR_ETP_msg },
{ CHAFSR_WDC, CHAFSR_WDC_msg },
{ CHAFSR_CPC, CHAFSR_CPC_msg },
{ CHAFSR_TO, CHAFSR_TO_msg },
{ CHAFSR_BERR, CHAFSR_BERR_msg },
{ JPAFSR_UMS, JPAFSR_UMS_msg },
{ JPAFSR_RUE, JPAFSR_RUE_msg },
{ JPAFSR_RCE, JPAFSR_RCE_msg },
{ JPAFSR_BP, JPAFSR_BP_msg },
{ JPAFSR_WBP, JPAFSR_WBP_msg },
{ JPAFSR_FRC, JPAFSR_FRC_msg },
{ JPAFSR_FRU, JPAFSR_FRU_msg },
/* These two do not update the AFAR. */
{ CHAFSR_IVU, CHAFSR_IVU_msg },
{ 0, NULL },
};
static struct afsr_error_table *cheetah_error_table;
static unsigned long cheetah_afsr_errors;
struct cheetah_err_info *cheetah_error_log;
[SPARC64]: __inline__ --> inline Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-27 15:13:04 +08:00
static inline struct cheetah_err_info *cheetah_get_error_log(unsigned long afsr)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct cheetah_err_info *p;
int cpu = smp_processor_id();
if (!cheetah_error_log)
return NULL;
p = cheetah_error_log + (cpu * 2);
if ((afsr & CHAFSR_TL1) != 0UL)
p++;
return p;
}
extern unsigned int tl0_icpe[], tl1_icpe[];
extern unsigned int tl0_dcpe[], tl1_dcpe[];
extern unsigned int tl0_fecc[], tl1_fecc[];
extern unsigned int tl0_cee[], tl1_cee[];
extern unsigned int tl0_iae[], tl1_iae[];
extern unsigned int tl0_dae[], tl1_dae[];
extern unsigned int cheetah_plus_icpe_trap_vector[], cheetah_plus_icpe_trap_vector_tl1[];
extern unsigned int cheetah_plus_dcpe_trap_vector[], cheetah_plus_dcpe_trap_vector_tl1[];
extern unsigned int cheetah_fecc_trap_vector[], cheetah_fecc_trap_vector_tl1[];
extern unsigned int cheetah_cee_trap_vector[], cheetah_cee_trap_vector_tl1[];
extern unsigned int cheetah_deferred_trap_vector[], cheetah_deferred_trap_vector_tl1[];
void __init cheetah_ecache_flush_init(void)
{
unsigned long largest_size, smallest_linesize, order, ver;
[SPARC64]: Use machine description and OBP properly for cpu probing. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-26 06:49:59 +08:00
int i, sz;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Scan all cpu device tree nodes, note two values:
* 1) largest E-cache size
* 2) smallest E-cache line size
*/
largest_size = 0UL;
smallest_linesize = ~0UL;
[SPARC64]: Use machine description and OBP properly for cpu probing. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-26 06:49:59 +08:00
for (i = 0; i < NR_CPUS; i++) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned long val;
[SPARC64]: Use machine description and OBP properly for cpu probing. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-26 06:49:59 +08:00
val = cpu_data(i).ecache_size;
if (!val)
continue;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (val > largest_size)
largest_size = val;
[SPARC64]: Use machine description and OBP properly for cpu probing. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-26 06:49:59 +08:00
val = cpu_data(i).ecache_line_size;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (val < smallest_linesize)
smallest_linesize = val;
[SPARC64]: Use machine description and OBP properly for cpu probing. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-26 06:49:59 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
if (largest_size == 0UL || smallest_linesize == ~0UL) {
prom_printf("cheetah_ecache_flush_init: Cannot probe cpu E-cache "
"parameters.\n");
prom_halt();
}
ecache_flush_size = (2 * largest_size);
ecache_flush_linesize = smallest_linesize;
[SPARC64]: Kill all external references to sp_banks[] Thus, we can mark sp_banks[] static in arch/sparc64/mm/init.c Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:46:43 +08:00
ecache_flush_physbase = find_ecache_flush_span(ecache_flush_size);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Kill all external references to sp_banks[] Thus, we can mark sp_banks[] static in arch/sparc64/mm/init.c Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:46:43 +08:00
if (ecache_flush_physbase == ~0UL) {
sparc: fix format string argument for prom_printf() prom_printf() takes printf style arguments. Specifing GCC's format attribute reveals that there are several wrong usages of prom_printf(). This fixes those wrong format strings and arguments, and also leaves format attributes in order to detect similar mistakes at compile time. Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: sparclinux@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net>
2012-09-29 11:14:49 +08:00
prom_printf("cheetah_ecache_flush_init: Cannot find %ld byte "
[SPARC64]: Kill all external references to sp_banks[] Thus, we can mark sp_banks[] static in arch/sparc64/mm/init.c Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:46:43 +08:00
"contiguous physical memory.\n",
ecache_flush_size);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
prom_halt();
}
/* Now allocate error trap reporting scoreboard. */
[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 14:34:02 +08:00
sz = NR_CPUS * (2 * sizeof(struct cheetah_err_info));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
for (order = 0; order < MAX_ORDER; order++) {
[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 14:34:02 +08:00
if ((PAGE_SIZE << order) >= sz)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
break;
}
cheetah_error_log = (struct cheetah_err_info *)
__get_free_pages(GFP_KERNEL, order);
if (!cheetah_error_log) {
prom_printf("cheetah_ecache_flush_init: Failed to allocate "
[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 14:34:02 +08:00
"error logging scoreboard (%d bytes).\n", sz);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
prom_halt();
}
memset(cheetah_error_log, 0, PAGE_SIZE << order);
/* Mark all AFSRs as invalid so that the trap handler will
* log new new information there.
*/
for (i = 0; i < 2 * NR_CPUS; i++)
cheetah_error_log[i].afsr = CHAFSR_INVALID;
__asm__ ("rdpr %%ver, %0" : "=r" (ver));
[SPARC64]: Refine code sequences to get the cpu id. On uniprocessor, it's always zero for optimize that. On SMP, the jmpl to the stub kills the return address stack in the cpu branch prediction logic, so expand the code sequence inline and use a code patching section to fix things up. This also always better and explicit register selection, which will be taken advantage of in a future changeset. The hard_smp_processor_id() function is big, so do not inline it. Fix up tests for Jalapeno to also test for Serrano chips too. These tests want "jbus Ultra-IIIi" cases to match, so that is what we should test for. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:27:19 +08:00
if ((ver >> 32) == __JALAPENO_ID ||
(ver >> 32) == __SERRANO_ID) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
cheetah_error_table = &__jalapeno_error_table[0];
cheetah_afsr_errors = JPAFSR_ERRORS;
} else if ((ver >> 32) == 0x003e0015) {
cheetah_error_table = &__cheetah_plus_error_table[0];
cheetah_afsr_errors = CHPAFSR_ERRORS;
} else {
cheetah_error_table = &__cheetah_error_table[0];
cheetah_afsr_errors = CHAFSR_ERRORS;
}
/* Now patch trap tables. */
memcpy(tl0_fecc, cheetah_fecc_trap_vector, (8 * 4));
memcpy(tl1_fecc, cheetah_fecc_trap_vector_tl1, (8 * 4));
memcpy(tl0_cee, cheetah_cee_trap_vector, (8 * 4));
memcpy(tl1_cee, cheetah_cee_trap_vector_tl1, (8 * 4));
memcpy(tl0_iae, cheetah_deferred_trap_vector, (8 * 4));
memcpy(tl1_iae, cheetah_deferred_trap_vector_tl1, (8 * 4));
memcpy(tl0_dae, cheetah_deferred_trap_vector, (8 * 4));
memcpy(tl1_dae, cheetah_deferred_trap_vector_tl1, (8 * 4));
if (tlb_type == cheetah_plus) {
memcpy(tl0_dcpe, cheetah_plus_dcpe_trap_vector, (8 * 4));
memcpy(tl1_dcpe, cheetah_plus_dcpe_trap_vector_tl1, (8 * 4));
memcpy(tl0_icpe, cheetah_plus_icpe_trap_vector, (8 * 4));
memcpy(tl1_icpe, cheetah_plus_icpe_trap_vector_tl1, (8 * 4));
}
flushi(PAGE_OFFSET);
}
static void cheetah_flush_ecache(void)
{
unsigned long flush_base = ecache_flush_physbase;
unsigned long flush_linesize = ecache_flush_linesize;
unsigned long flush_size = ecache_flush_size;
__asm__ __volatile__("1: subcc %0, %4, %0\n\t"
" bne,pt %%xcc, 1b\n\t"
" ldxa [%2 + %0] %3, %%g0\n\t"
: "=&r" (flush_size)
: "0" (flush_size), "r" (flush_base),
"i" (ASI_PHYS_USE_EC), "r" (flush_linesize));
}
static void cheetah_flush_ecache_line(unsigned long physaddr)
{
unsigned long alias;
physaddr &= ~(8UL - 1UL);
physaddr = (ecache_flush_physbase +
(physaddr & ((ecache_flush_size>>1UL) - 1UL)));
alias = physaddr + (ecache_flush_size >> 1UL);
__asm__ __volatile__("ldxa [%0] %2, %%g0\n\t"
"ldxa [%1] %2, %%g0\n\t"
"membar #Sync"
: /* no outputs */
: "r" (physaddr), "r" (alias),
"i" (ASI_PHYS_USE_EC));
}
/* Unfortunately, the diagnostic access to the I-cache tags we need to
* use to clear the thing interferes with I-cache coherency transactions.
*
* So we must only flush the I-cache when it is disabled.
*/
static void __cheetah_flush_icache(void)
{
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
unsigned int icache_size, icache_line_size;
unsigned long addr;
icache_size = local_cpu_data().icache_size;
icache_line_size = local_cpu_data().icache_line_size;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Clear the valid bits in all the tags. */
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
for (addr = 0; addr < icache_size; addr += icache_line_size) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
__asm__ __volatile__("stxa %%g0, [%0] %1\n\t"
"membar #Sync"
: /* no outputs */
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
: "r" (addr | (2 << 3)),
"i" (ASI_IC_TAG));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
static void cheetah_flush_icache(void)
{
unsigned long dcu_save;
/* Save current DCU, disable I-cache. */
__asm__ __volatile__("ldxa [%%g0] %1, %0\n\t"
"or %0, %2, %%g1\n\t"
"stxa %%g1, [%%g0] %1\n\t"
"membar #Sync"
: "=r" (dcu_save)
: "i" (ASI_DCU_CONTROL_REG), "i" (DCU_IC)
: "g1");
__cheetah_flush_icache();
/* Restore DCU register */
__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
"membar #Sync"
: /* no outputs */
: "r" (dcu_save), "i" (ASI_DCU_CONTROL_REG));
}
static void cheetah_flush_dcache(void)
{
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
unsigned int dcache_size, dcache_line_size;
unsigned long addr;
dcache_size = local_cpu_data().dcache_size;
dcache_line_size = local_cpu_data().dcache_line_size;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
for (addr = 0; addr < dcache_size; addr += dcache_line_size) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
__asm__ __volatile__("stxa %%g0, [%0] %1\n\t"
"membar #Sync"
: /* no outputs */
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
: "r" (addr), "i" (ASI_DCACHE_TAG));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
/* In order to make the even parity correct we must do two things.
* First, we clear DC_data_parity and set DC_utag to an appropriate value.
* Next, we clear out all 32-bytes of data for that line. Data of
* all-zero + tag parity value of zero == correct parity.
*/
static void cheetah_plus_zap_dcache_parity(void)
{
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
unsigned int dcache_size, dcache_line_size;
unsigned long addr;
dcache_size = local_cpu_data().dcache_size;
dcache_line_size = local_cpu_data().dcache_line_size;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
for (addr = 0; addr < dcache_size; addr += dcache_line_size) {
unsigned long tag = (addr >> 14);
unsigned long line;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
__asm__ __volatile__("membar #Sync\n\t"
"stxa %0, [%1] %2\n\t"
"membar #Sync"
: /* no outputs */
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
: "r" (tag), "r" (addr),
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
"i" (ASI_DCACHE_UTAG));
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
for (line = addr; line < addr + dcache_line_size; line += 8)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
__asm__ __volatile__("membar #Sync\n\t"
"stxa %%g0, [%0] %1\n\t"
"membar #Sync"
: /* no outputs */
[SPARC64]: Probe D/I/E-cache config and use. At boot time, determine the D-cache, I-cache and E-cache size and line-size. Use them in cache flushes when appropriate. This change was motivated by discovering that the D-cache on UltraSparc-IIIi and later are 64K not 32K, and the flushes done by the Cheetah error handlers were assuming a 32K size. There are still some pieces of code that are hard coding things and will need to be fixed up at some point. While we're here, fix the D-cache and I-cache parity error handlers to run with interrupts disabled, and when the trap occurs at trap level > 1 log the event via a counter displayed in /proc/cpuinfo. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-26 15:32:17 +08:00
: "r" (line),
"i" (ASI_DCACHE_DATA));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
/* Conversion tables used to frob Cheetah AFSR syndrome values into
* something palatable to the memory controller driver get_unumber
* routine.
