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

497 lines
13 KiB
C
Raw Normal View History

powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
// SPDX-License-Identifier: GPL-2.0-or-later
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
#include <linux/context_tracking.h>
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#include <linux/err.h>
powerpc/syscall: Use is_compat_task() Instead of hard comparing task flags with _TIF_32BIT, use is_compat_task(). The advantage is that it returns 0 on PPC32 allthough _TIF_32BIT is always set. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/c8094662199337a7200fea9f6e1d1f8b1b6d5f69.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:29 +08:00
#include <linux/compat.h>
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#include <asm/asm-prototypes.h>
powerpc: Only include kup-radix.h for 64-bit Book3S In kup.h we currently include kup-radix.h for all 64-bit builds, which includes Book3S and Book3E. The latter doesn't make sense, Book3E never uses the Radix MMU. This has worked up until now, but almost by accident, and the recent uaccess flush changes introduced a build breakage on Book3E because of the bad structure of the code. So disentangle things so that we only use kup-radix.h for Book3S. This requires some more stubs in kup.h and fixing an include in syscall_64.c. Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2020-11-19 20:43:53 +08:00
#include <asm/kup.h>
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#include <asm/cputime.h>
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
#include <asm/interrupt.h>
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#include <asm/hw_irq.h>
powerpc: convert interrupt handlers to use wrappers Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-29-npiggin@gmail.com
2021-01-30 21:08:38 +08:00
#include <asm/interrupt.h>
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#include <asm/kprobes.h>
#include <asm/paca.h>
#include <asm/ptrace.h>
#include <asm/reg.h>
#include <asm/signal.h>
#include <asm/switch_to.h>
#include <asm/syscall.h>
#include <asm/time.h>
#include <asm/unistd.h>
powerpc/32: Replace ASM exception exit by C exception exit from ppc64 This patch replaces the PPC32 ASM exception exit by C exception exit. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/48f8bae91da899d8e73fc0d75c9af66cc97b4d5b.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:34 +08:00
#if defined(CONFIG_PPC_ADV_DEBUG_REGS) && defined(CONFIG_PPC32)
unsigned long global_dbcr0[NR_CPUS];
#endif
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
typedef long (*syscall_fn)(long, long, long, long, long, long);
powerpc/64/syscall: Reconcile interrupts This reconciles interrupts in the system call case like all other interrupts. This allows system_call_common to be shared with the scv system call implementation in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-31-npiggin@gmail.com
2020-02-26 01:35:39 +08:00
/* Has to run notrace because it is entered not completely "reconciled" */
notrace long system_call_exception(long r3, long r4, long r5,
long r6, long r7, long r8,
unsigned long r0, struct pt_regs *regs)
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
{
syscall_fn f;
powerpc/32s: Move KUEP locking/unlocking in C This can be done in C, do it. Unrolling the loop gains approx. 15% performance. From now on, prepare_transfer_to_handler() is only for interrupts from kernel. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/4eadd873927e9a73c3d1dfe2f9497353465514cf.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:47 +08:00
kuep_lock();
powerpc/syscall: Save r3 in regs->orig_r3 Save r3 in regs->orig_r3 in system_call_exception() Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/9a90805ab6b9101b46daf56470f457a57acd86fc.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:30 +08:00
regs->orig_gpr3 = r3;
powerpc/64/syscall: Reconcile interrupts This reconciles interrupts in the system call case like all other interrupts. This allows system_call_common to be shared with the scv system call implementation in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-31-npiggin@gmail.com
2020-02-26 01:35:39 +08:00
if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
BUG_ON(irq_soft_mask_return() != IRQS_ALL_DISABLED);
powerpc/syscall: switch user_exit_irqoff and trace_hardirqs_off order user_exit_irqoff() -> __context_tracking_exit -> vtime_user_exit warns in __seqprop_assert due to lockdep thinking preemption is enabled because trace_hardirqs_off() has not yet been called. Switch the order of these two calls, which matches their ordering in interrupt_enter_prepare. Fixes: 5f0b6ac3905f ("powerpc/64/syscall: Reconcile interrupts") Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210316104206.407354-2-npiggin@gmail.com
2021-03-16 18:41:55 +08:00
trace_hardirqs_off(); /* finish reconciling */
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
CT_WARN_ON(ct_state() == CONTEXT_KERNEL);
user_exit_irqoff();
powerpc/syscall: Change condition to check MSR_RI In system_call_exception(), MSR_RI also needs to be checked on 8xx. Only booke and 40x doesn't have MSR_RI. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/67820fada8dd6a8fe9d7b666f175d4cc9d8de87e.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:31 +08:00
if (!IS_ENABLED(CONFIG_BOOKE) && !IS_ENABLED(CONFIG_40x))
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
BUG_ON(!(regs->msr & MSR_RI));
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
BUG_ON(!(regs->msr & MSR_PR));
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
BUG_ON(arch_irq_disabled_regs(regs));
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
#ifdef CONFIG_PPC_PKEY
if (mmu_has_feature(MMU_FTR_PKEY)) {
unsigned long amr, iamr;
powerpc/book3s64/pkeys: Optimize KUAP and KUEP feature disabled case If FTR_BOOK3S_KUAP is disabled, kernel will continue to run with the same AMR value with which it was entered. Hence there is a high chance that we can return without restoring the AMR value. This also helps the case when applications are not using the pkey feature. In this case, different applications will have the same AMR values and hence we can avoid restoring AMR in this case too. Also avoid isync() if not really needed. Do the same for IAMR. null-syscall benchmark results: With smap/smep disabled: Without patch: 957.95 ns 2778.17 cycles With patch: 858.38 ns 2489.30 cycles With smap/smep enabled: Without patch: 1017.26 ns 2950.36 cycles With patch: 1021.51 ns 2962.44 cycles Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-23-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:24 +08:00
bool flush_needed = false;
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
/*
* When entering from userspace we mostly have the AMR/IAMR
* different from kernel default values. Hence don't compare.
