Wireless extensions have the unfortunate problem that events
are multicast netlink messages, and are not independent of
pointer size. Thus, currently 32-bit tasks on 64-bit platforms
cannot properly receive events and fail with all kinds of
strange problems, for instance wpa_supplicant never notices
disassociations, due to the way the 64-bit event looks (to a
32-bit process), the fact that the address is all zeroes is
lost, it thinks instead it is 00:00:00:00:01:00.
The same problem existed with the ioctls, until David Miller
fixed those some time ago in an heroic effort.
A different problem caused by this is that we cannot send the
ASSOCREQIE/ASSOCRESPIE events because sending them causes a
32-bit wpa_supplicant on a 64-bit system to overwrite its
internal information, which is worse than it not getting the
information at all -- so we currently resort to sending a
custom string event that it then parses. This, however, has a
severe size limitation we are frequently hitting with modern
access points; this limitation would can be lifted after this
patch by sending the correct binary, not custom, event.
A similar problem apparently happens for some other netlink
users on x86_64 with 32-bit tasks due to the alignment for
64-bit quantities.
In order to fix these problems, I have implemented a way to
send compat messages to tasks. When sending an event, we send
the non-compat event data together with a compat event data in
skb_shinfo(main_skb)->frag_list. Then, when the event is read
from the socket, the netlink code makes sure to pass out only
the skb that is compatible with the task. This approach was
suggested by David Miller, my original approach required
always sending two skbs but that had various small problems.
To determine whether compat is needed or not, I have used the
MSG_CMSG_COMPAT flag, and adjusted the call path for recv and
recvfrom to include it, even if those calls do not have a cmsg
parameter.
I have not solved one small part of the problem, and I don't
think it is necessary to: if a 32-bit application uses read()
rather than any form of recvmsg() it will still get the wrong
(64-bit) event. However, neither do applications actually do
this, nor would it be a regression.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding smp_mb__after_lock define to be used as a smp_mb call after
a lock.
Making it nop for x86, since {read|write|spin}_lock() on x86 are
full memory barriers.
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch converts the remaining occurences of raw return values to their
symbolic counterparts in ndo_start_xmit() functions that were missed by the
previous automatic conversion.
Additionally code that assumed the symbolic value of NETDEV_TX_OK to be zero
is changed to explicitly use NETDEV_TX_OK.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is the result of an automatic spatch transformation to convert
all ndo_start_xmit() return values of 0 to NETDEV_TX_OK.
Some occurences are missed by the automatic conversion, those will be
handled in a seperate patch.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
MIPS: Fix CONFIG_FLATMEM version of pfn_valid()
MIPS: Reorganize Cavium OCTEON PCI support.
Update Yoichi Yuasa's e-mail address
MIPS: Allow suspend and hibernation again on uniprocessor kernels.
MIPS: 64-bit: Fix o32 core dump
MIPS: BC47xx: Fix SSB irq setup
MIPS: CMP: Update sync-r4k for current kernel
MIPS: CMP: Move gcmp_probe to before the SMP ops
MIPS: CMP: activate CMP support
MIPS: CMP: Extend IPI handling to CPU number
MIPS: CMP: Extend the GIC IPI interrupts beyond 32
MIPS: Define __arch_swab64 for all mips r2 cpus
MIPS: Update VR41xx GPIO driver to use gpiolib
MIPS: Hookup new syscalls sys_rt_tgsigqueueinfo and sys_perf_counter_open.
MIPS: Malta: Remove unnecessary function prototypes
MIPS: MT: Remove unnecessary semicolons
MIPS: Add support for Texas Instruments AR7 System-on-a-Chip
* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild-fixes:
kbuild: finally remove the obsolete variable $TOPDIR
gitignore: ignore scripts/ihex2fw
Kbuild: Disable the -Wformat-security gcc flag
gitignore: ignore gcov output files
kbuild: deb-pkg ship changelog
Add new __init_task_data macro to be used in arch init_task.c files.
asm-generic/vmlinux.lds.h: shuffle INIT_TASK* macro names in vmlinux.lds.h
Add new macros for page-aligned data and bss sections.
asm-generic/vmlinux.lds.h: Fix up RW_DATA_SECTION definition.
For systems which do not define PHYS_OFFSET as 0 pfn_valid() may falsely
have returned 0 on most configurations. Bug introduced by commit
752fbeb2e3555c0d236e992f1195fd7ce30e728d (linux-mips.org) rsp.