*/
#define MT0 137
#define MT1 138
#define MT2 139
#define NONE 254
#define MTC0 140
#define MTC1 141
#define MTC2 142
#define MTC3 143
#define C0 128
#define C1 129
#define C2 130
#define C3 131
#define C4 132
#define C5 133
#define C6 134
#define C7 135
#define C8 136
#define M2 144
#define M3 145
#define M4 146
#define M 147
static unsigned char cheetah_ecc_syntab[] = {
/*00*/NONE, C0, C1, M2, C2, M2, M3, 47, C3, M2, M2, 53, M2, 41, 29, M,
/*01*/C4, M, M, 50, M2, 38, 25, M2, M2, 33, 24, M2, 11, M, M2, 16,
/*02*/C5, M, M, 46, M2, 37, 19, M2, M, 31, 32, M, 7, M2, M2, 10,
/*03*/M2, 40, 13, M2, 59, M, M2, 66, M, M2, M2, 0, M2, 67, 71, M,
/*04*/C6, M, M, 43, M, 36, 18, M, M2, 49, 15, M, 63, M2, M2, 6,
/*05*/M2, 44, 28, M2, M, M2, M2, 52, 68, M2, M2, 62, M2, M3, M3, M4,
/*06*/M2, 26, 106, M2, 64, M, M2, 2, 120, M, M2, M3, M, M3, M3, M4,
/*07*/116, M2, M2, M3, M2, M3, M, M4, M2, 58, 54, M2, M, M4, M4, M3,
/*08*/C7, M2, M, 42, M, 35, 17, M2, M, 45, 14, M2, 21, M2, M2, 5,
/*09*/M, 27, M, M, 99, M, M, 3, 114, M2, M2, 20, M2, M3, M3, M,
/*0a*/M2, 23, 113, M2, 112, M2, M, 51, 95, M, M2, M3, M2, M3, M3, M2,
/*0b*/103, M, M2, M3, M2, M3, M3, M4, M2, 48, M, M, 73, M2, M, M3,
/*0c*/M2, 22, 110, M2, 109, M2, M, 9, 108, M2, M, M3, M2, M3, M3, M,
/*0d*/102, M2, M, M, M2, M3, M3, M, M2, M3, M3, M2, M, M4, M, M3,
/*0e*/98, M, M2, M3, M2, M, M3, M4, M2, M3, M3, M4, M3, M, M, M,
/*0f*/M2, M3, M3, M, M3, M, M, M, 56, M4, M, M3, M4, M, M, M,
/*10*/C8, M, M2, 39, M, 34, 105, M2, M, 30, 104, M, 101, M, M, 4,
/*11*/M, M, 100, M, 83, M, M2, 12, 87, M, M, 57, M2, M, M3, M,
/*12*/M2, 97, 82, M2, 78, M2, M2, 1, 96, M, M, M, M, M, M3, M2,
/*13*/94, M, M2, M3, M2, M, M3, M, M2, M, 79, M, 69, M, M4, M,
/*14*/M2, 93, 92, M, 91, M, M2, 8, 90, M2, M2, M, M, M, M, M4,
/*15*/89, M, M, M3, M2, M3, M3, M, M, M, M3, M2, M3, M2, M, M3,
/*16*/86, M, M2, M3, M2, M, M3, M, M2, M, M3, M, M3, M, M, M3,
/*17*/M, M, M3, M2, M3, M2, M4, M, 60, M, M2, M3, M4, M, M, M2,
/*18*/M2, 88, 85, M2, 84, M, M2, 55, 81, M2, M2, M3, M2, M3, M3, M4,
/*19*/77, M, M, M, M2, M3, M, M, M2, M3, M3, M4, M3, M2, M, M,
/*1a*/74, M, M2, M3, M, M, M3, M, M, M, M3, M, M3, M, M4, M3,
/*1b*/M2, 70, 107, M4, 65, M2, M2, M, 127, M, M, M, M2, M3, M3, M,
/*1c*/80, M2, M2, 72, M, 119, 118, M, M2, 126, 76, M, 125, M, M4, M3,
/*1d*/M2, 115, 124, M, 75, M, M, M3, 61, M, M4, M, M4, M, M, M,
/*1e*/M, 123, 122, M4, 121, M4, M, M3, 117, M2, M2, M3, M4, M3, M, M,
/*1f*/111, M, M, M, M4, M3, M3, M, M, M, M3, M, M3, M2, M, M
};
static unsigned char cheetah_mtag_syntab[] = {
NONE, MTC0,
MTC1, NONE,
MTC2, NONE,
NONE, MT0,
MTC3, NONE,
NONE, MT1,
NONE, MT2,
NONE, NONE
};
/* Return the highest priority error conditon mentioned. */
[SPARC64]: __inline__ --> inline Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-27 15:13:04 +08:00
static inline unsigned long cheetah_get_hipri(unsigned long afsr)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
unsigned long tmp = 0;
int i;
for (i = 0; cheetah_error_table[i].mask; i++) {
if ((tmp = (afsr & cheetah_error_table[i].mask)) != 0UL)
return tmp;
}
return tmp;
}
static const char *cheetah_get_string(unsigned long bit)
{
int i;
for (i = 0; cheetah_error_table[i].mask; i++) {
if ((bit & cheetah_error_table[i].mask) != 0UL)
return cheetah_error_table[i].name;
}
return "???";
}
static void cheetah_log_errors(struct pt_regs *regs, struct cheetah_err_info *info,
unsigned long afsr, unsigned long afar, int recoverable)
{
unsigned long hipri;
char unum[256];
printk("%s" "ERROR(%d): Cheetah error trap taken afsr[%016lx] afar[%016lx] TL1(%d)\n",
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
afsr, afar,
(afsr & CHAFSR_TL1) ? 1 : 0);
[SPARC64]: Print out return PC in cheetah_log_errors(). This makes debugging things a little bit easier. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-04-02 15:29:56 +08:00
printk("%s" "ERROR(%d): TPC[%lx] TNPC[%lx] O7[%lx] TSTATE[%lx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
[SPARC64]: Print out return PC in cheetah_log_errors(). This makes debugging things a little bit easier. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-04-02 15:29:56 +08:00
regs->tpc, regs->tnpc, regs->u_regs[UREG_I7], regs->tstate);
[SPARC64]: Add some missing print_symbol() calls. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-10-30 17:10:20 +08:00
printk("%s" "ERROR(%d): ",
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id());
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk("TPC<%pS>\n", (void *) regs->tpc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
printk("%s" "ERROR(%d): M_SYND(%lx), E_SYND(%lx)%s%s\n",
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
(afsr & CHAFSR_M_SYNDROME) >> CHAFSR_M_SYNDROME_SHIFT,
(afsr & CHAFSR_E_SYNDROME) >> CHAFSR_E_SYNDROME_SHIFT,
(afsr & CHAFSR_ME) ? ", Multiple Errors" : "",
(afsr & CHAFSR_PRIV) ? ", Privileged" : "");
hipri = cheetah_get_hipri(afsr);
printk("%s" "ERROR(%d): Highest priority error (%016lx) \"%s\"\n",
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
hipri, cheetah_get_string(hipri));
/* Try to get unumber if relevant. */
#define ESYND_ERRORS (CHAFSR_IVC | CHAFSR_IVU | \
CHAFSR_CPC | CHAFSR_CPU | \
CHAFSR_UE | CHAFSR_CE | \
CHAFSR_EDC | CHAFSR_EDU | \
CHAFSR_UCC | CHAFSR_UCU | \
CHAFSR_WDU | CHAFSR_WDC)
#define MSYND_ERRORS (CHAFSR_EMC | CHAFSR_EMU)
if (afsr & ESYND_ERRORS) {
int syndrome;
int ret;
syndrome = (afsr & CHAFSR_E_SYNDROME) >> CHAFSR_E_SYNDROME_SHIFT;
syndrome = cheetah_ecc_syntab[syndrome];
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
ret = sprintf_dimm(syndrome, afar, unum, sizeof(unum));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (ret != -1)
printk("%s" "ERROR(%d): AFAR E-syndrome [%s]\n",
(recoverable ? KERN_WARNING : KERN_CRIT),
smp_processor_id(), unum);
} else if (afsr & MSYND_ERRORS) {
int syndrome;
int ret;
syndrome = (afsr & CHAFSR_M_SYNDROME) >> CHAFSR_M_SYNDROME_SHIFT;
syndrome = cheetah_mtag_syntab[syndrome];
sparc64: Add generic interface for registering a dimm printing handler. The way to do this varies by platform type and the exact memory controller the cpu uses. For Spitfire cpus we currently just use prom_getunumber() and hope that works. For Cheetah cpus we have a memory controller driver that can compute this information. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-25 13:08:34 +08:00
ret = sprintf_dimm(syndrome, afar, unum, sizeof(unum));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (ret != -1)
printk("%s" "ERROR(%d): AFAR M-syndrome [%s]\n",
(recoverable ? KERN_WARNING : KERN_CRIT),
smp_processor_id(), unum);
}
/* Now dump the cache snapshots. */
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): D-cache idx[%x] tag[%016llx] utag[%016llx] stag[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
(int) info->dcache_index,
info->dcache_tag,
info->dcache_utag,
info->dcache_stag);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): D-cache data0[%016llx] data1[%016llx] data2[%016llx] data3[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
info->dcache_data[0],
info->dcache_data[1],
info->dcache_data[2],
info->dcache_data[3]);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): I-cache idx[%x] tag[%016llx] utag[%016llx] stag[%016llx] "
"u[%016llx] l[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
(int) info->icache_index,
info->icache_tag,
info->icache_utag,
info->icache_stag,
info->icache_upper,
info->icache_lower);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): I-cache INSN0[%016llx] INSN1[%016llx] INSN2[%016llx] INSN3[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
info->icache_data[0],
info->icache_data[1],
info->icache_data[2],
info->icache_data[3]);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): I-cache INSN4[%016llx] INSN5[%016llx] INSN6[%016llx] INSN7[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
info->icache_data[4],
info->icache_data[5],
info->icache_data[6],
info->icache_data[7]);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): E-cache idx[%x] tag[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
(int) info->ecache_index, info->ecache_tag);
sparc64: Use unsigned long long for u64. Andrew Morton wrote: People keep on doing printk("%llu", some_u64); testing it only on x86_64 and this generates a warning storm on powerpc, sparc64, etc. Because they use `long', not `long long'. Quite a few 64-bit architectures are using `long' for their s64/u64 types. We should convert them all to `long long'. Update types.h so we use unsigned long long for u64 and fix all warnings in sparc64 code. Tested with an allnoconfig, defconfig and allmodconfig builds. This patch introduces additional warnings in several drivers. These will be dealt with in separate patches. Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-07 05:19:28 +08:00
printk("%s" "ERROR(%d): E-cache data0[%016llx] data1[%016llx] data2[%016llx] data3[%016llx]\n",
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
(recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
info->ecache_data[0],
info->ecache_data[1],
info->ecache_data[2],
info->ecache_data[3]);
afsr = (afsr & ~hipri) & cheetah_afsr_errors;
while (afsr != 0UL) {
unsigned long bit = cheetah_get_hipri(afsr);
printk("%s" "ERROR: Multiple-error (%016lx) \"%s\"\n",
(recoverable ? KERN_WARNING : KERN_CRIT),
bit, cheetah_get_string(bit));
afsr &= ~bit;
}
if (!recoverable)
printk(KERN_CRIT "ERROR: This condition is not recoverable.\n");
}
static int cheetah_recheck_errors(struct cheetah_err_info *logp)
{
unsigned long afsr, afar;
int ret = 0;
__asm__ __volatile__("ldxa [%%g0] %1, %0\n\t"
: "=r" (afsr)
: "i" (ASI_AFSR));
if ((afsr & cheetah_afsr_errors) != 0) {
if (logp != NULL) {
__asm__ __volatile__("ldxa [%%g0] %1, %0\n\t"
: "=r" (afar)
: "i" (ASI_AFAR));
logp->afsr = afsr;
logp->afar = afar;
}
ret = 1;
}
__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
"membar #Sync\n\t"
: : "r" (afsr), "i" (ASI_AFSR));
return ret;
}
void cheetah_fecc_handler(struct pt_regs *regs, unsigned long afsr, unsigned long afar)
{
struct cheetah_err_info local_snapshot, *p;
int recoverable;
/* Flush E-cache */
cheetah_flush_ecache();
p = cheetah_get_error_log(afsr);
if (!p) {
prom_printf("ERROR: Early Fast-ECC error afsr[%016lx] afar[%016lx]\n",
afsr, afar);
prom_printf("ERROR: CPU(%d) TPC[%016lx] TNPC[%016lx] TSTATE[%016lx]\n",
smp_processor_id(), regs->tpc, regs->tnpc, regs->tstate);
prom_halt();
}
/* Grab snapshot of logged error. */
memcpy(&local_snapshot, p, sizeof(local_snapshot));
/* If the current trap snapshot does not match what the
* trap handler passed along into our args, big trouble.
* In such a case, mark the local copy as invalid.
*
* Else, it matches and we mark the afsr in the non-local
* copy as invalid so we may log new error traps there.
*/
if (p->afsr != afsr || p->afar != afar)
local_snapshot.afsr = CHAFSR_INVALID;
else
p->afsr = CHAFSR_INVALID;
cheetah_flush_icache();
cheetah_flush_dcache();
/* Re-enable I-cache/D-cache */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_DCU_CONTROL_REG),
"i" (DCU_DC | DCU_IC)
: "g1");
/* Re-enable error reporting */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_ESTATE_ERROR_EN),
"i" (ESTATE_ERROR_NCEEN | ESTATE_ERROR_CEEN)
: "g1");
/* Decide if we can continue after handling this trap and
* logging the error.
*/
recoverable = 1;
if (afsr & (CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP))
recoverable = 0;
/* Re-check AFSR/AFAR. What we are looking for here is whether a new
* error was logged while we had error reporting traps disabled.
*/
if (cheetah_recheck_errors(&local_snapshot)) {
unsigned long new_afsr = local_snapshot.afsr;
/* If we got a new asynchronous error, die... */
if (new_afsr & (CHAFSR_EMU | CHAFSR_EDU |
CHAFSR_WDU | CHAFSR_CPU |
CHAFSR_IVU | CHAFSR_UE |
CHAFSR_BERR | CHAFSR_TO))
recoverable = 0;
}
/* Log errors. */
cheetah_log_errors(regs, &local_snapshot, afsr, afar, recoverable);
if (!recoverable)
panic("Irrecoverable Fast-ECC error trap.\n");
/* Flush E-cache to kick the error trap handlers out. */
cheetah_flush_ecache();
}
/* Try to fix a correctable error by pushing the line out from
* the E-cache. Recheck error reporting registers to see if the
* problem is intermittent.
*/
static int cheetah_fix_ce(unsigned long physaddr)
{
unsigned long orig_estate;
unsigned long alias1, alias2;
int ret;
/* Make sure correctable error traps are disabled. */
__asm__ __volatile__("ldxa [%%g0] %2, %0\n\t"
"andn %0, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %2\n\t"
"membar #Sync"
: "=&r" (orig_estate)
: "i" (ESTATE_ERROR_CEEN),
"i" (ASI_ESTATE_ERROR_EN)
: "g1");
/* We calculate alias addresses that will force the
* cache line in question out of the E-cache. Then
* we bring it back in with an atomic instruction so
* that we get it in some modified/exclusive state,
* then we displace it again to try and get proper ECC
* pushed back into the system.