*/
amr = mfspr(SPRN_AMR);
iamr = mfspr(SPRN_IAMR);
regs->amr = amr;
regs->iamr = iamr;
powerpc/book3s64/pkeys: Optimize KUAP and KUEP feature disabled case If FTR_BOOK3S_KUAP is disabled, kernel will continue to run with the same AMR value with which it was entered. Hence there is a high chance that we can return without restoring the AMR value. This also helps the case when applications are not using the pkey feature. In this case, different applications will have the same AMR values and hence we can avoid restoring AMR in this case too. Also avoid isync() if not really needed. Do the same for IAMR. null-syscall benchmark results: With smap/smep disabled: Without patch: 957.95 ns 2778.17 cycles With patch: 858.38 ns 2489.30 cycles With smap/smep enabled: Without patch: 1017.26 ns 2950.36 cycles With patch: 1021.51 ns 2962.44 cycles Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-23-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:24 +08:00
if (mmu_has_feature(MMU_FTR_BOOK3S_KUAP)) {
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
mtspr(SPRN_AMR, AMR_KUAP_BLOCKED);
powerpc/book3s64/pkeys: Optimize KUAP and KUEP feature disabled case If FTR_BOOK3S_KUAP is disabled, kernel will continue to run with the same AMR value with which it was entered. Hence there is a high chance that we can return without restoring the AMR value. This also helps the case when applications are not using the pkey feature. In this case, different applications will have the same AMR values and hence we can avoid restoring AMR in this case too. Also avoid isync() if not really needed. Do the same for IAMR. null-syscall benchmark results: With smap/smep disabled: Without patch: 957.95 ns 2778.17 cycles With patch: 858.38 ns 2489.30 cycles With smap/smep enabled: Without patch: 1017.26 ns 2950.36 cycles With patch: 1021.51 ns 2962.44 cycles Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-23-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:24 +08:00
flush_needed = true;
}
if (mmu_has_feature(MMU_FTR_BOOK3S_KUEP)) {
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
mtspr(SPRN_IAMR, AMR_KUEP_BLOCKED);
powerpc/book3s64/pkeys: Optimize KUAP and KUEP feature disabled case If FTR_BOOK3S_KUAP is disabled, kernel will continue to run with the same AMR value with which it was entered. Hence there is a high chance that we can return without restoring the AMR value. This also helps the case when applications are not using the pkey feature. In this case, different applications will have the same AMR values and hence we can avoid restoring AMR in this case too. Also avoid isync() if not really needed. Do the same for IAMR. null-syscall benchmark results: With smap/smep disabled: Without patch: 957.95 ns 2778.17 cycles With patch: 858.38 ns 2489.30 cycles With smap/smep enabled: Without patch: 1017.26 ns 2950.36 cycles With patch: 1021.51 ns 2962.44 cycles Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-23-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:24 +08:00
flush_needed = true;
}
if (flush_needed)
isync();
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
} else
#endif
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
kuap_assert_locked();
powerpc/64/kuap: Move kuap checks out of MSR[RI]=0 regions of exit code Any kind of WARN causes a program check that will crash with unrecoverable exception if it occurs when RI is clear. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429065654.1677541-2-npiggin@gmail.com
2020-04-29 14:56:49 +08:00
powerpc/32: Handle bookE debugging in C in syscall entry/exit The handling of SPRN_DBCR0 and other registers can easily be done in C instead of ASM. For that, create booke_load_dbcr0() and booke_restore_dbcr0(). Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1a7515f9258b27a9177de88491a8bb79b255ceb7.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:28 +08:00
booke_restore_dbcr0();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
account_cpu_user_entry();
powerpc/64: move account_stolen_time into its own function This will be used by interrupt entry as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-38-npiggin@gmail.com
2021-01-30 21:08:47 +08:00
account_stolen_time();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
/*
* This is not required for the syscall exit path, but makes the
* stack frame look nicer. If this was initialised in the first stack
* frame, or if the unwinder was taught the first stack frame always
* returns to user with IRQS_ENABLED, this store could be avoided!
*/
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
irq_soft_mask_regs_set_state(regs, IRQS_ENABLED);
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
powerpc/64/syscall: Reconcile interrupts This reconciles interrupts in the system call case like all other interrupts. This allows system_call_common to be shared with the scv system call implementation in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-31-npiggin@gmail.com
2020-02-26 01:35:39 +08:00
local_irq_enable();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
powerpc/64: make buildable without CONFIG_COMPAT There are numerous references to 32bit functions in generic and 64bit code so ifdef them out. Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/e5619617020ef3a1f54f0c076e7d74cb9ec9f3bf.1584699455.git.msuchanek@suse.de
2020-03-20 18:20:16 +08:00
if (unlikely(current_thread_info()->flags & _TIF_SYSCALL_DOTRACE)) {
powerpc/syscall: Do not check unsupported scv vector on PPC32 Only book3s/64 has scv. No need to check the 0x7ff0 trap on 32 or 64e. For that, add a helper trap_is_unsupported_scv() similar to trap_is_scv(). And ignore the scv parameter in syscall_exit_prepare (Save 14 cycles 346 => 332 cycles) Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/fb87b205ae8eb8c623f33bb316801acf95a831e6.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:27 +08:00
if (unlikely(trap_is_unsupported_scv(regs))) {
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
/* Unsupported scv vector */
_exception(SIGILL, regs, ILL_ILLOPC, regs->nip);
return regs->gpr[3];
}
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
/*
* We use the return value of do_syscall_trace_enter() as the
* syscall number. If the syscall was rejected for any reason
* do_syscall_trace_enter() returns an invalid syscall number
* and the test against NR_syscalls will fail and the return
* value to be used is in regs->gpr[3].