6f284a2ce7 (kernel.org) titled "[MIPS]
FLATMEM: introduce PHYS_OFFSET."
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Move the cavium PCI files to the arch/mips/pci directory. Also cleanup
comment formatting and code layout. Code from pci-common.c, was moved
into other files.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
If an o32 process generates a core dump on a 64 bit kernel, the core file
will not be correctly recognized. This is because ELF_CORE_COPY_REGS and
ELF_CORE_COPY_TASK_REGS are not correctly defined for o32 and will use
the default register set which would be CONFIG_64BIT in asm/elf.h.
So we'll switch to use the right register defines in this situation by
checking for WANT_COMPAT_REG_H and use the right defines of
ELF_CORE_COPY_REGS and ELF_CORE_COPY_TASK_REGS.
[Ralf: made ELF_CORE_COPY_TASK_REGS() bullet-proof against funny arguments.]
Signed-off-by: Yong Zhang <yong.zhang@windriver.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This revises the sync-4k so it will boot and operate since the removal of
expirelo from the timer code.
Signed-off-by: Tim Anderson <tanderson@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This is to move the gcmp_probe call to before the use of and selection of
the smp_ops functions. This allows malta with 1004K to work.
Signed-off-by: Tim Anderson <tanderson@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Most of the CMP support was added before, this mostly correct compile
problems but adds a platform specific translation for the interrupt number
based on cpu number.
Signed-off-by: Tim Anderson <tanderson@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This takes the current IPI interrupt assignment from the fix number of 4
to the number of CPUs defined in the system.
Signed-off-by: Tim Anderson <tanderson@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This patch extends the GIC interrupt handling beyond the current 32 bit
range as well as extending the number of interrupts based on the number
of CPUs.
Signed-off-by: Tim Anderson <tanderson@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Some CPUs implement mipsr2, but because they are a super-set of mips64r2 do
not define CONFIG_CPU_MIPS64_R2. Cavium OCTEON falls into this category.
We would still like to use the optimized implementation, so since we have
already checked for CONFIG_CPU_MIPSR2, checking for CONFIG_64BIT instead of
CONFIG_CPU_MIPS64_R2 is sufficient.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
[Ralf: I fixed up the numbering in the comment in scall64-n32.S.]
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
This patch adds support for the Texas Instruments AR7 System-on-a-Chip.
It supports the TNETD7100, 7200 and 7300 versions of the SoC.
Signed-off-by: Matteo Croce <matteo@openwrt.org>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Eugene Konev <ejka@openwrt.org>
Signed-off-by: Nicolas Thill <nico@openwrt.org>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Somewhat redundant since our atomic_t uses hashed-locks on 32-bit
anyway... Maybe we can clean those up to be generic too someday.
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
arch/parisc/mm/init.c: In function 'free_initmem':
381: warning: passing argument 1 of 'memset' makes pointer from integer without a cast
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
The TLB flushing functions on hppa, which causes PxTLB broadcasts on the system
bus, needs to be protected by irq-safe spinlocks to avoid irq handlers to deadlock
the kernel. The deadlocks only happened during I/O intensive loads and triggered
pretty seldom, which is why this bug went so long unnoticed.
Signed-off-by: Helge Deller <deller@gmx.de>
[edited to use spin_lock_irqsave on UP as well since we'd been locking there
all this time anyway, --kyle]
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Rewrote timer_interrupt() to properly handle the "delayed!" case.
If we used floating point math to compute the number of ticks that had
elapsed since the last timer interrupt, it could take up to 12K cycles
(emperical!) to handle the interrupt. Existing code assumed it would
never take more than 8k cycles. We end up programming Interval Timer
to a value less than "current" cycle counter. Thus have to wait until
Interval Timer "wrapped" and would then get the "delayed!" printk that
I moved below.
Since we don't really know what the upper limit is, I prefer to read
CR16 again after we've programmed it to make sure we won't have to
wait for CR16 to wrap.
Further, the printk was between reading CR16 (cycle couner) and writing CR16
(the interval timer). This would cause us to continue to set the interval
timer to a value that was "behind" the cycle counter. Rinse and repeat.
So no printk's between reading CR16 and setting next interval timer.
Tested on A500 (550 Mhz PA8600).
Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Tested-by: Kyle McMartin <kyle@mcmartin.ca>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
----
Kyle, Helge, and other parisc's,
Please test on 32-bit before committing.
I think I have it right but recognize I might not.