*/
physaddr &= ~(8UL - 1UL);
alias1 = (ecache_flush_physbase +
(physaddr & ((ecache_flush_size >> 1) - 1)));
alias2 = alias1 + (ecache_flush_size >> 1);
__asm__ __volatile__("ldxa [%0] %3, %%g0\n\t"
"ldxa [%1] %3, %%g0\n\t"
"casxa [%2] %3, %%g0, %%g0\n\t"
"ldxa [%0] %3, %%g0\n\t"
"ldxa [%1] %3, %%g0\n\t"
"membar #Sync"
: /* no outputs */
: "r" (alias1), "r" (alias2),
"r" (physaddr), "i" (ASI_PHYS_USE_EC));
/* Did that trigger another error? */
if (cheetah_recheck_errors(NULL)) {
/* Try one more time. */
__asm__ __volatile__("ldxa [%0] %1, %%g0\n\t"
"membar #Sync"
: : "r" (physaddr), "i" (ASI_PHYS_USE_EC));
if (cheetah_recheck_errors(NULL))
ret = 2;
else
ret = 1;
} else {
/* No new error, intermittent problem. */
ret = 0;
}
/* Restore error enables. */
__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
"membar #Sync"
: : "r" (orig_estate), "i" (ASI_ESTATE_ERROR_EN));
return ret;
}
/* Return non-zero if PADDR is a valid physical memory address. */
static int cheetah_check_main_memory(unsigned long paddr)
{
[SPARC64]: Kill all external references to sp_banks[] Thus, we can mark sp_banks[] static in arch/sparc64/mm/init.c Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:46:43 +08:00
unsigned long vaddr = PAGE_OFFSET + paddr;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Rewrite convoluted physical memory probing. Delete all of the code working with sp_banks[] and replace with clean acquisition and sorting of physical memory parameters from the firmware. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-30 08:58:26 +08:00
if (vaddr > (unsigned long) high_memory)
[SPARC64]: Solidify check in cheetah_check_main_memory(). Need to make sure the address is below high_memory before passing it to kern_addr_valid(). Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:48:25 +08:00
return 0;
[SPARC64]: Kill all external references to sp_banks[] Thus, we can mark sp_banks[] static in arch/sparc64/mm/init.c Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 12:46:43 +08:00
return kern_addr_valid(vaddr);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void cheetah_cee_handler(struct pt_regs *regs, unsigned long afsr, unsigned long afar)
{
struct cheetah_err_info local_snapshot, *p;
int recoverable, is_memory;
p = cheetah_get_error_log(afsr);
if (!p) {
prom_printf("ERROR: Early CEE error afsr[%016lx] afar[%016lx]\n",
afsr, afar);
prom_printf("ERROR: CPU(%d) TPC[%016lx] TNPC[%016lx] TSTATE[%016lx]\n",
smp_processor_id(), regs->tpc, regs->tnpc, regs->tstate);
prom_halt();
}
/* Grab snapshot of logged error. */
memcpy(&local_snapshot, p, sizeof(local_snapshot));
/* If the current trap snapshot does not match what the
* trap handler passed along into our args, big trouble.
* In such a case, mark the local copy as invalid.
*
* Else, it matches and we mark the afsr in the non-local
* copy as invalid so we may log new error traps there.
*/
if (p->afsr != afsr || p->afar != afar)
local_snapshot.afsr = CHAFSR_INVALID;
else
p->afsr = CHAFSR_INVALID;
is_memory = cheetah_check_main_memory(afar);
if (is_memory && (afsr & CHAFSR_CE) != 0UL) {
/* XXX Might want to log the results of this operation
* XXX somewhere... -DaveM
*/
cheetah_fix_ce(afar);
}
{
int flush_all, flush_line;
flush_all = flush_line = 0;
if ((afsr & CHAFSR_EDC) != 0UL) {
if ((afsr & cheetah_afsr_errors) == CHAFSR_EDC)
flush_line = 1;
else
flush_all = 1;
} else if ((afsr & CHAFSR_CPC) != 0UL) {
if ((afsr & cheetah_afsr_errors) == CHAFSR_CPC)
flush_line = 1;
else
flush_all = 1;
}
/* Trap handler only disabled I-cache, flush it. */
cheetah_flush_icache();
/* Re-enable I-cache */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_DCU_CONTROL_REG),
"i" (DCU_IC)
: "g1");
if (flush_all)
cheetah_flush_ecache();
else if (flush_line)
cheetah_flush_ecache_line(afar);
}
/* Re-enable error reporting */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_ESTATE_ERROR_EN),
"i" (ESTATE_ERROR_CEEN)
: "g1");
/* Decide if we can continue after handling this trap and
* logging the error.
*/
recoverable = 1;
if (afsr & (CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP))
recoverable = 0;
/* Re-check AFSR/AFAR */
(void) cheetah_recheck_errors(&local_snapshot);
/* Log errors. */
cheetah_log_errors(regs, &local_snapshot, afsr, afar, recoverable);
if (!recoverable)
panic("Irrecoverable Correctable-ECC error trap.\n");
}
void cheetah_deferred_handler(struct pt_regs *regs, unsigned long afsr, unsigned long afar)
{
struct cheetah_err_info local_snapshot, *p;
int recoverable, is_memory;
#ifdef CONFIG_PCI
/* Check for the special PCI poke sequence. */
if (pci_poke_in_progress && pci_poke_cpu == smp_processor_id()) {
cheetah_flush_icache();
cheetah_flush_dcache();
/* Re-enable I-cache/D-cache */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_DCU_CONTROL_REG),
"i" (DCU_DC | DCU_IC)
: "g1");
/* Re-enable error reporting */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_ESTATE_ERROR_EN),
"i" (ESTATE_ERROR_NCEEN | ESTATE_ERROR_CEEN)
: "g1");
(void) cheetah_recheck_errors(NULL);
pci_poke_faulted = 1;
regs->tpc += 4;
regs->tnpc = regs->tpc + 4;
return;
}
#endif
p = cheetah_get_error_log(afsr);
if (!p) {
prom_printf("ERROR: Early deferred error afsr[%016lx] afar[%016lx]\n",
afsr, afar);
prom_printf("ERROR: CPU(%d) TPC[%016lx] TNPC[%016lx] TSTATE[%016lx]\n",
smp_processor_id(), regs->tpc, regs->tnpc, regs->tstate);
prom_halt();
}
/* Grab snapshot of logged error. */
memcpy(&local_snapshot, p, sizeof(local_snapshot));
/* If the current trap snapshot does not match what the
* trap handler passed along into our args, big trouble.
* In such a case, mark the local copy as invalid.
*
* Else, it matches and we mark the afsr in the non-local
* copy as invalid so we may log new error traps there.
*/
if (p->afsr != afsr || p->afar != afar)
local_snapshot.afsr = CHAFSR_INVALID;
else
p->afsr = CHAFSR_INVALID;
is_memory = cheetah_check_main_memory(afar);
{
int flush_all, flush_line;
flush_all = flush_line = 0;
if ((afsr & CHAFSR_EDU) != 0UL) {
if ((afsr & cheetah_afsr_errors) == CHAFSR_EDU)
flush_line = 1;
else
flush_all = 1;
} else if ((afsr & CHAFSR_BERR) != 0UL) {
if ((afsr & cheetah_afsr_errors) == CHAFSR_BERR)
flush_line = 1;
else
flush_all = 1;
}
cheetah_flush_icache();
cheetah_flush_dcache();
/* Re-enable I/D caches */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_DCU_CONTROL_REG),
"i" (DCU_IC | DCU_DC)
: "g1");
if (flush_all)
cheetah_flush_ecache();
else if (flush_line)
cheetah_flush_ecache_line(afar);
}
/* Re-enable error reporting */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_ESTATE_ERROR_EN),
"i" (ESTATE_ERROR_NCEEN | ESTATE_ERROR_CEEN)
: "g1");
/* Decide if we can continue after handling this trap and
* logging the error.
*/
recoverable = 1;
if (afsr & (CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP))
recoverable = 0;
/* Re-check AFSR/AFAR. What we are looking for here is whether a new
* error was logged while we had error reporting traps disabled.
*/
if (cheetah_recheck_errors(&local_snapshot)) {
unsigned long new_afsr = local_snapshot.afsr;
/* If we got a new asynchronous error, die... */
if (new_afsr & (CHAFSR_EMU | CHAFSR_EDU |
CHAFSR_WDU | CHAFSR_CPU |
CHAFSR_IVU | CHAFSR_UE |
CHAFSR_BERR | CHAFSR_TO))
recoverable = 0;
}
/* Log errors. */
cheetah_log_errors(regs, &local_snapshot, afsr, afar, recoverable);
/* "Recoverable" here means we try to yank the page from ever
* being newly used again. This depends upon a few things:
* 1) Must be main memory, and AFAR must be valid.
* 2) If we trapped from user, OK.
* 3) Else, if we trapped from kernel we must find exception
* table entry (ie. we have to have been accessing user
* space).
*
* If AFAR is not in main memory, or we trapped from kernel
* and cannot find an exception table entry, it is unacceptable
* to try and continue.
*/
if (recoverable && is_memory) {
if ((regs->tstate & TSTATE_PRIV) == 0UL) {
/* OK, usermode access. */
recoverable = 1;
} else {
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
const struct exception_table_entry *entry;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
entry = search_exception_tables(regs->tpc);
if (entry) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* OK, kernel access to userspace. */
recoverable = 1;
} else {
/* BAD, privileged state is corrupted. */
recoverable = 0;
}
if (recoverable) {
if (pfn_valid(afar >> PAGE_SHIFT))
get_page(pfn_to_page(afar >> PAGE_SHIFT));
else
recoverable = 0;
/* Only perform fixup if we still have a
* recoverable condition.
*/
if (recoverable) {
[SPARC64]: Convert to use generic exception table support. The funny "range" exception table entries we had were only used by the compat layer socketcall assembly, and it wasn't even needed there. For free we now get proper exception table sorting and fast binary searching. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-09-29 11:21:11 +08:00
regs->tpc = entry->fixup;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
regs->tnpc = regs->tpc + 4;
}
}
}
} else {
recoverable = 0;
}
if (!recoverable)
panic("Irrecoverable deferred error trap.\n");
}
/* Handle a D/I cache parity error trap. TYPE is encoded as:
*
* Bit0: 0=dcache,1=icache
* Bit1: 0=recoverable,1=unrecoverable
*
* The hardware has disabled both the I-cache and D-cache in
* the %dcr register.
*/
void cheetah_plus_parity_error(int type, struct pt_regs *regs)
{
if (type & 0x1)
__cheetah_flush_icache();
else
cheetah_plus_zap_dcache_parity();
cheetah_flush_dcache();
/* Re-enable I-cache/D-cache */
__asm__ __volatile__("ldxa [%%g0] %0, %%g1\n\t"
"or %%g1, %1, %%g1\n\t"
"stxa %%g1, [%%g0] %0\n\t"
"membar #Sync"
: /* no outputs */
: "i" (ASI_DCU_CONTROL_REG),
"i" (DCU_DC | DCU_IC)
: "g1");
if (type & 0x2) {
printk(KERN_EMERG "CPU[%d]: Cheetah+ %c-cache parity error at TPC[%016lx]\n",
smp_processor_id(),
(type & 0x1) ? 'I' : 'D',
regs->tpc);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_EMERG "TPC<%pS>\n", (void *) regs->tpc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
panic("Irrecoverable Cheetah+ parity error.");
}
printk(KERN_WARNING "CPU[%d]: Cheetah+ %c-cache parity error at TPC[%016lx]\n",
smp_processor_id(),
(type & 0x1) ? 'I' : 'D',
regs->tpc);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_WARNING "TPC<%pS>\n", (void *) regs->tpc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
struct sun4v_error_entry {
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
/* Unique error handle */
/*0x00*/u64 err_handle;
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
/* %stick value at the time of the error */
/*0x08*/u64 err_stick;
/*0x10*/u8 reserved_1[3];
/* Error type */
/*0x13*/u8 err_type;
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
#define SUN4V_ERR_TYPE_UNDEFINED 0
#define SUN4V_ERR_TYPE_UNCORRECTED_RES 1
#define SUN4V_ERR_TYPE_PRECISE_NONRES 2
#define SUN4V_ERR_TYPE_DEFERRED_NONRES 3
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
#define SUN4V_ERR_TYPE_SHUTDOWN_RQST 4
#define SUN4V_ERR_TYPE_DUMP_CORE 5
#define SUN4V_ERR_TYPE_SP_STATE_CHANGE 6
#define SUN4V_ERR_TYPE_NUM 7
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
/* Error attributes */
/*0x14*/u32 err_attrs;
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
#define SUN4V_ERR_ATTRS_PROCESSOR 0x00000001
#define SUN4V_ERR_ATTRS_MEMORY 0x00000002
#define SUN4V_ERR_ATTRS_PIO 0x00000004
#define SUN4V_ERR_ATTRS_INT_REGISTERS 0x00000008
#define SUN4V_ERR_ATTRS_FPU_REGISTERS 0x00000010
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
#define SUN4V_ERR_ATTRS_SHUTDOWN_RQST 0x00000020
#define SUN4V_ERR_ATTRS_ASR 0x00000040
#define SUN4V_ERR_ATTRS_ASI 0x00000080
#define SUN4V_ERR_ATTRS_PRIV_REG 0x00000100
#define SUN4V_ERR_ATTRS_SPSTATE_MSK 0x00000600
#define SUN4V_ERR_ATTRS_SPSTATE_SHFT 9
#define SUN4V_ERR_ATTRS_MODE_MSK 0x03000000
#define SUN4V_ERR_ATTRS_MODE_SHFT 24
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
#define SUN4V_ERR_ATTRS_RES_QUEUE_FULL 0x80000000
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
#define SUN4V_ERR_SPSTATE_FAULTED 0
#define SUN4V_ERR_SPSTATE_AVAILABLE 1
#define SUN4V_ERR_SPSTATE_NOT_PRESENT 2
#define SUN4V_ERR_MODE_USER 1
#define SUN4V_ERR_MODE_PRIV 2
/* Real address of the memory region or PIO transaction */
/*0x18*/u64 err_raddr;
/* Size of the operation triggering the error, in bytes */
/*0x20*/u32 err_size;
/* ID of the CPU */
/*0x24*/u16 err_cpu;
/* Grace periof for shutdown, in seconds */
/*0x26*/u16 err_secs;
/* Value of the %asi register */
/*0x28*/u8 err_asi;
/*0x29*/u8 reserved_2;
/* Value of the ASR register number */
/*0x2a*/u16 err_asr;
#define SUN4V_ERR_ASR_VALID 0x8000
/*0x2c*/u32 reserved_3;
/*0x30*/u64 reserved_4;
/*0x38*/u64 reserved_5;
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
};
static atomic_t sun4v_resum_oflow_cnt = ATOMIC_INIT(0);
static atomic_t sun4v_nonresum_oflow_cnt = ATOMIC_INIT(0);
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
static const char *sun4v_err_type_to_str(u8 type)
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
{
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
static const char *types[SUN4V_ERR_TYPE_NUM] = {
"undefined",
"uncorrected resumable",
"precise nonresumable",
"deferred nonresumable",
"shutdown request",
"dump core",
"SP state change",
};
if (type < SUN4V_ERR_TYPE_NUM)
return types[type];
return "unknown";
}
static void sun4v_emit_err_attr_strings(u32 attrs)
{
static const char *attr_names[] = {
"processor",
"memory",
"PIO",
"int-registers",
"fpu-registers",
"shutdown-request",
"ASR",
"ASI",
"priv-reg",
};
static const char *sp_states[] = {
"sp-faulted",
"sp-available",
"sp-not-present",
"sp-state-reserved",
};
static const char *modes[] = {
"mode-reserved0",
"user",
"priv",
"mode-reserved1",
};
u32 sp_state, mode;
int i;
for (i = 0; i < ARRAY_SIZE(attr_names); i++) {
if (attrs & (1U << i)) {
const char *s = attr_names[i];
pr_cont("%s ", s);
}
sparc: Remove unnecessary semicolons Semicolons are not necessary after switch/while/for/if braces so remove them. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-06-03 22:45:23 +08:00
}
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
sp_state = ((attrs & SUN4V_ERR_ATTRS_SPSTATE_MSK) >>
SUN4V_ERR_ATTRS_SPSTATE_SHFT);
pr_cont("%s ", sp_states[sp_state]);
mode = ((attrs & SUN4V_ERR_ATTRS_MODE_MSK) >>
SUN4V_ERR_ATTRS_MODE_SHFT);
pr_cont("%s ", modes[mode]);
if (attrs & SUN4V_ERR_ATTRS_RES_QUEUE_FULL)
pr_cont("res-queue-full ");
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
}
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
/* When the report contains a real-address of "-1" it means that the
* hardware did not provide the address. So we compute the effective
* address of the load or store instruction at regs->tpc and report
* that. Usually when this happens it's a PIO and in such a case we
* are using physical addresses with bypass ASIs anyways, so what we
* report here is exactly what we want.