*/
r0 = do_syscall_trace_enter(regs);
if (unlikely(r0 >= NR_syscalls))
return regs->gpr[3];
r3 = regs->gpr[3];
r4 = regs->gpr[4];
r5 = regs->gpr[5];
r6 = regs->gpr[6];
r7 = regs->gpr[7];
r8 = regs->gpr[8];
} else if (unlikely(r0 >= NR_syscalls)) {
powerpc/syscall: Do not check unsupported scv vector on PPC32 Only book3s/64 has scv. No need to check the 0x7ff0 trap on 32 or 64e. For that, add a helper trap_is_unsupported_scv() similar to trap_is_scv(). And ignore the scv parameter in syscall_exit_prepare (Save 14 cycles 346 => 332 cycles) Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/fb87b205ae8eb8c623f33bb316801acf95a831e6.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:27 +08:00
if (unlikely(trap_is_unsupported_scv(regs))) {
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
/* Unsupported scv vector */
_exception(SIGILL, regs, ILL_ILLOPC, regs->nip);
return regs->gpr[3];
}
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
return -ENOSYS;
}
/* May be faster to do array_index_nospec? */
barrier_nospec();
powerpc/syscall: Use is_compat_task() Instead of hard comparing task flags with _TIF_32BIT, use is_compat_task(). The advantage is that it returns 0 on PPC32 allthough _TIF_32BIT is always set. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/c8094662199337a7200fea9f6e1d1f8b1b6d5f69.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:29 +08:00
if (unlikely(is_compat_task())) {
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
f = (void *)compat_sys_call_table[r0];
r3 &= 0x00000000ffffffffULL;
r4 &= 0x00000000ffffffffULL;
r5 &= 0x00000000ffffffffULL;
r6 &= 0x00000000ffffffffULL;
r7 &= 0x00000000ffffffffULL;
r8 &= 0x00000000ffffffffULL;
} else {
f = (void *)sys_call_table[r0];
}
return f(r3, r4, r5, r6, r7, r8);
}
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
/*
* local irqs must be disabled. Returns false if the caller must re-enable
* them, check for new work, and try again.
powerpc/64s: interrupt exit improve bounding of interrupt recursion When replaying pending soft-masked interrupts when an interrupt returns to an irqs-enabled context, there is a special case required if this was an asynchronous interrupt to avoid unbounded interrupt recursion. This case was not tested for in the case the asynchronous interrupt hit in user context, because a subsequent nested interrupt would by definition hit in kernel mode, which then exits via the kernel path which does test this case. There is no reason to allow this for such interrupts. While recursion is bounded at the next level, it's simpler and uses less stack to apply the replay logic consistently. This also expands the comment which was really pretty poor and didn't explain the problem (I can say that because I wrote it). Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-2-npiggin@gmail.com
2021-01-30 21:08:11 +08:00
*
* This should be called with local irqs disabled, but if they were previously
* enabled when the interrupt handler returns (indicating a process-context /
* synchronous interrupt) then irqs_enabled should be true.
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
*/
powerpc/syscall: Force inlining of __prep_irq_for_enabled_exit() As reported by kernel test robot, a randconfig with high amount of debuging options can lead to build failure for undefined reference to replay_soft_interrupts() on ppc32. This is due to gcc not seeing that __prep_irq_for_enabled_exit() always returns true on ppc32 because it doesn't inline it for some reason. Force inlining of __prep_irq_for_enabled_exit() to fix the build. Fixes: 344bb20b159d ("powerpc/syscall: Make interrupt.c buildable on PPC32") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/53f3a1f719441761000c41154602bf097d4350b5.1614148356.git.christophe.leroy@csgroup.eu
2021-02-24 14:34:22 +08:00
static notrace __always_inline bool __prep_irq_for_enabled_exit(bool clear_ri)
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
{
/* This must be done with RI=1 because tracing may touch vmaps */
trace_hardirqs_on();
/* This pattern matches prep_irq_for_idle */
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
if (clear_ri)
__hard_EE_RI_disable();
else
__hard_irq_disable();
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
#ifdef CONFIG_PPC64
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
if (unlikely(lazy_irq_pending_nocheck())) {
/* Took an interrupt, may have more exit work to do. */
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
if (clear_ri)
__hard_RI_enable();
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
trace_hardirqs_off();
local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
return false;
}
local_paca->irq_happened = 0;
irq_soft_mask_set(IRQS_ENABLED);
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
#endif
powerpc/64: Refactor interrupt exit irq disabling sequence The same complicated sequence for juggling EE, RI, soft mask, and irq tracing is repeated 3 times, tidy these up into one function. This differs qiute a bit between sub architectures, so this makes the ppc32 port cleaner as well. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429062421.1675400-1-npiggin@gmail.com
2020-04-29 14:24:21 +08:00
return true;
}
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
static notrace inline bool prep_irq_for_enabled_exit(bool clear_ri, bool irqs_enabled)
{
if (__prep_irq_for_enabled_exit(clear_ri))
return true;
/*
* Must replay pending soft-masked interrupts now. Don't just
* local_irq_enabe(); local_irq_disable(); because if we are
* returning from an asynchronous interrupt here, another one
* might hit after irqs are enabled, and it would exit via this
* same path allowing another to fire, and so on unbounded.