TODO: I wanted to use "do_div()" in order to get both remainder
and value back with one division op. That should help with the
latency alot but can be applied seperately from this patch.
thanks,
grant
>>>> I think this is what was intended? Note that this patch may affect
>>>> profiling.
>>> it really should be
>>>
>>> - if (likely(t1 & (sizeof(unsigned int)-1)) == 0) {
>>> + if (likely((t1 & (sizeof(unsigned int)-1)) == 0)) {
>>>
>>> randolph
Reported-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Randolph Chung <tausq@parisc-linux.org>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
The defines and typedefs (hw_interrupt_type, no_irq_type, irq_desc_t) have
been kept around for migration reasons. After more than two years it's
time to remove them finally.
This patch cleans up one of the remaining users. When all such patches
hit mainline we can remove the defines and typedefs finally.
Impact: cleanup
Convert the last remaining users to struct irq_chip and remove the
define.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Fix miscompilation in arch/parisc/kernel/irq.c:
123: warning: passing arg 1 of `cpumask_setall' from incompatible pointer type
141: warning: passing arg 1 of `cpumask_copy' from incompatible pointer type
300: warning: passing arg 1 of `cpumask_copy' from incompatible pointer type
357: warning: passing arg 2 of `cpumask_copy' from incompatible pointer type
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
There are two reasons to expose the memory *a in the asm:
1) To prevent the compiler from discarding a preceeding write to *a, and
2) to prevent it from caching *a in a register over the asm.
The change has had a few days testing with a SMP build of 2.6.22.19
running on a rp3440.
This patch is about the correctness of the __ldcw() macro itself.
The use of the macro should be confined to small inline functions
to try to limit the effect of clobbering memory on GCC's optimization
of loads and stores.
Signed-off-by: Dave Anglin <dave.anglin@nrc-cnrc.gc.ca>
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Doing an IPI with local interrupts off triggers a warning. We
don't need to be quite so ridiculously paranoid. Also, clean up
a bit of the code a little.
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
The atomic operations on parisc are defined as macros. The macros
includes casts which disallows the use of some syntax elements and
produces error like this:
net/phonet/pep.c: In function 'pipe_rcv_status':
net/phonet/pep.c:262: error: lvalue required as left operand of assignment
The patch removes this superfluous casts.
Signed-off-by: Bastian Blank <waldi@debian.org>
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
* git://git.infradead.org/iommu-2.6: (38 commits)
intel-iommu: Don't keep freeing page zero in dma_pte_free_pagetable()
intel-iommu: Introduce first_pte_in_page() to simplify PTE-setting loops
intel-iommu: Use cmpxchg64_local() for setting PTEs
intel-iommu: Warn about unmatched unmap requests
intel-iommu: Kill superfluous mapping_lock
intel-iommu: Ensure that PTE writes are 64-bit atomic, even on i386
intel-iommu: Make iommu=pt work on i386 too
intel-iommu: Performance improvement for dma_pte_free_pagetable()
intel-iommu: Don't free too much in dma_pte_free_pagetable()
intel-iommu: dump mappings but don't die on pte already set
intel-iommu: Combine domain_pfn_mapping() and domain_sg_mapping()
intel-iommu: Introduce domain_sg_mapping() to speed up intel_map_sg()
intel-iommu: Simplify __intel_alloc_iova()
intel-iommu: Performance improvement for domain_pfn_mapping()
intel-iommu: Performance improvement for dma_pte_clear_range()
intel-iommu: Clean up iommu_domain_identity_map()
intel-iommu: Remove last use of PHYSICAL_PAGE_MASK, for reserving PCI BARs
intel-iommu: Make iommu_flush_iotlb_psi() take pfn as argument
intel-iommu: Change aligned_size() to aligned_nrpages()
intel-iommu: Clean up intel_map_sg(), remove domain_page_mapping()
...
fix hang with HIGHMEM_64G and 32bit resource. According to hpa and
Linus, use (resource_size_t)-1 to fend off big ranges.
Analyzed by hpa
Reported-and-tested-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
These macros had two bugs:
- the type of the mask was not correctly expanded to the full size of
the argument being expanded, resulting in possible loss of high bits
when mixing types.
- the alignment argument was evaluated twice, despite the macro looking
like a fancy function (but it really does need to be a macro, since
it works on arbitrary integer types)
Noticed by Peter Anvin, and with a fix that is a modification of his
suggestion (bug noticed by Yinghai Lu).
Cc: Peter Anvin <hpa@zytor.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add basic performance counter support to the FRV arch.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Johannes Berg