*/
static void sun4v_report_real_raddr(const char *pfx, struct pt_regs *regs)
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
{
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
unsigned int insn;
u64 addr;
if (!(regs->tstate & TSTATE_PRIV))
return;
insn = *(unsigned int *) regs->tpc;
addr = compute_effective_address(regs, insn, 0);
printk("%s: insn effective address [0x%016llx]\n",
pfx, addr);
}
static void sun4v_log_error(struct pt_regs *regs, struct sun4v_error_entry *ent,
int cpu, const char *pfx, atomic_t *ocnt)
{
u64 *raw_ptr = (u64 *) ent;
u32 attrs;
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
int cnt;
printk("%s: Reporting on cpu %d\n", pfx, cpu);
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
printk("%s: TPC [0x%016lx] <%pS>\n",
pfx, regs->tpc, (void *) regs->tpc);
printk("%s: RAW [%016llx:%016llx:%016llx:%016llx\n",
pfx, raw_ptr[0], raw_ptr[1], raw_ptr[2], raw_ptr[3]);
printk("%s: %016llx:%016llx:%016llx:%016llx]\n",
pfx, raw_ptr[4], raw_ptr[5], raw_ptr[6], raw_ptr[7]);
printk("%s: handle [0x%016llx] stick [0x%016llx]\n",
pfx, ent->err_handle, ent->err_stick);
printk("%s: type [%s]\n", pfx, sun4v_err_type_to_str(ent->err_type));
attrs = ent->err_attrs;
printk("%s: attrs [0x%08x] < ", pfx, attrs);
sun4v_emit_err_attr_strings(attrs);
pr_cont(">\n");
/* Various fields in the error report are only valid if
* certain attribute bits are set.
*/
if (attrs & (SUN4V_ERR_ATTRS_MEMORY |
SUN4V_ERR_ATTRS_PIO |
SUN4V_ERR_ATTRS_ASI)) {
printk("%s: raddr [0x%016llx]\n", pfx, ent->err_raddr);
if (ent->err_raddr == ~(u64)0)
sun4v_report_real_raddr(pfx, regs);
}
if (attrs & (SUN4V_ERR_ATTRS_MEMORY | SUN4V_ERR_ATTRS_ASI))
printk("%s: size [0x%x]\n", pfx, ent->err_size);
if (attrs & (SUN4V_ERR_ATTRS_PROCESSOR |
SUN4V_ERR_ATTRS_INT_REGISTERS |
SUN4V_ERR_ATTRS_FPU_REGISTERS |
SUN4V_ERR_ATTRS_PRIV_REG))
printk("%s: cpu[%u]\n", pfx, ent->err_cpu);
if (attrs & SUN4V_ERR_ATTRS_ASI)
printk("%s: asi [0x%02x]\n", pfx, ent->err_asi);
if ((attrs & (SUN4V_ERR_ATTRS_INT_REGISTERS |
SUN4V_ERR_ATTRS_FPU_REGISTERS |
SUN4V_ERR_ATTRS_PRIV_REG)) &&
(ent->err_asr & SUN4V_ERR_ASR_VALID) != 0)
printk("%s: reg [0x%04x]\n",
pfx, ent->err_asr & ~SUN4V_ERR_ASR_VALID);
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
sparc64: Kill __show_regs(). The story is that what we used to do when we actually used smp_report_regs() is that if you specifically only wanted to have the current cpu's registers dumped you would call "__show_regs()" otherwise you would call show_regs() which also invoked smp_report_regs(). Now that we killed off smp_report_regs() there is no longer any reason to have these two routines, just show_regs() is sufficient. Also kill off a stray declaration of show_regs() in sparc64_ksym.c Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-01 11:33:43 +08:00
show_regs(regs);
[SPARC64]: Dump local cpu registers in sun4v_log_error() This makes the debugging information more usable. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-07 08:37:41 +08:00
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
if ((cnt = atomic_read(ocnt)) != 0) {
atomic_set(ocnt, 0);
wmb();
printk("%s: Queue overflowed %d times.\n",
pfx, cnt);
}
}
sparc64: Make %pil level 15 a pseudo-NMI. So that we can profile code even in a local_irq_disable() section, only write 14 (instead of 15) into the %pil register to disable IRQs. This allows PIL level 15 to serve as a pseudo NMI. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 13:55:29 +08:00
/* We run with %pil set to PIL_NORMAL_MAX and PSTATE_IE enabled in %pstate.
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
* Log the event and clear the first word of the entry.
*/
void sun4v_resum_error(struct pt_regs *regs, unsigned long offset)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
struct sun4v_error_entry *ent, local_copy;
struct trap_per_cpu *tb;
unsigned long paddr;
int cpu;
cpu = get_cpu();
tb = &trap_block[cpu];
paddr = tb->resum_kernel_buf_pa + offset;
ent = __va(paddr);
memcpy(&local_copy, ent, sizeof(struct sun4v_error_entry));
/* We have a local copy now, so release the entry. */
ent->err_handle = 0;
wmb();
put_cpu();
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
if (local_copy.err_type == SUN4V_ERR_TYPE_SHUTDOWN_RQST) {
/* We should really take the seconds field of
* the error report and use it for the shutdown
* invocation, but for now do the same thing we
* do for a DS shutdown request.
[SPARC64]: Run ctrl-alt-del action for sun4v powerdown request. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-30 13:16:21 +08:00
*/
sparc64: Fix deficiencies in sun4v error reporting. Missing error types, attributes, and report fields. Pad out to 64-bytes. Make string reporting cleaner and easier to extend in the future using "const char *" arrays that index by either bit position, or absolute field value. Report the raw 64-byte error report as a sequence of u64s before the annotated version. Only report fields which are valid, given the context and the attribute bits which are set. For shutdown requests, use the local copy of the error report not the one we just freed up back to the queue. Also, use orderly_poweroff() just like the Domain Services shutdown request code does. If the real-address reported is "-1" (unknown) try to disassemble the instruction to report the effective address of the access. Only do this in privileged mode. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-11 08:19:32 +08:00
pr_info("Shutdown request, %u seconds...\n",
local_copy.err_secs);
orderly_poweroff(true);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
[SPARC64]: Run ctrl-alt-del action for sun4v powerdown request. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-30 13:16:21 +08:00
}
[SPARC64]: Dump local cpu registers in sun4v_log_error() This makes the debugging information more usable. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-07 08:37:41 +08:00
sun4v_log_error(regs, &local_copy, cpu,
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
KERN_ERR "RESUMABLE ERROR",
&sun4v_resum_oflow_cnt);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
}
/* If we try to printk() we'll probably make matters worse, by trying
* to retake locks this cpu already holds or causing more errors. So
* just bump a counter, and we'll report these counter bumps above.
*/
void sun4v_resum_overflow(struct pt_regs *regs)
{
atomic_inc(&sun4v_resum_oflow_cnt);
}
sparc64: Handle PIO & MEM non-resumable errors. User processes trying to access an invalid memory address via PIO will receive a SIGBUS signal instead of causing a panic. Memory errors will receive a SIGKILL since a SIGBUS may result in a coredump which may attempt to repeat the faulting access. Signed-off-by: Liam R. Howlett <Liam.Howlett@Oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-17 23:59:03 +08:00
/* Given a set of registers, get the virtual addressi that was being accessed
* by the faulting instructions at tpc.
*/
static unsigned long sun4v_get_vaddr(struct pt_regs *regs)
{
unsigned int insn;
if (!copy_from_user(&insn, (void __user *)regs->tpc, 4)) {
return compute_effective_address(regs, insn,
(insn >> 25) & 0x1f);
}
return 0;
}
/* Attempt to handle non-resumable errors generated from userspace.
* Returns true if the signal was handled, false otherwise.
*/
bool sun4v_nonresum_error_user_handled(struct pt_regs *regs,
struct sun4v_error_entry *ent) {
unsigned int attrs = ent->err_attrs;
if (attrs & SUN4V_ERR_ATTRS_MEMORY) {
unsigned long addr = ent->err_raddr;
siginfo_t info;
if (addr == ~(u64)0) {
/* This seems highly unlikely to ever occur */
pr_emerg("SUN4V NON-RECOVERABLE ERROR: Memory error detected in unknown location!\n");
} else {
unsigned long page_cnt = DIV_ROUND_UP(ent->err_size,
PAGE_SIZE);
/* Break the unfortunate news. */
pr_emerg("SUN4V NON-RECOVERABLE ERROR: Memory failed at %016lX\n",
addr);
pr_emerg("SUN4V NON-RECOVERABLE ERROR: Claiming %lu ages.\n",
page_cnt);
while (page_cnt-- > 0) {
if (pfn_valid(addr >> PAGE_SHIFT))
get_page(pfn_to_page(addr >> PAGE_SHIFT));
addr += PAGE_SIZE;
}
}
info.si_signo = SIGKILL;
info.si_errno = 0;
info.si_trapno = 0;
force_sig_info(info.si_signo, &info, current);
return true;
}
if (attrs & SUN4V_ERR_ATTRS_PIO) {
siginfo_t info;
info.si_signo = SIGBUS;
info.si_code = BUS_ADRERR;
info.si_addr = (void __user *)sun4v_get_vaddr(regs);
force_sig_info(info.si_signo, &info, current);
return true;
}
/* Default to doing nothing */
return false;
}
sparc64: Make %pil level 15 a pseudo-NMI. So that we can profile code even in a local_irq_disable() section, only write 14 (instead of 15) into the %pil register to disable IRQs. This allows PIL level 15 to serve as a pseudo NMI. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-11-24 13:55:29 +08:00
/* We run with %pil set to PIL_NORMAL_MAX and PSTATE_IE enabled in %pstate.
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
* Log the event, clear the first word of the entry, and die.
*/
void sun4v_nonresum_error(struct pt_regs *regs, unsigned long offset)
{
struct sun4v_error_entry *ent, local_copy;
struct trap_per_cpu *tb;
unsigned long paddr;
int cpu;
cpu = get_cpu();
tb = &trap_block[cpu];
paddr = tb->nonresum_kernel_buf_pa + offset;
ent = __va(paddr);
memcpy(&local_copy, ent, sizeof(struct sun4v_error_entry));
/* We have a local copy now, so release the entry. */
ent->err_handle = 0;
wmb();
put_cpu();
sparc64: Handle PIO & MEM non-resumable errors. User processes trying to access an invalid memory address via PIO will receive a SIGBUS signal instead of causing a panic. Memory errors will receive a SIGKILL since a SIGBUS may result in a coredump which may attempt to repeat the faulting access. Signed-off-by: Liam R. Howlett <Liam.Howlett@Oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-17 23:59:03 +08:00
if (!(regs->tstate & TSTATE_PRIV) &&
sun4v_nonresum_error_user_handled(regs, &local_copy)) {
/* DON'T PANIC: This userspace error was handled. */
return;
}
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
#ifdef CONFIG_PCI
/* Check for the special PCI poke sequence. */
if (pci_poke_in_progress && pci_poke_cpu == cpu) {
pci_poke_faulted = 1;
regs->tpc += 4;
regs->tnpc = regs->tpc + 4;
return;
}
#endif
[SPARC64]: Dump local cpu registers in sun4v_log_error() This makes the debugging information more usable. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-07 08:37:41 +08:00
sun4v_log_error(regs, &local_copy, cpu,
[SPARC64]: Sun4v interrupt handling. Sun4v has 4 interrupt queues: cpu, device, resumable errors, and non-resumable errors. A set of head/tail offset pointers help maintain a work queue in physical memory. The entries are 64-bytes in size. Each queue is allocated then registered with the hypervisor as we bring cpus up. The two error queues each get a kernel side buffer that we use to quickly empty the main interrupt queue before we call up to C code to log the event and possibly take evasive action. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-08 18:53:50 +08:00
KERN_EMERG "NON-RESUMABLE ERROR",
&sun4v_nonresum_oflow_cnt);
panic("Non-resumable error.");
}
/* If we try to printk() we'll probably make matters worse, by trying
* to retake locks this cpu already holds or causing more errors. So
* just bump a counter, and we'll report these counter bumps above.
*/
void sun4v_nonresum_overflow(struct pt_regs *regs)
{
/* XXX Actually even this can make not that much sense. Perhaps
* XXX we should just pull the plug and panic directly from here?