*
* If interrupts were enabled when this interrupt exited,
* indicating a process context (synchronous) interrupt,
* local_irq_enable/disable can be used, which will enable
* interrupts rather than keeping them masked (unclear how
* much benefit this is over just replaying for all cases,
* because we immediately disable again, so all we're really
* doing is allowing hard interrupts to execute directly for
* a very small time, rather than being masked and replayed).
*/
if (irqs_enabled) {
local_irq_enable();
local_irq_disable();
} else {
replay_soft_interrupts();
}
return false;
}
powerpc/32: Handle bookE debugging in C in syscall entry/exit The handling of SPRN_DBCR0 and other registers can easily be done in C instead of ASM. For that, create booke_load_dbcr0() and booke_restore_dbcr0(). Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1a7515f9258b27a9177de88491a8bb79b255ceb7.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:28 +08:00
static notrace void booke_load_dbcr0(void)
{
#ifdef CONFIG_PPC_ADV_DEBUG_REGS
unsigned long dbcr0 = current->thread.debug.dbcr0;
if (likely(!(dbcr0 & DBCR0_IDM)))
return;
/*
* Check to see if the dbcr0 register is set up to debug.
* Use the internal debug mode bit to do this.
*/
mtmsr(mfmsr() & ~MSR_DE);
if (IS_ENABLED(CONFIG_PPC32)) {
isync();
global_dbcr0[smp_processor_id()] = mfspr(SPRN_DBCR0);
}
mtspr(SPRN_DBCR0, dbcr0);
mtspr(SPRN_DBSR, -1);
#endif
}
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
/*
* This should be called after a syscall returns, with r3 the return value
* from the syscall. If this function returns non-zero, the system call
* exit assembly should additionally load all GPR registers and CTR and XER
* from the interrupt frame.
*
* The function graph tracer can not trace the return side of this function,
* because RI=0 and soft mask state is "unreconciled", so it is marked notrace.
*/
notrace unsigned long syscall_exit_prepare(unsigned long r3,
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
struct pt_regs *regs,
long scv)
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
{
unsigned long ti_flags;
unsigned long ret = 0;
powerpc/syscall: Do not check unsupported scv vector on PPC32 Only book3s/64 has scv. No need to check the 0x7ff0 trap on 32 or 64e. For that, add a helper trap_is_unsupported_scv() similar to trap_is_scv(). And ignore the scv parameter in syscall_exit_prepare (Save 14 cycles 346 => 332 cycles) Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/fb87b205ae8eb8c623f33bb316801acf95a831e6.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:27 +08:00
bool is_not_scv = !IS_ENABLED(CONFIG_PPC_BOOK3S_64) || !scv;
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
CT_WARN_ON(ct_state() == CONTEXT_USER);
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
kuap_assert_locked();
powerpc/64/kuap: Move kuap checks out of MSR[RI]=0 regions of exit code Any kind of WARN causes a program check that will crash with unrecoverable exception if it occurs when RI is clear. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429065654.1677541-2-npiggin@gmail.com
2020-04-29 14:56:49 +08:00
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
regs->result = r3;
/* Check whether the syscall is issued inside a restartable sequence */
rseq_syscall(regs);
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
ti_flags = current_thread_info()->flags;
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
powerpc/syscall: Do not check unsupported scv vector on PPC32 Only book3s/64 has scv. No need to check the 0x7ff0 trap on 32 or 64e. For that, add a helper trap_is_unsupported_scv() similar to trap_is_scv(). And ignore the scv parameter in syscall_exit_prepare (Save 14 cycles 346 => 332 cycles) Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/fb87b205ae8eb8c623f33bb316801acf95a831e6.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:27 +08:00
if (unlikely(r3 >= (unsigned long)-MAX_ERRNO) && is_not_scv) {
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
if (likely(!(ti_flags & (_TIF_NOERROR | _TIF_RESTOREALL)))) {
r3 = -r3;
regs->ccr |= 0x10000000; /* Set SO bit in CR */
}
}
if (unlikely(ti_flags & _TIF_PERSYSCALL_MASK)) {
if (ti_flags & _TIF_RESTOREALL)
ret = _TIF_RESTOREALL;
else
regs->gpr[3] = r3;
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
clear_bits(_TIF_PERSYSCALL_MASK, &current_thread_info()->flags);
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
} else {
regs->gpr[3] = r3;
}
if (unlikely(ti_flags & _TIF_SYSCALL_DOTRACE)) {
do_syscall_trace_leave(regs);
ret |= _TIF_RESTOREALL;
}
local_irq_disable();
powerpc/64s: interrupt exit improve bounding of interrupt recursion When replaying pending soft-masked interrupts when an interrupt returns to an irqs-enabled context, there is a special case required if this was an asynchronous interrupt to avoid unbounded interrupt recursion. This case was not tested for in the case the asynchronous interrupt hit in user context, because a subsequent nested interrupt would by definition hit in kernel mode, which then exits via the kernel path which does test this case. There is no reason to allow this for such interrupts. While recursion is bounded at the next level, it's simpler and uses less stack to apply the replay logic consistently. This also expands the comment which was really pretty poor and didn't explain the problem (I can say that because I wrote it). Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-2-npiggin@gmail.com
2021-01-30 21:08:11 +08:00
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
again:
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
ti_flags = READ_ONCE(current_thread_info()->flags);
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
while (unlikely(ti_flags & (_TIF_USER_WORK_MASK & ~_TIF_RESTORE_TM))) {
local_irq_enable();
if (ti_flags & _TIF_NEED_RESCHED) {
schedule();
} else {
/*
* SIGPENDING must restore signal handler function
* argument GPRs, and some non-volatiles (e.g., r1).
* Restore all for now. This could be made lighter.
*/
if (ti_flags & _TIF_SIGPENDING)
ret |= _TIF_RESTOREALL;
do_notify_resume(regs, ti_flags);
}
local_irq_disable();
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
ti_flags = READ_ONCE(current_thread_info()->flags);
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
}
powerpc/interrupt: Fix CONFIG ifdef typo CONFIG_PPC_BOOK3S should be CONFIG_PPC_BOOK3S_64. restore_math is a no-op for other configurations. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> [np: split from another patch] Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210617155116.2167984-2-npiggin@gmail.com
2021-06-17 23:51:00 +08:00
if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && IS_ENABLED(CONFIG_PPC_FPU)) {
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
if (IS_ENABLED(CONFIG_PPC_TRANSACTIONAL_MEM) &&
unlikely((ti_flags & _TIF_RESTORE_TM))) {
restore_tm_state(regs);
} else {
unsigned long mathflags = MSR_FP;
if (cpu_has_feature(CPU_FTR_VSX))
mathflags |= MSR_VEC | MSR_VSX;
else if (cpu_has_feature(CPU_FTR_ALTIVEC))
mathflags |= MSR_VEC;
powerpc/64s: Fix restore_math unnecessarily changing MSR Before returning to user, if there are missing FP/VEC/VSX bits from the user MSR then those registers had been saved and must be restored again before use. restore_math will decide whether to restore immediately, or skip the restore and let fp/vec/vsx unavailable faults demand load the registers. Each time restore_math restores one of the FP/VSX or VEC register sets is loaded, an 8-bit counter is incremented (load_fp and load_vec). When these wrap to zero, restore_math no longer restores that register set until after they are next demand faulted. It's quite usual for those counters to have different values, so if one wraps to zero and restore_math no longer restores its registers or user MSR bit but the other is not zero yet does not need to be restored (because the kernel is not frequently using the FPU), then restore_math will be called and it will also not return in the early exit check. This causes msr_check_and_set to test and set the MSR at every kernel exit despite having no work to do. This can cause workloads (e.g., a NULL syscall microbenchmark) to run fast for a time while both counters are non-zero, then slow down when one of the counters reaches zero, then speed up again after the second counter reaches zero. The cost is significant, about 10% slowdown on a NULL syscall benchmark, and the jittery behaviour is very undesirable. Fix this by having restore_math test all conditions first, and only update MSR if we will be loading registers. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200623234139.2262227-2-npiggin@gmail.com
2020-06-24 07:41:38 +08:00
/*
* If userspace MSR has all available FP bits set,
* then they are live and no need to restore. If not,
* it means the regs were given up and restore_math
* may decide to restore them (to avoid taking an FP
* fault).
*/
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
if ((regs->msr & mathflags) != mathflags)
restore_math(regs);
}
}
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
user_enter_irqoff();
powerpc/64s: system call support for scv/rfscv instructions Add support for the scv instruction on POWER9 and later CPUs. For now this implements the zeroth scv vector 'scv 0', as identical to 'sc' system calls, with the exception that LR is not preserved, nor are volatile CR registers, and error is not indicated with CR0[SO], but by returning a negative errno. rfscv is implemented to return from scv type system calls. It can not be used to return from sc system calls because those are defined to preserve LR. getpid syscall throughput on POWER9 is improved by 26% (428 to 318 cycles), largely due to reducing mtmsr and mtspr. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> [mpe: Fix ppc64e build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200611081203.995112-3-npiggin@gmail.com
2020-06-11 16:12:03 +08:00
/* scv need not set RI=0 because SRRs are not used */
powerpc/syscall: Do not check unsupported scv vector on PPC32 Only book3s/64 has scv. No need to check the 0x7ff0 trap on 32 or 64e. For that, add a helper trap_is_unsupported_scv() similar to trap_is_scv(). And ignore the scv parameter in syscall_exit_prepare (Save 14 cycles 346 => 332 cycles) Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/fb87b205ae8eb8c623f33bb316801acf95a831e6.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:27 +08:00
if (unlikely(!__prep_irq_for_enabled_exit(is_not_scv))) {
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
user_exit_irqoff();
local_irq_enable();
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
local_irq_disable();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
goto again;
powerpc/64: context tracking remove _TIF_NOHZ Add context tracking to the system call handler explicitly, and remove _TIF_NOHZ. This improves system call performance when nohz_full is enabled. On a POWER9, gettid scv system call cost on a nohz_full CPU improves from 1129 cycles to 1004 cycles and on a housekeeping CPU from 550 cycles to 430 cycles. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-31-npiggin@gmail.com
2021-01-30 21:08:40 +08:00
}
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
local_paca->tm_scratch = regs->msr;
#endif
powerpc/32: Handle bookE debugging in C in syscall entry/exit The handling of SPRN_DBCR0 and other registers can easily be done in C instead of ASM. For that, create booke_load_dbcr0() and booke_restore_dbcr0(). Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1a7515f9258b27a9177de88491a8bb79b255ceb7.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:28 +08:00
booke_load_dbcr0();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
account_cpu_user_exit();
powerpc/interrupt: update common interrupt code for This makes adjustments to 64-bit asm and common C interrupt return code to be usable by the 64e subarchitecture. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210316104206.407354-4-npiggin@gmail.com