*/
atomic_inc(&sun4v_nonresum_oflow_cnt);
}
sparc64: sun4v TLB error power off events We've witnessed a few TLB events causing the machine to power off because of prom_halt. In one case it was some nfs related area during rmmod. Another was an mmapper of /dev/mem. A more recent one is an ITLB issue with a bad pagesize which could be a hardware bug. Bugs happen but we should attempt to not power off the machine and/or hang it when possible. This is a DTLB error from an mmapper of /dev/mem: [root@sparcie ~]# SUN4V-DTLB: Error at TPC[fffff80100903e6c], tl 1 SUN4V-DTLB: TPC<0xfffff80100903e6c> SUN4V-DTLB: O7[fffff801081979d0] SUN4V-DTLB: O7<0xfffff801081979d0> SUN4V-DTLB: vaddr[fffff80100000000] ctx[1250] pte[98000000000f0610] error[2] . This is recent mainline for ITLB: [ 3708.179864] SUN4V-ITLB: TPC<0xfffffc010071cefc> [ 3708.188866] SUN4V-ITLB: O7[fffffc010071cee8] [ 3708.197377] SUN4V-ITLB: O7<0xfffffc010071cee8> [ 3708.206539] SUN4V-ITLB: vaddr[e0003] ctx[1a3c] pte[2900000dcc800eeb] error[4] . Normally sun4v_itlb_error_report() and sun4v_dtlb_error_report() would call prom_halt() and drop us to OF command prompt "ok". This isn't the case for LDOMs and the machine powers off. For the HV reported error of HV_ENORADDR for HV HV_MMU_MAP_ADDR_TRAP we cause a SIGBUS error by qualifying it within do_sparc64_fault() for fault code mask of FAULT_CODE_BAD_RA. This is done when trap level (%tl) is less or equal one("1"). Otherwise, for %tl > 1, we proceed eventually to die_if_kernel(). The logic of this patch was partially inspired by David Miller's feedback. Power off of large sparc64 machines is painful. Plus die_if_kernel provides more context. A reset sequence isn't a brief period on large sparc64 but better than power-off/power-on sequence. Cc: sparclinux@vger.kernel.org Signed-off-by: Bob Picco <bob.picco@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-16 21:26:47 +08:00
static void sun4v_tlb_error(struct pt_regs *regs)
{
die_if_kernel("TLB/TSB error", regs);
}
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
unsigned long sun4v_err_itlb_vaddr;
unsigned long sun4v_err_itlb_ctx;
unsigned long sun4v_err_itlb_pte;
unsigned long sun4v_err_itlb_error;
void sun4v_itlb_error_report(struct pt_regs *regs, int tl)
{
sparc64: sun4v TLB error power off events We've witnessed a few TLB events causing the machine to power off because of prom_halt. In one case it was some nfs related area during rmmod. Another was an mmapper of /dev/mem. A more recent one is an ITLB issue with a bad pagesize which could be a hardware bug. Bugs happen but we should attempt to not power off the machine and/or hang it when possible. This is a DTLB error from an mmapper of /dev/mem: [root@sparcie ~]# SUN4V-DTLB: Error at TPC[fffff80100903e6c], tl 1 SUN4V-DTLB: TPC<0xfffff80100903e6c> SUN4V-DTLB: O7[fffff801081979d0] SUN4V-DTLB: O7<0xfffff801081979d0> SUN4V-DTLB: vaddr[fffff80100000000] ctx[1250] pte[98000000000f0610] error[2] . This is recent mainline for ITLB: [ 3708.179864] SUN4V-ITLB: TPC<0xfffffc010071cefc> [ 3708.188866] SUN4V-ITLB: O7[fffffc010071cee8] [ 3708.197377] SUN4V-ITLB: O7<0xfffffc010071cee8> [ 3708.206539] SUN4V-ITLB: vaddr[e0003] ctx[1a3c] pte[2900000dcc800eeb] error[4] . Normally sun4v_itlb_error_report() and sun4v_dtlb_error_report() would call prom_halt() and drop us to OF command prompt "ok". This isn't the case for LDOMs and the machine powers off. For the HV reported error of HV_ENORADDR for HV HV_MMU_MAP_ADDR_TRAP we cause a SIGBUS error by qualifying it within do_sparc64_fault() for fault code mask of FAULT_CODE_BAD_RA. This is done when trap level (%tl) is less or equal one("1"). Otherwise, for %tl > 1, we proceed eventually to die_if_kernel(). The logic of this patch was partially inspired by David Miller's feedback. Power off of large sparc64 machines is painful. Plus die_if_kernel provides more context. A reset sequence isn't a brief period on large sparc64 but better than power-off/power-on sequence. Cc: sparclinux@vger.kernel.org Signed-off-by: Bob Picco <bob.picco@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-16 21:26:47 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG "SUN4V-ITLB: Error at TPC[%lx], tl %d\n",
regs->tpc, tl);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_EMERG "SUN4V-ITLB: TPC<%pS>\n", (void *) regs->tpc);
[SPARC64]: Fix hypervisor TLB operation error reporting. 1) Trap level wasn't being passed down properly, we need to move it from %l4 into the correct outgoing arg register. 2) Although the TPC often provides the most direct clue, we have the caller PC so we should provide that as well. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-01-17 17:32:09 +08:00
printk(KERN_EMERG "SUN4V-ITLB: O7[%lx]\n", regs->u_regs[UREG_I7]);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_EMERG "SUN4V-ITLB: O7<%pS>\n",
(void *) regs->u_regs[UREG_I7]);
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG "SUN4V-ITLB: vaddr[%lx] ctx[%lx] "
"pte[%lx] error[%lx]\n",
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
sun4v_err_itlb_vaddr, sun4v_err_itlb_ctx,
sun4v_err_itlb_pte, sun4v_err_itlb_error);
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
sparc64: sun4v TLB error power off events We've witnessed a few TLB events causing the machine to power off because of prom_halt. In one case it was some nfs related area during rmmod. Another was an mmapper of /dev/mem. A more recent one is an ITLB issue with a bad pagesize which could be a hardware bug. Bugs happen but we should attempt to not power off the machine and/or hang it when possible. This is a DTLB error from an mmapper of /dev/mem: [root@sparcie ~]# SUN4V-DTLB: Error at TPC[fffff80100903e6c], tl 1 SUN4V-DTLB: TPC<0xfffff80100903e6c> SUN4V-DTLB: O7[fffff801081979d0] SUN4V-DTLB: O7<0xfffff801081979d0> SUN4V-DTLB: vaddr[fffff80100000000] ctx[1250] pte[98000000000f0610] error[2] . This is recent mainline for ITLB: [ 3708.179864] SUN4V-ITLB: TPC<0xfffffc010071cefc> [ 3708.188866] SUN4V-ITLB: O7[fffffc010071cee8] [ 3708.197377] SUN4V-ITLB: O7<0xfffffc010071cee8> [ 3708.206539] SUN4V-ITLB: vaddr[e0003] ctx[1a3c] pte[2900000dcc800eeb] error[4] . Normally sun4v_itlb_error_report() and sun4v_dtlb_error_report() would call prom_halt() and drop us to OF command prompt "ok". This isn't the case for LDOMs and the machine powers off. For the HV reported error of HV_ENORADDR for HV HV_MMU_MAP_ADDR_TRAP we cause a SIGBUS error by qualifying it within do_sparc64_fault() for fault code mask of FAULT_CODE_BAD_RA. This is done when trap level (%tl) is less or equal one("1"). Otherwise, for %tl > 1, we proceed eventually to die_if_kernel(). The logic of this patch was partially inspired by David Miller's feedback. Power off of large sparc64 machines is painful. Plus die_if_kernel provides more context. A reset sequence isn't a brief period on large sparc64 but better than power-off/power-on sequence. Cc: sparclinux@vger.kernel.org Signed-off-by: Bob Picco <bob.picco@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-16 21:26:47 +08:00
sun4v_tlb_error(regs);
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
}
unsigned long sun4v_err_dtlb_vaddr;
unsigned long sun4v_err_dtlb_ctx;
unsigned long sun4v_err_dtlb_pte;
unsigned long sun4v_err_dtlb_error;
void sun4v_dtlb_error_report(struct pt_regs *regs, int tl)
{
sparc64: sun4v TLB error power off events We've witnessed a few TLB events causing the machine to power off because of prom_halt. In one case it was some nfs related area during rmmod. Another was an mmapper of /dev/mem. A more recent one is an ITLB issue with a bad pagesize which could be a hardware bug. Bugs happen but we should attempt to not power off the machine and/or hang it when possible. This is a DTLB error from an mmapper of /dev/mem: [root@sparcie ~]# SUN4V-DTLB: Error at TPC[fffff80100903e6c], tl 1 SUN4V-DTLB: TPC<0xfffff80100903e6c> SUN4V-DTLB: O7[fffff801081979d0] SUN4V-DTLB: O7<0xfffff801081979d0> SUN4V-DTLB: vaddr[fffff80100000000] ctx[1250] pte[98000000000f0610] error[2] . This is recent mainline for ITLB: [ 3708.179864] SUN4V-ITLB: TPC<0xfffffc010071cefc> [ 3708.188866] SUN4V-ITLB: O7[fffffc010071cee8] [ 3708.197377] SUN4V-ITLB: O7<0xfffffc010071cee8> [ 3708.206539] SUN4V-ITLB: vaddr[e0003] ctx[1a3c] pte[2900000dcc800eeb] error[4] . Normally sun4v_itlb_error_report() and sun4v_dtlb_error_report() would call prom_halt() and drop us to OF command prompt "ok". This isn't the case for LDOMs and the machine powers off. For the HV reported error of HV_ENORADDR for HV HV_MMU_MAP_ADDR_TRAP we cause a SIGBUS error by qualifying it within do_sparc64_fault() for fault code mask of FAULT_CODE_BAD_RA. This is done when trap level (%tl) is less or equal one("1"). Otherwise, for %tl > 1, we proceed eventually to die_if_kernel(). The logic of this patch was partially inspired by David Miller's feedback. Power off of large sparc64 machines is painful. Plus die_if_kernel provides more context. A reset sequence isn't a brief period on large sparc64 but better than power-off/power-on sequence. Cc: sparclinux@vger.kernel.org Signed-off-by: Bob Picco <bob.picco@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-16 21:26:47 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG "SUN4V-DTLB: Error at TPC[%lx], tl %d\n",
regs->tpc, tl);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_EMERG "SUN4V-DTLB: TPC<%pS>\n", (void *) regs->tpc);
[SPARC64]: Fix hypervisor TLB operation error reporting. 1) Trap level wasn't being passed down properly, we need to move it from %l4 into the correct outgoing arg register. 2) Although the TPC often provides the most direct clue, we have the caller PC so we should provide that as well. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-01-17 17:32:09 +08:00
printk(KERN_EMERG "SUN4V-DTLB: O7[%lx]\n", regs->u_regs[UREG_I7]);
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(KERN_EMERG "SUN4V-DTLB: O7<%pS>\n",
(void *) regs->u_regs[UREG_I7]);
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
printk(KERN_EMERG "SUN4V-DTLB: vaddr[%lx] ctx[%lx] "
"pte[%lx] error[%lx]\n",
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
sun4v_err_dtlb_vaddr, sun4v_err_dtlb_ctx,
sun4v_err_dtlb_pte, sun4v_err_dtlb_error);
[SPARC64]: Use KERN_EMERG in dump_tl1_traplog() and sun4v TLB errors. We're about to seriously die in these cases so it is important that the messages make it to the console. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-19 09:06:28 +08:00
sparc64: sun4v TLB error power off events We've witnessed a few TLB events causing the machine to power off because of prom_halt. In one case it was some nfs related area during rmmod. Another was an mmapper of /dev/mem. A more recent one is an ITLB issue with a bad pagesize which could be a hardware bug. Bugs happen but we should attempt to not power off the machine and/or hang it when possible. This is a DTLB error from an mmapper of /dev/mem: [root@sparcie ~]# SUN4V-DTLB: Error at TPC[fffff80100903e6c], tl 1 SUN4V-DTLB: TPC<0xfffff80100903e6c> SUN4V-DTLB: O7[fffff801081979d0] SUN4V-DTLB: O7<0xfffff801081979d0> SUN4V-DTLB: vaddr[fffff80100000000] ctx[1250] pte[98000000000f0610] error[2] . This is recent mainline for ITLB: [ 3708.179864] SUN4V-ITLB: TPC<0xfffffc010071cefc> [ 3708.188866] SUN4V-ITLB: O7[fffffc010071cee8] [ 3708.197377] SUN4V-ITLB: O7<0xfffffc010071cee8> [ 3708.206539] SUN4V-ITLB: vaddr[e0003] ctx[1a3c] pte[2900000dcc800eeb] error[4] . Normally sun4v_itlb_error_report() and sun4v_dtlb_error_report() would call prom_halt() and drop us to OF command prompt "ok". This isn't the case for LDOMs and the machine powers off. For the HV reported error of HV_ENORADDR for HV HV_MMU_MAP_ADDR_TRAP we cause a SIGBUS error by qualifying it within do_sparc64_fault() for fault code mask of FAULT_CODE_BAD_RA. This is done when trap level (%tl) is less or equal one("1"). Otherwise, for %tl > 1, we proceed eventually to die_if_kernel(). The logic of this patch was partially inspired by David Miller's feedback. Power off of large sparc64 machines is painful. Plus die_if_kernel provides more context. A reset sequence isn't a brief period on large sparc64 but better than power-off/power-on sequence. Cc: sparclinux@vger.kernel.org Signed-off-by: Bob Picco <bob.picco@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-16 21:26:47 +08:00
sun4v_tlb_error(regs);
[SPARC64]: Fix some SUN4V TLB handling bugs. 1) Add error return checking for TLB load hypervisor calls. 2) Don't fallthru to dtlb tsb miss handler from itlb tsb miss handler, oops. 3) On window fixups, propagate fault information to fixup handler correctly. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-18 06:58:02 +08:00
}
[SPARC64]: Bulletproof hypervisor TLB flushing. Check TLB flush hypervisor calls for errors and report them. Pass HV_MMU_ALL always for now, we can add back the optimization to avoid the I-TLB flush later. Always explicitly page align the virtual address arguments. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 11:31:49 +08:00
void hypervisor_tlbop_error(unsigned long err, unsigned long op)
{
printk(KERN_CRIT "SUN4V: TLB hv call error %lu for op %lu\n",
err, op);
}
void hypervisor_tlbop_error_xcall(unsigned long err, unsigned long op)
{
printk(KERN_CRIT "SUN4V: XCALL TLB hv call error %lu for op %lu\n",
err, op);
}
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
static void do_fpe_common(struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (regs->tstate & TSTATE_PRIV) {
regs->tpc = regs->tnpc;
regs->tnpc += 4;
} else {
unsigned long fsr = current_thread_info()->xfsr[0];
siginfo_t info;
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGFPE;
info.si_errno = 0;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = 0;