2021-03-16 18:41:57 +08:00
/* Restore user access locks last */
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
kuap_user_restore(regs);
powerpc/32s: Move KUEP locking/unlocking in C This can be done in C, do it. Unrolling the loop gains approx. 15% performance. From now on, prepare_transfer_to_handler() is only for interrupts from kernel. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/4eadd873927e9a73c3d1dfe2f9497353465514cf.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:47 +08:00
kuep_unlock();
powerpc/64/sycall: Implement syscall entry/exit logic in C System call entry and particularly exit code is beyond the limit of what is reasonable to implement in asm. This conversion moves all conditional branches out of the asm code, except for the case that all GPRs should be restored at exit. Null syscall test is about 5% faster after this patch, because the exit work is handled under local_irq_disable, and the hard mask and pending interrupt replay is handled after that, which avoids games with MSR. mpe: Includes subsequent fixes from Nick: This fixes 4 issues caught by TM selftests. First was a tm-syscall bug that hit due to tabort_syscall being called after interrupts were reconciled (in a subsequent patch), which led to interrupts being enabled before tabort_syscall was called. Rather than going through an un-reconciling interrupts for the return, I just go back to putting the test early in asm, the C-ification of that wasn't a big win anyway. Second is the syscall return _TIF_USER_WORK_MASK check would go into an infinite loop if _TIF_RESTORE_TM became set. The asm code uses _TIF_USER_WORK_MASK to brach to slowpath which includes restore_tm_state. Third is system call return was not calling restore_tm_state, I missed this completely (alhtough it's in the return from interrupt C conversion because when the asm syscall code encountered problems it would branch to the interrupt return code. Fourth is MSR_VEC missing from restore_math, which was caught by tm-unavailable selftest taking an unexpected facility unavailable interrupt when testing VSX unavailble exception with MSR.FP=1 MSR.VEC=1. Fourth case also has a fixup in a subsequent patch. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-26-npiggin@gmail.com
2020-02-26 01:35:34 +08:00
return ret;
}
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
powerpc: remove interrupt exit helpers unused argument The msr argument is not used, remove it. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210617155116.2167984-3-npiggin@gmail.com
2021-06-17 23:51:01 +08:00
notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs)
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
{
unsigned long ti_flags;
unsigned long flags;
unsigned long ret = 0;
powerpc/syscall: Change condition to check MSR_RI In system_call_exception(), MSR_RI also needs to be checked on 8xx. Only booke and 40x doesn't have MSR_RI. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/67820fada8dd6a8fe9d7b666f175d4cc9d8de87e.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:31 +08:00
if (!IS_ENABLED(CONFIG_BOOKE) && !IS_ENABLED(CONFIG_40x))
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
BUG_ON(!(regs->msr & MSR_RI));
BUG_ON(!(regs->msr & MSR_PR));
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
BUG_ON(arch_irq_disabled_regs(regs));
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
CT_WARN_ON(ct_state() == CONTEXT_USER);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
powerpc/64/kuap: Conditionally restore AMR in interrupt exit The AMR update is made conditional on AMR actually changing, which should be the less common case on most workloads (though kernel page faults on uaccess could be frequent, this doesn't significantly slow down that case). Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429065654.1677541-4-npiggin@gmail.com
2020-04-29 14:56:51 +08:00
/*
* We don't need to restore AMR on the way back to userspace for KUAP.
* AMR can only have been unlocked if we interrupted the kernel.
*/
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
kuap_assert_locked();
powerpc/64/kuap: Move kuap checks out of MSR[RI]=0 regions of exit code Any kind of WARN causes a program check that will crash with unrecoverable exception if it occurs when RI is clear. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429065654.1677541-2-npiggin@gmail.com
2020-04-29 14:56:49 +08:00
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
local_irq_save(flags);
again:
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
ti_flags = READ_ONCE(current_thread_info()->flags);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
while (unlikely(ti_flags & (_TIF_USER_WORK_MASK & ~_TIF_RESTORE_TM))) {
local_irq_enable(); /* returning to user: may enable */
if (ti_flags & _TIF_NEED_RESCHED) {
schedule();
} else {
if (ti_flags & _TIF_SIGPENDING)
ret |= _TIF_RESTOREALL;
do_notify_resume(regs, ti_flags);
}
local_irq_disable();
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
ti_flags = READ_ONCE(current_thread_info()->flags);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
}
powerpc/32: Replace ASM exception exit by C exception exit from ppc64 This patch replaces the PPC32 ASM exception exit by C exception exit. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/48f8bae91da899d8e73fc0d75c9af66cc97b4d5b.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:34 +08:00
if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && IS_ENABLED(CONFIG_PPC_FPU)) {
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
if (IS_ENABLED(CONFIG_PPC_TRANSACTIONAL_MEM) &&
unlikely((ti_flags & _TIF_RESTORE_TM))) {
restore_tm_state(regs);
} else {
unsigned long mathflags = MSR_FP;
if (cpu_has_feature(CPU_FTR_VSX))
mathflags |= MSR_VEC | MSR_VSX;
else if (cpu_has_feature(CPU_FTR_ALTIVEC))
mathflags |= MSR_VEC;
powerpc/64s: Fix restore_math unnecessarily changing MSR Before returning to user, if there are missing FP/VEC/VSX bits from the user MSR then those registers had been saved and must be restored again before use. restore_math will decide whether to restore immediately, or skip the restore and let fp/vec/vsx unavailable faults demand load the registers. Each time restore_math restores one of the FP/VSX or VEC register sets is loaded, an 8-bit counter is incremented (load_fp and load_vec). When these wrap to zero, restore_math no longer restores that register set until after they are next demand faulted. It's quite usual for those counters to have different values, so if one wraps to zero and restore_math no longer restores its registers or user MSR bit but the other is not zero yet does not need to be restored (because the kernel is not frequently using the FPU), then restore_math will be called and it will also not return in the early exit check. This causes msr_check_and_set to test and set the MSR at every kernel exit despite having no work to do. This can cause workloads (e.g., a NULL syscall microbenchmark) to run fast for a time while both counters are non-zero, then slow down when one of the counters reaches zero, then speed up again after the second counter reaches zero. The cost is significant, about 10% slowdown on a NULL syscall benchmark, and the jittery behaviour is very undesirable. Fix this by having restore_math test all conditions first, and only update MSR if we will be loading registers. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200623234139.2262227-2-npiggin@gmail.com
2020-06-24 07:41:38 +08:00
/* See above restore_math comment */
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
if ((regs->msr & mathflags) != mathflags)
restore_math(regs);
}
}
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
user_enter_irqoff();
if (unlikely(!__prep_irq_for_enabled_exit(true))) {
user_exit_irqoff();
local_irq_enable();
local_irq_disable();
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
goto again;
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
}
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
powerpc/32: Handle bookE debugging in C in syscall entry/exit The handling of SPRN_DBCR0 and other registers can easily be done in C instead of ASM. For that, create booke_load_dbcr0() and booke_restore_dbcr0(). Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1a7515f9258b27a9177de88491a8bb79b255ceb7.1612898425.git.christophe.leroy@csgroup.eu
2021-02-10 03:29:28 +08:00
booke_load_dbcr0();
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
local_paca->tm_scratch = regs->msr;
#endif
account_cpu_user_exit();
powerpc/interrupt: update common interrupt code for This makes adjustments to 64-bit asm and common C interrupt return code to be usable by the 64e subarchitecture. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210316104206.407354-4-npiggin@gmail.com
2021-03-16 18:41:57 +08:00
/* Restore user access locks last */
powerpc/book3s64/pkeys: Store/restore userspace AMR/IAMR correctly on entry and exit from kernel This prepare kernel to operate with a different value than userspace AMR/IAMR. For this, AMR/IAMR need to be saved and restored on entry and return from the kernel. With KUAP we modify kernel AMR when accessing user address from the kernel via copy_to/from_user interfaces. We don't need to modify IAMR value in similar fashion. If MMU_FTR_PKEY is enabled we need to save AMR/IAMR in pt_regs on entering kernel from userspace. If not we can assume that AMR/IAMR is not modified from userspace. We need to save AMR if we have MMU_FTR_BOOK3S_KUAP feature enabled and we are interrupted within kernel. This is required so that if we get interrupted within copy_to/from_user we continue with the right AMR value. If we hae MMU_FTR_BOOK3S_KUEP enabled we need to restore IAMR on return to userspace beause kernel will be running with a different IAMR value. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20201127044424.40686-11-aneesh.kumar@linux.ibm.com
2020-11-27 12:44:12 +08:00
kuap_user_restore(regs);
powerpc/interrupts: Fix kuep_unlock() call Same as kuap_user_restore(), kuep_unlock() has to be called when really returning to user, that is in interrupt_exit_user_prepare(), not in interrupt_exit_prepare(). Fixes: b5efec00b671 ("powerpc/32s: Move KUEP locking/unlocking in C") Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/b831e54a2579db24fbef836ed415588ce2b3e825.1620312573.git.christophe.leroy@csgroup.eu
2021-05-06 22:49:45 +08:00
kuep_unlock();
powerpc/interrupt: update common interrupt code for This makes adjustments to 64-bit asm and common C interrupt return code to be usable by the 64e subarchitecture. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210316104206.407354-4-npiggin@gmail.com
2021-03-16 18:41:57 +08:00
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
return ret;
}
void preempt_schedule_irq(void);
powerpc: remove interrupt exit helpers unused argument The msr argument is not used, remove it. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210617155116.2167984-3-npiggin@gmail.com
2021-06-17 23:51:01 +08:00
notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs)
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
{
unsigned long flags;
unsigned long ret = 0;
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
unsigned long kuap;
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
powerpc/syscall: Change condition to check MSR_RI In system_call_exception(), MSR_RI also needs to be checked on 8xx. Only booke and 40x doesn't have MSR_RI. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/67820fada8dd6a8fe9d7b666f175d4cc9d8de87e.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:31 +08:00
if (!IS_ENABLED(CONFIG_BOOKE) && !IS_ENABLED(CONFIG_40x) &&
unlikely(!(regs->msr & MSR_RI)))
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
unrecoverable_exception(regs);
BUG_ON(regs->msr & MSR_PR);
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
/*
* CT_WARN_ON comes here via program_check_exception,
* so avoid recursion.