signal/sparc: Document a conflict with SI_USER with SIGFPE Setting si_code to __SI_FAULT results in a userspace seeing an si_code of 0. This is the same si_code as SI_USER. Posix and common sense requires that SI_USER not be a signal specific si_code. As such this use of 0 for the si_code is a pretty horribly broken ABI. This was introduced in 2.3.41 so this mess has had a long time for people to be able to start depending on it. As this bug has existed for 17 years already I don't know if it is worth fixing. It is definitely worth documenting what is going on so that no one decides to copy this bad decision. Cc: "David S. Miller" <davem@davemloft.net> Cc: sparclinux@vger.kernel.org Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
2017-07-17 09:04:58 +08:00
info.si_code = FPE_FIXME;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if ((fsr & 0x1c000) == (1 << 14)) {
if (fsr & 0x10)
info.si_code = FPE_FLTINV;
else if (fsr & 0x08)
info.si_code = FPE_FLTOVF;
else if (fsr & 0x04)
info.si_code = FPE_FLTUND;
else if (fsr & 0x02)
info.si_code = FPE_FLTDIV;
else if (fsr & 0x01)
info.si_code = FPE_FLTRES;
}
force_sig_info(SIGFPE, &info, current);
}
}
void do_fpieee(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (notify_die(DIE_TRAP, "fpu exception ieee", regs,
0, 0x24, SIGFPE) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
do_fpe_common(regs);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void do_fpother(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
struct fpustate *f = FPUSTATE;
int ret = 0;
if (notify_die(DIE_TRAP, "fpu exception other", regs,
0, 0x25, SIGFPE) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
switch ((current_thread_info()->xfsr[0] & 0x1c000)) {
case (2 << 14): /* unfinished_FPop */
case (3 << 14): /* unimplemented_FPop */
sparc64: Fix several bugs in quad floating point emulation. UltraSPARC-T2 and later do not use the fp_exception_other trap and do not set the floating point trap type field in the %fsr at all when you try to execute an unimplemented FPU operation. Instead, it uses the illegal_instruction trap and it leaves the floating point trap type field clear. So we should not validate the %fsr trap type field when do_mathemu() is invoked from the illegal instruction handler. Also, the floating point trap type field is 3 bits, not 4 bits. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-05-25 15:31:56 +08:00
ret = do_mathemu(regs, f, false);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
break;
}
if (ret)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
do_fpe_common(regs);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void do_tof(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "tagged arithmetic overflow", regs,
0, 0x26, SIGEMT) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV)
die_if_kernel("Penguin overflow trap from kernel mode", regs);
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGEMT;
info.si_errno = 0;
info.si_code = EMT_TAGOVF;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = 0;
force_sig_info(SIGEMT, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void do_div0(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "integer division by zero", regs,
0, 0x28, SIGFPE) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV)
die_if_kernel("TL0: Kernel divide by zero.", regs);
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGFPE;
info.si_errno = 0;
info.si_code = FPE_INTDIV;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = 0;
force_sig_info(SIGFPE, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Fix sparse errors in arch/sparc64/kernel/traps.c Add 'UL' markers to DCU_* macros. Declare C functions called from assembler in entry.h Declare C functions called from within the sparc64 arch code in include/asm-sparc64/*.h headers as appropriate. Remove unused routines in traps.c Signed-off-by: David S. Miller <davem@davemloft.net>
2008-03-26 15:19:43 +08:00
static void instruction_dump(unsigned int *pc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
int i;
if ((((unsigned long) pc) & 3))
return;
printk("Instruction DUMP:");
for (i = -3; i < 6; i++)
printk("%c%08x%c",i?' ':'<',pc[i],i?' ':'>');
printk("\n");
}
[SPARC64]: Fix sparse errors in arch/sparc64/kernel/traps.c Add 'UL' markers to DCU_* macros. Declare C functions called from assembler in entry.h Declare C functions called from within the sparc64 arch code in include/asm-sparc64/*.h headers as appropriate. Remove unused routines in traps.c Signed-off-by: David S. Miller <davem@davemloft.net>
2008-03-26 15:19:43 +08:00
static void user_instruction_dump(unsigned int __user *pc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
int i;
unsigned int buf[9];
if ((((unsigned long) pc) & 3))
return;
if (copy_from_user(buf, pc - 3, sizeof(buf)))
return;
printk("Instruction DUMP:");
for (i = 0; i < 9; i++)
printk("%c%08x%c",i==3?' ':'<',buf[i],i==3?' ':'>');
printk("\n");
}
void show_stack(struct task_struct *tsk, unsigned long *_ksp)
{
sparc64: Fix build errors with gcc-4.6.0 Most of the warnings emitted (we fail arch/sparc file builds with -Werror) were legitimate but harmless, however one case (n2_pcr_write) was a genuine bug. Based almost entirely upon a patch by Sam Ravnborg. Reported-by: Dennis Gilmore <dennis@ausil.us> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-02-27 15:40:02 +08:00
unsigned long fp, ksp;
[SPARC64]: Fix show_stack() when stack argument is NULL. It didn't handle that case at all, and now dump_stack() can be implemented directly as show_stack(current, NULL) Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-30 15:17:12 +08:00
struct thread_info *tp;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
int count = 0;
sparc64: Fix stack dumping and tracing when function graph is enabled. Like x86, when the function graph tracer is enabled, emit the ftrace stub as well as the program counter it will be transformed back into. We duplicate a lot of similar stack walking logic in 3 or 4 spots, so eventually we should consolidate things like x86 does. Thanks to Frederic Weisbecker for pointing this out. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-21 18:08:11 +08:00
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
int graph = 0;
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ksp = (unsigned long) _ksp;
[SPARC64]: Fix show_stack() when stack argument is NULL. It didn't handle that case at all, and now dump_stack() can be implemented directly as show_stack(current, NULL) Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-30 15:17:12 +08:00
if (!tsk)
tsk = current;
tp = task_thread_info(tsk);
if (ksp == 0UL) {
if (tsk == current)
asm("mov %%fp, %0" : "=r" (ksp));
else
ksp = tp->ksp;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (tp == current_thread_info())
flushw_all();
fp = ksp + STACK_BIAS;
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk("Call Trace:\n");
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
do {
sparc64: Fix stack tracing through trap frames. The offset to the pt_regs area was wrong, so we weren't looking at the right location for the magic cookie. A trap frame is composed of a "struct sparc_stackf" then a "struct pt_regs", the code was using "struct reg_window" instead of "struct sparc_stackf". Signed-off-by: David S. Miller <davem@davemloft.net>
2008-05-22 09:15:53 +08:00
struct sparc_stackf *sf;
[SPARC64]: Detect trap frames in stack backtraces. Now that we have a magic cookie in the pt_regs, we can properly detect trap frames in stack bactraces. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 18:28:52 +08:00
struct pt_regs *regs;
unsigned long pc;
sparc64: Implement IRQ stacks. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-13 09:33:56 +08:00
if (!kstack_valid(tp, fp))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
break;
sparc64: Fix stack tracing through trap frames. The offset to the pt_regs area was wrong, so we weren't looking at the right location for the magic cookie. A trap frame is composed of a "struct sparc_stackf" then a "struct pt_regs", the code was using "struct reg_window" instead of "struct sparc_stackf". Signed-off-by: David S. Miller <davem@davemloft.net>
2008-05-22 09:15:53 +08:00
sf = (struct sparc_stackf *) fp;
regs = (struct pt_regs *) (sf + 1);
[SPARC64]: Detect trap frames in stack backtraces. Now that we have a magic cookie in the pt_regs, we can properly detect trap frames in stack bactraces. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 18:28:52 +08:00
sparc64: Implement IRQ stacks. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-13 09:33:56 +08:00
if (kstack_is_trap_frame(tp, regs)) {
sparc64: Fix stack tracing through trap frames. The offset to the pt_regs area was wrong, so we weren't looking at the right location for the magic cookie. A trap frame is composed of a "struct sparc_stackf" then a "struct pt_regs", the code was using "struct reg_window" instead of "struct sparc_stackf". Signed-off-by: David S. Miller <davem@davemloft.net>
2008-05-22 09:15:53 +08:00
if (!(regs->tstate & TSTATE_PRIV))
break;
[SPARC64]: Detect trap frames in stack backtraces. Now that we have a magic cookie in the pt_regs, we can properly detect trap frames in stack bactraces. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 18:28:52 +08:00
pc = regs->tpc;
fp = regs->u_regs[UREG_I6] + STACK_BIAS;
} else {
sparc64: Fix stack tracing through trap frames. The offset to the pt_regs area was wrong, so we weren't looking at the right location for the magic cookie. A trap frame is composed of a "struct sparc_stackf" then a "struct pt_regs", the code was using "struct reg_window" instead of "struct sparc_stackf". Signed-off-by: David S. Miller <davem@davemloft.net>
2008-05-22 09:15:53 +08:00
pc = sf->callers_pc;
fp = (unsigned long)sf->fp + STACK_BIAS;
[SPARC64]: Detect trap frames in stack backtraces. Now that we have a magic cookie in the pt_regs, we can properly detect trap frames in stack bactraces. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 18:28:52 +08:00
}
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk(" [%016lx] %pS\n", pc, (void *) pc);
sparc64: Fix stack dumping and tracing when function graph is enabled. Like x86, when the function graph tracer is enabled, emit the ftrace stub as well as the program counter it will be transformed back into. We duplicate a lot of similar stack walking logic in 3 or 4 spots, so eventually we should consolidate things like x86 does. Thanks to Frederic Weisbecker for pointing this out. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-21 18:08:11 +08:00
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
if ((pc + 8UL) == (unsigned long) &return_to_handler) {
int index = tsk->curr_ret_stack;
if (tsk->ret_stack && index >= graph) {
pc = tsk->ret_stack[index - graph].ret;
printk(" [%016lx] %pS\n", pc, (void *) pc);
graph++;
}
}
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
} while (++count < 16);
}
static inline struct reg_window *kernel_stack_up(struct reg_window *rw)
{
unsigned long fp = rw->ins[6];
if (!fp)
return NULL;
return (struct reg_window *) (fp + STACK_BIAS);
}
sparc64: fix sparse warning in traps_64.c Fix following warning: traps_64.c:2384:6: error: symbol 'die_if_kernel' redeclared with different type (originally declared at include/asm/bug.h:23) - different modifiers Add proper __noreturn to the implementation to fix this Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-17 05:25:52 +08:00
void __noreturn die_if_kernel(char *str, struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
static int die_counter;
int count = 0;
/* Amuse the user. */
printk(
" \\|/ ____ \\|/\n"
" \"@'/ .. \\`@\"\n"
" /_| \\__/ |_\\\n"
" \\__U_/\n");
Use helpers to obtain task pid in printks (arch code) One of the easiest things to isolate is the pid printed in kernel log. There was a patch, that made this for arch-independent code, this one makes so for arch/xxx files. It took some time to cross-compile it, but hopefully these are all the printks in arch code. Signed-off-by: Alexey Dobriyan <adobriyan@openvz.org> Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Cc: <linux-arch@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:41 +08:00
printk("%s(%d): %s [#%d]\n", current->comm, task_pid_nr(current), str, ++die_counter);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
notify_die(DIE_OOPS, str, regs, 0, 255, SIGSEGV);
__asm__ __volatile__("flushw");
sparc64: Kill __show_regs(). The story is that what we used to do when we actually used smp_report_regs() is that if you specifically only wanted to have the current cpu's registers dumped you would call "__show_regs()" otherwise you would call show_regs() which also invoked smp_report_regs(). Now that we killed off smp_report_regs() there is no longer any reason to have these two routines, just show_regs() is sufficient. Also kill off a stray declaration of show_regs() in sparc64_ksym.c Signed-off-by: David S. Miller <davem@davemloft.net>
2008-08-01 11:33:43 +08:00
show_regs(regs);
taint: add explicit flag to show whether lock dep is still OK. Fix up all callers as they were before, with make one change: an unsigned module taints the kernel, but doesn't turn off lockdep. Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2013-01-21 14:47:39 +08:00
add_taint(TAINT_DIE, LOCKDEP_NOW_UNRELIABLE);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV) {
sparc64: Use kstack_valid() in die_if_kernel(). This gets rid of a local function (is_kernel_stack()) which tries to do the same thing, yet poorly in that it doesn't handle IRQ stacks properly. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-13 13:16:22 +08:00
struct thread_info *tp = current_thread_info();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
struct reg_window *rw = (struct reg_window *)
(regs->u_regs[UREG_FP] + STACK_BIAS);
/* Stop the back trace when we hit userland or we
* find some badly aligned kernel stack.