*/
powerpc/traps: Enhance readability for trap types Define macros to list ppc interrupt types in interttupt.h, replace the reference of the trap hex values with these macros. Referred the hex numbers in arch/powerpc/kernel/exceptions-64e.S, arch/powerpc/kernel/exceptions-64s.S, arch/powerpc/kernel/head_*.S, arch/powerpc/kernel/head_booke.h and arch/powerpc/include/asm/kvm_asm.h. Signed-off-by: Xiongwei Song <sxwjean@gmail.com> [mpe: Resolve conflicts in nmi_disables_ftrace(), fix 40x build] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1618398033-13025-1-git-send-email-sxwjean@me.com
2021-04-14 19:00:33 +08:00
if (TRAP(regs) != INTERRUPT_PROGRAM)
powerpc/64s: move context tracking exit to interrupt exit path The interrupt handler wrapper functions are not the ideal place to maintain context tracking because after they return, the low level exit code must then determine if there are interrupts to replay, or if the task should be preempted, etc. Those paths (e.g., schedule_user) include their own exception_enter/exit pairs to fix this up but it's a bit hacky (see schedule_user() comments). Ideally context tracking will go to user mode only when there are no more interrupts or context switches or other exit processing work to handle. 64e can not do this because it does not use the C interrupt exit code. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-36-npiggin@gmail.com
2021-01-30 21:08:45 +08:00
CT_WARN_ON(ct_state() == CONTEXT_USER);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
kuap = kuap_get_and_assert_locked();
powerpc/64/kuap: Move kuap checks out of MSR[RI]=0 regions of exit code Any kind of WARN causes a program check that will crash with unrecoverable exception if it occurs when RI is clear. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200429065654.1677541-2-npiggin@gmail.com
2020-04-29 14:56:49 +08:00
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
if (unlikely(current_thread_info()->flags & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, &current_thread_info()->flags);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
ret = 1;
}
local_irq_save(flags);
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
if (!arch_irq_disabled_regs(regs)) {
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
/* Returning to a kernel context with local irqs enabled. */
WARN_ON_ONCE(!(regs->msr & MSR_EE));
again:
if (IS_ENABLED(CONFIG_PREEMPT)) {
/* Return to preemptible kernel context */
powerpc/syscall: Avoid storing 'current' in another pointer By saving the pointer pointing to thread_info.flags, gcc copies r2 in a non-volatile register. We know 'current' doesn't change, so avoid that intermediaite pointer. Reduces null_syscall benchmark by 2 cycles (322 => 320 cycles) On PPC64, gcc seems to know that 'current' is not changing, and it keeps it in a non volatile register to avoid multiple read of 'current' in paca. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ad0363ff0ff8c125f40e1cdc589a85bbd7e31693.1612946484.git.christophe.leroy@csgroup.eu
2021-02-10 16:44:09 +08:00
if (unlikely(current_thread_info()->flags & _TIF_NEED_RESCHED)) {
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
if (preempt_count() == 0)
preempt_schedule_irq();
}
}
powerpc/64s: interrupt exit improve bounding of interrupt recursion When replaying pending soft-masked interrupts when an interrupt returns to an irqs-enabled context, there is a special case required if this was an asynchronous interrupt to avoid unbounded interrupt recursion. This case was not tested for in the case the asynchronous interrupt hit in user context, because a subsequent nested interrupt would by definition hit in kernel mode, which then exits via the kernel path which does test this case. There is no reason to allow this for such interrupts. While recursion is bounded at the next level, it's simpler and uses less stack to apply the replay logic consistently. This also expands the comment which was really pretty poor and didn't explain the problem (I can say that because I wrote it). Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-2-npiggin@gmail.com
2021-01-30 21:08:11 +08:00
if (unlikely(!prep_irq_for_enabled_exit(true, !irqs_disabled_flags(flags))))
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
goto again;
} else {
/* Returning to a kernel context with local irqs disabled. */
__hard_EE_RI_disable();
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
#ifdef CONFIG_PPC64
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
if (regs->msr & MSR_EE)
local_paca->irq_happened &= ~PACA_IRQ_HARD_DIS;
powerpc/syscall: Make interrupt.c buildable on PPC32 To allow building interrupt.c on PPC32, ifdef out specific PPC64 code or use helpers which are available on both PP32 and PPC64 Modify Makefile to always build interrupt.o Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/ba073ad67bd971a88ce331b65d6655523b54c794.1612796617.git.christophe.leroy@csgroup.eu
2021-02-08 23:10:28 +08:00
#endif
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
}
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
local_paca->tm_scratch = regs->msr;
#endif
/*
powerpc/interrupt: update common interrupt code for This makes adjustments to 64-bit asm and common C interrupt return code to be usable by the 64e subarchitecture. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210316104206.407354-4-npiggin@gmail.com
2021-03-16 18:41:57 +08:00
* 64s does not want to mfspr(SPRN_AMR) here, because this comes after
* mtmsr, which would cause Read-After-Write stalls. Hence, take the
* AMR value from the check above.
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
*/
powerpc/64s: Make kuap_check_amr() and kuap_get_and_check_amr() generic In preparation of porting powerpc32 to C syscall entry/exit, rename kuap_check_amr() and kuap_get_and_check_amr() as kuap_assert_locked() and kuap_get_and_assert_locked(), and move in the generic asm/kup.h the stub for when CONFIG_PPC_KUAP is not selected. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/f82614d9b17b83abd739aa18fc08811815d0c2e3.1615552867.git.christophe.leroy@csgroup.eu
2021-03-12 20:50:48 +08:00
kuap_kernel_restore(regs, kuap);
powerpc/64s: Implement interrupt exit logic in C Implement the bulk of interrupt return logic in C. The asm return code must handle a few cases: restoring full GPRs, and emulating stack store. The stack store emulation is significantly simplfied, rather than creating a new return frame and switching to that before performing the store, it uses the PACA to keep a scratch register around to perform the store. The asm return code is moved into 64e for now. The new logic has made allowance for 64e, but I don't have a full environment that works well to test it, and even booting in emulated qemu is not great for stress testing. 64e shouldn't be too far off working with this, given a bit more testing and auditing of the logic. This is slightly faster on a POWER9 (page fault speed increases about 1.1%), probably due to reduced mtmsrd. mpe: Includes fixes from Nick for _TIF_EMULATE_STACK_STORE handling (including the fast_interrupt_return path), to remove trace_hardirqs_on(), and fixes the interrupt-return part of the MSR_VSX restore bug caught by tm-unavailable selftest. mpe: Incorporate fix from Nick: The return-to-kernel path has to replay any soft-pending interrupts if it is returning to a context that had interrupts soft-enabled. It has to do this carefully and avoid plain enabling interrupts if this is an irq context, which can cause multiple nesting of interrupts on the stack, and other unexpected issues. The code which avoided this case got the soft-mask state wrong, and marked interrupts as enabled before going around again to retry. This seems to be mostly harmless except when PREEMPT=y, this calls preempt_schedule_irq with irqs apparently enabled and runs into a BUG in kernel/sched/core.c Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200225173541.1549955-29-npiggin@gmail.com
2020-02-26 01:35:37 +08:00
return ret;
}