*/
while (rw &&
sparc64: Use kstack_valid() in die_if_kernel(). This gets rid of a local function (is_kernel_stack()) which tries to do the same thing, yet poorly in that it doesn't handle IRQ stacks properly. Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-13 13:16:22 +08:00
count++ < 30 &&
kstack_valid(tp, (unsigned long) rw)) {
sparc: Use new '%pS' infrastructure to print symbols. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-18 13:11:32 +08:00
printk("Caller[%016lx]: %pS\n", rw->ins[7],
(void *) rw->ins[7]);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
rw = kernel_stack_up(rw);
}
instruction_dump ((unsigned int *) regs->tpc);
} else {
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
user_instruction_dump ((unsigned int __user *) regs->tpc);
}
sparc64: fatal trap should stop all cpus "echo c > /proc/sysrq-trigger" does not result in a system crash. There are two problems. One is that the trap handler ignores the global variable, panic_on_oops. The other is that smp_send_stop() is a no-op which leaves the other cpus running normally when one cpu panics. Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-07 08:31:39 +08:00
if (panic_on_oops)
panic("Fatal exception");
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV)
do_exit(SIGKILL);
do_exit(SIGSEGV);
}
sparc64: move EXPORT_SYMBOL to the symbols definition Move all applicable EXPORT_SYMBOL()s to the file where the respective symbol is defined. Removed all the includes that are no longer needed in sparc_ksyms_64.c Comment all remaining EXPORT_SYMBOL()s in sparc_ksyms_64.c Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Additions by Julian Calaby: * Moved EXPORT_SYMBOL()s for prom functions to their rightful places. * Made some minor cleanups to the includes and comments of sparc_ksyms_64.c * Updated and tidied commit message. * Rebased patch over sparc-2.6.git HEAD. * Ensured that all modified files have the correct includes. Signed-off-by: Julian Calaby <julian.calaby@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-01-09 08:58:20 +08:00
EXPORT_SYMBOL(die_if_kernel);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SPARC64]: Call do_mathemu on illegal instruction traps too. To add this logic, put the VIS instruction check at the vis_emul() call site instead of inside of vis_emul(). Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-20 06:38:25 +08:00
#define VIS_OPCODE_MASK ((0x3 << 30) | (0x3f << 19))
#define VIS_OPCODE_VAL ((0x2 << 30) | (0x36 << 19))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
void do_illegal_instruction(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned long pc = regs->tpc;
unsigned long tstate = regs->tstate;
u32 insn;
siginfo_t info;
if (notify_die(DIE_TRAP, "illegal instruction", regs,
0, 0x10, SIGILL) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (tstate & TSTATE_PRIV)
die_if_kernel("Kernel illegal instruction", regs);
if (test_thread_flag(TIF_32BIT))
pc = (u32)pc;
if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
if ((insn & 0xc1ffc000) == 0x81700000) /* POPC */ {
if (handle_popc(insn, regs))
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
} else if ((insn & 0xc1580000) == 0xc1100000) /* LDQ/STQ */ {
if (handle_ldf_stq(insn, regs))
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
[SPARC64]: First cut at VIS simulator for Niagara. Niagara does not implement some of the VIS instructions in hardware, so we have to emulate them. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-13 17:27:34 +08:00
} else if (tlb_type == hypervisor) {
[SPARC64]: Call do_mathemu on illegal instruction traps too. To add this logic, put the VIS instruction check at the vis_emul() call site instead of inside of vis_emul(). Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-20 06:38:25 +08:00
if ((insn & VIS_OPCODE_MASK) == VIS_OPCODE_VAL) {
if (!vis_emul(regs, insn))
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
[SPARC64]: Call do_mathemu on illegal instruction traps too. To add this logic, put the VIS instruction check at the vis_emul() call site instead of inside of vis_emul(). Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-20 06:38:25 +08:00
} else {
struct fpustate *f = FPUSTATE;
[SPARC64]: First cut at VIS simulator for Niagara. Niagara does not implement some of the VIS instructions in hardware, so we have to emulate them. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-13 17:27:34 +08:00
sparc64: Fix several bugs in quad floating point emulation. UltraSPARC-T2 and later do not use the fp_exception_other trap and do not set the floating point trap type field in the %fsr at all when you try to execute an unimplemented FPU operation. Instead, it uses the illegal_instruction trap and it leaves the floating point trap type field clear. So we should not validate the %fsr trap type field when do_mathemu() is invoked from the illegal instruction handler. Also, the floating point trap type field is 3 bits, not 4 bits. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-05-25 15:31:56 +08:00
/* On UltraSPARC T2 and later, FPU insns which
* are not implemented in HW signal an illegal
* instruction trap and do not set the FP Trap
* Trap in the %fsr to unimplemented_FPop.
[SPARC64]: Call do_mathemu on illegal instruction traps too. To add this logic, put the VIS instruction check at the vis_emul() call site instead of inside of vis_emul(). Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-20 06:38:25 +08:00
*/
sparc64: Fix several bugs in quad floating point emulation. UltraSPARC-T2 and later do not use the fp_exception_other trap and do not set the floating point trap type field in the %fsr at all when you try to execute an unimplemented FPU operation. Instead, it uses the illegal_instruction trap and it leaves the floating point trap type field clear. So we should not validate the %fsr trap type field when do_mathemu() is invoked from the illegal instruction handler. Also, the floating point trap type field is 3 bits, not 4 bits. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-05-25 15:31:56 +08:00
if (do_mathemu(regs, f, true))
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
[SPARC64]: Call do_mathemu on illegal instruction traps too. To add this logic, put the VIS instruction check at the vis_emul() call site instead of inside of vis_emul(). Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-20 06:38:25 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
info.si_signo = SIGILL;
info.si_errno = 0;
info.si_code = ILL_ILLOPC;
info.si_addr = (void __user *)pc;
info.si_trapno = 0;
force_sig_info(SIGILL, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void mem_address_unaligned(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "memory address unaligned", regs,
0, 0x34, SIGSEGV) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (regs->tstate & TSTATE_PRIV) {
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
kernel_unaligned_trap(regs, *((unsigned int *)regs->tpc));
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
sparc64: Handle additional cases of no fault loads Load instructions using ASI_PNF or other no-fault ASIs should not cause a SIGSEGV or SIGBUS. A garden variety unmapped address follows the TSB miss path, and when no valid mapping is found in the process page tables, the miss handler checks to see if the access was via a no-fault ASI. It then fixes up the target register with a zero, and skips the no-fault load instruction. But different paths are taken for data access exceptions and alignment traps, and these do not respect the no-fault ASI. We add checks in these paths for the no-fault ASI, and fix up the target register and TPC just like in the TSB miss case. Signed-off-by: Rob Gardner <rob.gardner@oracle.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-09-09 06:34:21 +08:00
if (is_no_fault_exception(regs))
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
info.si_signo = SIGBUS;
info.si_errno = 0;
info.si_code = BUS_ADRALN;
info.si_addr = (void __user *)sfar;
info.si_trapno = 0;
force_sig_info(SIGBUS, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[SPARC64]: Fix C-function name called by sun4v_mna trap code. The trap code was calling itself :-) Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-15 08:39:22 +08:00
void sun4v_do_mna(struct pt_regs *regs, unsigned long addr, unsigned long type_ctx)
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
{
siginfo_t info;
if (notify_die(DIE_TRAP, "memory address unaligned", regs,
0, 0x34, SIGSEGV) == NOTIFY_STOP)
return;
if (regs->tstate & TSTATE_PRIV) {
kernel_unaligned_trap(regs, *((unsigned int *)regs->tpc));
return;
}
sparc64: Handle additional cases of no fault loads Load instructions using ASI_PNF or other no-fault ASIs should not cause a SIGSEGV or SIGBUS. A garden variety unmapped address follows the TSB miss path, and when no valid mapping is found in the process page tables, the miss handler checks to see if the access was via a no-fault ASI. It then fixes up the target register with a zero, and skips the no-fault load instruction. But different paths are taken for data access exceptions and alignment traps, and these do not respect the no-fault ASI. We add checks in these paths for the no-fault ASI, and fix up the target register and TPC just like in the TSB miss case. Signed-off-by: Rob Gardner <rob.gardner@oracle.com> Acked-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-09-09 06:34:21 +08:00
if (is_no_fault_exception(regs))
return;
[SPARC64]: SUN4V memory exception trap handlers. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-10 12:20:34 +08:00
info.si_signo = SIGBUS;
info.si_errno = 0;
info.si_code = BUS_ADRALN;
info.si_addr = (void __user *) addr;
info.si_trapno = 0;
force_sig_info(SIGBUS, &info, current);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
void do_privop(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
enum ctx_state prev_state = exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
siginfo_t info;
if (notify_die(DIE_TRAP, "privileged operation", regs,
0, 0x11, SIGILL) == NOTIFY_STOP)
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
info.si_signo = SIGILL;
info.si_errno = 0;
info.si_code = ILL_PRVOPC;
info.si_addr = (void __user *)regs->tpc;
info.si_trapno = 0;
force_sig_info(SIGILL, &info, current);
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
out:
exception_exit(prev_state);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void do_privact(struct pt_regs *regs)
{
do_privop(regs);
}
/* Trap level 1 stuff or other traps we should never see... */
void do_cee(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
die_if_kernel("TL0: Cache Error Exception", regs);
}
void do_div0_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: DIV0 Exception", regs);
}
void do_fpieee_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: FPU IEEE Exception", regs);
}
void do_fpother_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: FPU Other Exception", regs);
}
void do_ill_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: Illegal Instruction Exception", regs);
}
void do_irq_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: IRQ Exception", regs);
}
void do_lddfmna_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: LDDF Exception", regs);
}
void do_stdfmna_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: STDF Exception", regs);
}
void do_paw(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
die_if_kernel("TL0: Phys Watchpoint Exception", regs);
}
void do_paw_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: Phys Watchpoint Exception", regs);
}
void do_vaw(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
die_if_kernel("TL0: Virt Watchpoint Exception", regs);
}
void do_vaw_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: Virt Watchpoint Exception", regs);
}
void do_tof_tl1(struct pt_regs *regs)
{
sparc64: Implement HAVE_CONTEXT_TRACKING Mark the places when the system are in user or are in kernel. This is used to make full dynticks system (tickless) -- CONFIG_NO_HZ_FULL dependence. Signed-off-by: Kirill Tkhai <tkhai@yandex.ru> CC: David Miller <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-09-14 20:02:11 +08:00
exception_enter();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dump_tl1_traplog((struct tl1_traplog *)(regs + 1));
die_if_kernel("TL1: Tag Overflow Exception", regs);
}
void do_getpsr(struct pt_regs *regs)
{
regs->u_regs[UREG_I0] = tstate_to_psr(regs->tstate);
regs->tpc = regs->tnpc;
regs->tnpc += 4;
if (test_thread_flag(TIF_32BIT)) {
regs->tpc &= 0xffffffff;
regs->tnpc &= 0xffffffff;
}
}
sparc64: Measure receiver forward progress to avoid send mondo timeout A large sun4v SPARC system may have moments of intensive xcall activities, usually caused by unmapping many pages on many CPUs concurrently. This can flood receivers with CPU mondo interrupts for an extended period, causing some unlucky senders to hit send-mondo timeout. This problem gets worse as cpu count increases because sometimes mappings must be invalidated on all CPUs, and sometimes all CPUs may gang up on a single CPU. But a busy system is not a broken system. In the above scenario, as long as the receiver is making forward progress processing mondo interrupts, the sender should continue to retry. This patch implements the receiver's forward progress meter by introducing a per cpu counter 'cpu_mondo_counter[cpu]' where 'cpu' is in the range of 0..NR_CPUS. The receiver increments its counter as soon as it receives a mondo and the sender tracks the receiver's counter. If the receiver has stopped making forward progress when the retry limit is reached, the sender declares send-mondo-timeout and panic; otherwise, the receiver is allowed to keep making forward progress. In addition, it's been observed that PCIe hotplug events generate Correctable Errors that are handled by hypervisor and then OS. Hypervisor 'borrows' a guest cpu strand briefly to provide the service. If the cpu strand is simultaneously the only cpu targeted by a mondo, it may not be available for the mondo in 20msec, causing SUN4V mondo timeout. It appears that 1 second is the agreed wait time between hypervisor and guest OS, this patch makes the adjustment. Orabug: 25476541 Orabug: 26417466 Signed-off-by: Jane Chu <jane.chu@oracle.com> Reviewed-by: Steve Sistare <steven.sistare@oracle.com> Reviewed-by: Anthony Yznaga <anthony.yznaga@oracle.com> Reviewed-by: Rob Gardner <rob.gardner@oracle.com> Reviewed-by: Thomas Tai <thomas.tai@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-12 02:00:54 +08:00
u64 cpu_mondo_counter[NR_CPUS] = {0};
[SPARC64]: Elminate all usage of hard-coded trap globals. UltraSPARC has special sets of global registers which are switched to for certain trap types. There is one set for MMU related traps, one set of Interrupt Vector processing, and another set (called the Alternate globals) for all other trap types. For what seems like forever we've hard coded the values in some of these trap registers. Some examples include: 1) Interrupt Vector global %g6 holds current processors interrupt work struct where received interrupts are managed for IRQ handler dispatch. 2) MMU global %g7 holds the base of the page tables of the currently active address space. 3) Alternate global %g6 held the current_thread_info() value. Such hardcoding has resulted in some serious issues in many areas. There are some code sequences where having another register available would help clean up the implementation. Taking traps such as cross-calls from the OBP firmware requires some trick code sequences wherein we have to save away and restore all of the special sets of global registers when we enter/exit OBP. We were also using the IMMU TSB register on SMP to hold the per-cpu area base address, which doesn't work any longer now that we actually use the TSB facility of the cpu. The implementation is pretty straight forward. One tricky bit is getting the current processor ID as that is different on different cpu variants. We use a stub with a fancy calling convention which we patch at boot time. The calling convention is that the stub is branched to and the (PC - 4) to return to is in register %g1. The cpu number is left in %g6. This stub can be invoked by using the __GET_CPUID macro. We use an array of per-cpu trap state to store the current thread and physical address of the current address space's page tables. The TRAP_LOAD_THREAD_REG loads %g6 with the current thread from this table, it uses __GET_CPUID and also clobbers %g1. TRAP_LOAD_IRQ_WORK is used by the interrupt vector processing to load the current processor's IRQ software state into %g6. It also uses __GET_CPUID and clobbers %g1. Finally, TRAP_LOAD_PGD_PHYS loads the physical address base of the current address space's page tables into %g7, it clobbers %g1 and uses __GET_CPUID. Many refinements are possible, as well as some tuning, with this stuff in place. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:24:22 +08:00
struct trap_per_cpu trap_block[NR_CPUS];
sparc64: Store per-cpu offset in trap_block[] Surprisingly this actually makes LOAD_PER_CPU_BASE() a little more efficient. Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-01 16:47:10 +08:00
EXPORT_SYMBOL(trap_block);
[SPARC64]: Elminate all usage of hard-coded trap globals. UltraSPARC has special sets of global registers which are switched to for certain trap types. There is one set for MMU related traps, one set of Interrupt Vector processing, and another set (called the Alternate globals) for all other trap types. For what seems like forever we've hard coded the values in some of these trap registers. Some examples include: 1) Interrupt Vector global %g6 holds current processors interrupt work struct where received interrupts are managed for IRQ handler dispatch. 2) MMU global %g7 holds the base of the page tables of the currently active address space. 3) Alternate global %g6 held the current_thread_info() value. Such hardcoding has resulted in some serious issues in many areas. There are some code sequences where having another register available would help clean up the implementation. Taking traps such as cross-calls from the OBP firmware requires some trick code sequences wherein we have to save away and restore all of the special sets of global registers when we enter/exit OBP. We were also using the IMMU TSB register on SMP to hold the per-cpu area base address, which doesn't work any longer now that we actually use the TSB facility of the cpu. The implementation is pretty straight forward. One tricky bit is getting the current processor ID as that is different on different cpu variants. We use a stub with a fancy calling convention which we patch at boot time. The calling convention is that the stub is branched to and the (PC - 4) to return to is in register %g1. The cpu number is left in %g6. This stub can be invoked by using the __GET_CPUID macro. We use an array of per-cpu trap state to store the current thread and physical address of the current address space's page tables. The TRAP_LOAD_THREAD_REG loads %g6 with the current thread from this table, it uses __GET_CPUID and also clobbers %g1. TRAP_LOAD_IRQ_WORK is used by the interrupt vector processing to load the current processor's IRQ software state into %g6. It also uses __GET_CPUID and clobbers %g1. Finally, TRAP_LOAD_PGD_PHYS loads the physical address base of the current address space's page tables into %g7, it clobbers %g1 and uses __GET_CPUID. Many refinements are possible, as well as some tuning, with this stuff in place. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:24:22 +08:00
/* This can get invoked before sched_init() so play it super safe
* and use hard_smp_processor_id().
*/
sparc64: Fix SMP bootup with CONFIG_STACK_DEBUG or ftrace. Based upon a report by Meelis Roos. Any function call can try to access the current thread register via the _mcount hooks when the kernel is built with -pg (via ftrace or STACK_DEBUG). That can't be setup properly very early on during the bootup of other cpus for sun4u and some early sun4v systems. So add notrace markers to these specific functions, so that _mcount doesn't get invoked too early. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-09-17 02:44:00 +08:00
void notrace init_cur_cpu_trap(struct thread_info *t)
[SPARC64]: Elminate all usage of hard-coded trap globals. UltraSPARC has special sets of global registers which are switched to for certain trap types. There is one set for MMU related traps, one set of Interrupt Vector processing, and another set (called the Alternate globals) for all other trap types. For what seems like forever we've hard coded the values in some of these trap registers. Some examples include: 1) Interrupt Vector global %g6 holds current processors interrupt work struct where received interrupts are managed for IRQ handler dispatch. 2) MMU global %g7 holds the base of the page tables of the currently active address space. 3) Alternate global %g6 held the current_thread_info() value. Such hardcoding has resulted in some serious issues in many areas. There are some code sequences where having another register available would help clean up the implementation. Taking traps such as cross-calls from the OBP firmware requires some trick code sequences wherein we have to save away and restore all of the special sets of global registers when we enter/exit OBP. We were also using the IMMU TSB register on SMP to hold the per-cpu area base address, which doesn't work any longer now that we actually use the TSB facility of the cpu. The implementation is pretty straight forward. One tricky bit is getting the current processor ID as that is different on different cpu variants. We use a stub with a fancy calling convention which we patch at boot time. The calling convention is that the stub is branched to and the (PC - 4) to return to is in register %g1. The cpu number is left in %g6. This stub can be invoked by using the __GET_CPUID macro. We use an array of per-cpu trap state to store the current thread and physical address of the current address space's page tables. The TRAP_LOAD_THREAD_REG loads %g6 with the current thread from this table, it uses __GET_CPUID and also clobbers %g1. TRAP_LOAD_IRQ_WORK is used by the interrupt vector processing to load the current processor's IRQ software state into %g6. It also uses __GET_CPUID and clobbers %g1. Finally, TRAP_LOAD_PGD_PHYS loads the physical address base of the current address space's page tables into %g7, it clobbers %g1 and uses __GET_CPUID. Many refinements are possible, as well as some tuning, with this stuff in place. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:24:22 +08:00
{
int cpu = hard_smp_processor_id();
struct trap_per_cpu *p = &trap_block[cpu];
[SPARC64]: Get SUN4V SMP working. The sibling cpu bringup is extremely fragile. We can only perform the most basic calls until we take over the trap table from the firmware/hypervisor on the new cpu. This means no accesses to %g4, %g5, %g6 since those can't be TLB translated without our trap handlers. In order to achieve this: 1) Change sun4v_init_mondo_queues() so that it can operate in several modes. It can allocate the queues, or install them in the current processor, or both. The boot cpu does both in it's call early on. Later, the boot cpu allocates the sibling cpu queue, starts the sibling cpu, then the sibling cpu loads them in. 2) init_cur_cpu_trap() is changed to take the current_thread_info() as an argument instead of reading %g6 directly on the current cpu. 3) Create a trampoline stack for the sibling cpus. We do our basic kernel calls using this stack, which is locked into the kernel image, then go to our proper thread stack after taking over the trap table. 4) While we are in this delicate startup state, we put 0xdeadbeef into %g4/%g5/%g6 in order to catch accidental accesses. 5) On the final prom_set_trap_table*() call, we put &init_thread_union into %g6. This is a hack to make prom_world(0) work. All that wants to do is restore the %asi register using get_thread_current_ds(). Longer term we should just do the OBP calls to set the trap table by hand just like we do for everything else. This would avoid that silly prom_world(0) issue, then we can remove the init_thread_union hack. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-17 17:29:17 +08:00
p->thread = t;
[SPARC64]: Elminate all usage of hard-coded trap globals. UltraSPARC has special sets of global registers which are switched to for certain trap types. There is one set for MMU related traps, one set of Interrupt Vector processing, and another set (called the Alternate globals) for all other trap types. For what seems like forever we've hard coded the values in some of these trap registers. Some examples include: 1) Interrupt Vector global %g6 holds current processors interrupt work struct where received interrupts are managed for IRQ handler dispatch. 2) MMU global %g7 holds the base of the page tables of the currently active address space. 3) Alternate global %g6 held the current_thread_info() value. Such hardcoding has resulted in some serious issues in many areas. There are some code sequences where having another register available would help clean up the implementation. Taking traps such as cross-calls from the OBP firmware requires some trick code sequences wherein we have to save away and restore all of the special sets of global registers when we enter/exit OBP. We were also using the IMMU TSB register on SMP to hold the per-cpu area base address, which doesn't work any longer now that we actually use the TSB facility of the cpu. The implementation is pretty straight forward. One tricky bit is getting the current processor ID as that is different on different cpu variants. We use a stub with a fancy calling convention which we patch at boot time. The calling convention is that the stub is branched to and the (PC - 4) to return to is in register %g1. The cpu number is left in %g6. This stub can be invoked by using the __GET_CPUID macro. We use an array of per-cpu trap state to store the current thread and physical address of the current address space's page tables. The TRAP_LOAD_THREAD_REG loads %g6 with the current thread from this table, it uses __GET_CPUID and also clobbers %g1. TRAP_LOAD_IRQ_WORK is used by the interrupt vector processing to load the current processor's IRQ software state into %g6. It also uses __GET_CPUID and clobbers %g1. Finally, TRAP_LOAD_PGD_PHYS loads the physical address base of the current address space's page tables into %g7, it clobbers %g1 and uses __GET_CPUID. Many refinements are possible, as well as some tuning, with this stuff in place. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:24:22 +08:00
p->pgd_paddr = 0;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
extern void thread_info_offsets_are_bolixed_dave(void);
[SPARC64]: Elminate all usage of hard-coded trap globals. UltraSPARC has special sets of global registers which are switched to for certain trap types. There is one set for MMU related traps, one set of Interrupt Vector processing, and another set (called the Alternate globals) for all other trap types. For what seems like forever we've hard coded the values in some of these trap registers. Some examples include: 1) Interrupt Vector global %g6 holds current processors interrupt work struct where received interrupts are managed for IRQ handler dispatch. 2) MMU global %g7 holds the base of the page tables of the currently active address space. 3) Alternate global %g6 held the current_thread_info() value. Such hardcoding has resulted in some serious issues in many areas. There are some code sequences where having another register available would help clean up the implementation. Taking traps such as cross-calls from the OBP firmware requires some trick code sequences wherein we have to save away and restore all of the special sets of global registers when we enter/exit OBP. We were also using the IMMU TSB register on SMP to hold the per-cpu area base address, which doesn't work any longer now that we actually use the TSB facility of the cpu. The implementation is pretty straight forward. One tricky bit is getting the current processor ID as that is different on different cpu variants. We use a stub with a fancy calling convention which we patch at boot time. The calling convention is that the stub is branched to and the (PC - 4) to return to is in register %g1. The cpu number is left in %g6. This stub can be invoked by using the __GET_CPUID macro. We use an array of per-cpu trap state to store the current thread and physical address of the current address space's page tables. The TRAP_LOAD_THREAD_REG loads %g6 with the current thread from this table, it uses __GET_CPUID and also clobbers %g1. TRAP_LOAD_IRQ_WORK is used by the interrupt vector processing to load the current processor's IRQ software state into %g6. It also uses __GET_CPUID and clobbers %g1. Finally, TRAP_LOAD_PGD_PHYS loads the physical address base of the current address space's page tables into %g7, it clobbers %g1 and uses __GET_CPUID. Many refinements are possible, as well as some tuning, with this stuff in place. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-27 15:24:22 +08:00
extern void trap_per_cpu_offsets_are_bolixed_dave(void);
[SPARC64]: Add a secondary TSB for hugepage mappings. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-22 16:49:59 +08:00
extern void tsb_config_offsets_are_bolixed_dave(void);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Only invoked on boot processor. */
void __init trap_init(void)
{
/* Compile time sanity check. */
sparc64: Use BUILD_BUG_ON() in trap_init(). Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-01 16:57:36 +08:00
BUILD_BUG_ON(TI_TASK != offsetof(struct thread_info, task) ||
TI_FLAGS != offsetof(struct thread_info, flags) ||
TI_CPU != offsetof(struct thread_info, cpu) ||
TI_FPSAVED != offsetof(struct thread_info, fpsaved) ||
TI_KSP != offsetof(struct thread_info, ksp) ||
TI_FAULT_ADDR != offsetof(struct thread_info,
fault_address) ||
TI_KREGS != offsetof(struct thread_info, kregs) ||
TI_UTRAPS != offsetof(struct thread_info, utraps) ||
TI_REG_WINDOW != offsetof(struct thread_info,
reg_window) ||
TI_RWIN_SPTRS != offsetof(struct thread_info,
rwbuf_stkptrs) ||
TI_GSR != offsetof(struct thread_info, gsr) ||
TI_XFSR != offsetof(struct thread_info, xfsr) ||
TI_PRE_COUNT != offsetof(struct thread_info,
preempt_count) ||
TI_NEW_CHILD != offsetof(struct thread_info, new_child) ||
sparc64: clear syscall_noerror on the entry to syscall, not on the exit Move that sucker to just before TI_FPDEPTH and replace stb with sth in etrap_save(). Take current_ds to its old place, so that we don't push wsaved into TI_... flags. That allows to lose clearing syscall_noerror on return from syscall. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2012-09-26 13:21:14 +08:00
TI_CURRENT_DS != offsetof(struct thread_info,
current_ds) ||
sparc64: Use BUILD_BUG_ON() in trap_init(). Signed-off-by: David S. Miller <davem@davemloft.net>
2009-04-01 16:57:36 +08:00
TI_KUNA_REGS != offsetof(struct thread_info,
kern_una_regs) ||
TI_KUNA_INSN != offsetof(struct thread_info,
kern_una_insn) ||
TI_FPREGS != offsetof(struct thread_info, fpregs) ||
(TI_FPREGS & (64 - 1)));
BUILD_BUG_ON(TRAP_PER_CPU_THREAD != offsetof(struct trap_per_cpu,
thread) ||
(TRAP_PER_CPU_PGD_PADDR !=
offsetof(struct trap_per_cpu, pgd_paddr)) ||
(TRAP_PER_CPU_CPU_MONDO_PA !=
offsetof(struct trap_per_cpu, cpu_mondo_pa)) ||
(TRAP_PER_CPU_DEV_MONDO_PA !=
offsetof(struct trap_per_cpu, dev_mondo_pa)) ||
(TRAP_PER_CPU_RESUM_MONDO_PA !=
offsetof(struct trap_per_cpu, resum_mondo_pa)) ||
(TRAP_PER_CPU_RESUM_KBUF_PA !=
offsetof(struct trap_per_cpu, resum_kernel_buf_pa)) ||
(TRAP_PER_CPU_NONRESUM_MONDO_PA !=
offsetof(struct trap_per_cpu, nonresum_mondo_pa)) ||
(TRAP_PER_CPU_NONRESUM_KBUF_PA !=
offsetof(struct trap_per_cpu, nonresum_kernel_buf_pa)) ||
(TRAP_PER_CPU_FAULT_INFO !=
offsetof(struct trap_per_cpu, fault_info)) ||
(TRAP_PER_CPU_CPU_MONDO_BLOCK_PA !=
offsetof(struct trap_per_cpu, cpu_mondo_block_pa)) ||
(TRAP_PER_CPU_CPU_LIST_PA !=
offsetof(struct trap_per_cpu, cpu_list_pa)) ||
(TRAP_PER_CPU_TSB_HUGE !=
offsetof(struct trap_per_cpu, tsb_huge)) ||
(TRAP_PER_CPU_TSB_HUGE_TEMP !=
offsetof(struct trap_per_cpu, tsb_huge_temp)) ||
(TRAP_PER_CPU_IRQ_WORKLIST_PA !=
offsetof(struct trap_per_cpu, irq_worklist_pa)) ||
(TRAP_PER_CPU_CPU_MONDO_QMASK !=
offsetof(struct trap_per_cpu, cpu_mondo_qmask)) ||
(TRAP_PER_CPU_DEV_MONDO_QMASK !=
offsetof(struct trap_per_cpu, dev_mondo_qmask)) ||
(TRAP_PER_CPU_RESUM_QMASK !=
offsetof(struct trap_per_cpu, resum_qmask)) ||
(TRAP_PER_CPU_NONRESUM_QMASK !=
offsetof(struct trap_per_cpu, nonresum_qmask)) ||
(TRAP_PER_CPU_PER_CPU_BASE !=
offsetof(struct trap_per_cpu, __per_cpu_base)));
BUILD_BUG_ON((TSB_CONFIG_TSB !=
offsetof(struct tsb_config, tsb)) ||
(TSB_CONFIG_RSS_LIMIT !=
offsetof(struct tsb_config, tsb_rss_limit)) ||
(TSB_CONFIG_NENTRIES !=
offsetof(struct tsb_config, tsb_nentries)) ||
(TSB_CONFIG_REG_VAL !=
offsetof(struct tsb_config, tsb_reg_val)) ||
(TSB_CONFIG_MAP_VADDR !=
offsetof(struct tsb_config, tsb_map_vaddr)) ||
(TSB_CONFIG_MAP_PTE !=
offsetof(struct tsb_config, tsb_map_pte)));
[SPARC64]: Add a secondary TSB for hugepage mappings. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-22 16:49:59 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Attach to the address space of init_task. On SMP we
* do this in smp.c:smp_callin for other cpus.
*/
mm: add new mmgrab() helper Apart from adding the helper function itself, the rest of the kernel is converted mechanically using: git grep -l 'atomic_inc.*mm_count' | xargs sed -i 's/atomic_inc(&\(.*\)->mm_count);/mmgrab\(\1\);/' git grep -l 'atomic_inc.*mm_count' | xargs sed -i 's/atomic_inc(&\(.*\)\.mm_count);/mmgrab\(\&\1\);/' This is needed for a later patch that hooks into the helper, but might be a worthwhile cleanup on its own. (Michal Hocko provided most of the kerneldoc comment.) Link: http://lkml.kernel.org/r/20161218123229.22952-1-vegard.nossum@oracle.com Signed-off-by: Vegard Nossum <vegard.nossum@oracle.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-28 06:30:07 +08:00
mmgrab(&init_mm);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
current->active_mm = &init_mm;
}