| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Very long RCU read-side critical sections (50 milliseconds or
so) can cause a race between force_quiescent_state() and
rcu_start_gp() as follows on kernel builds with multi-level
rcu_node hierarchies:
1. CPU 0 calls force_quiescent_state(), sees that there is a
grace period in progress, and acquires ->fsqlock.
2. CPU 1 detects the end of the grace period, and so
cpu_quiet_msk_finish() sets rsp->completed to rsp->gpnum.
This operation is carried out under the root rnp->lock,
but CPU 0 has not yet acquired that lock. Note that
rsp->signaled is still RCU_SAVE_DYNTICK from the last
grace period.
3. CPU 1 calls rcu_start_gp(), but no one wants a new grace
period, so it drops the root rnp->lock and returns.
4. CPU 0 acquires the root rnp->lock and picks up rsp->completed
and rsp->signaled, then drops rnp->lock. It then enters the
RCU_SAVE_DYNTICK leg of the switch statement.
5. CPU 2 invokes call_rcu(), and now needs a new grace period.
It calls rcu_start_gp(), which acquires the root rnp->lock, sets
rsp->signaled to RCU_GP_INIT (too bad that CPU 0 is already in
the RCU_SAVE_DYNTICK leg of the switch statement!) and starts
initializing the rcu_node hierarchy. If there are multiple
levels to the hierarchy, it will drop the root rnp->lock and
initialize the lower levels of the hierarchy.
6. CPU 0 notes that rsp->completed has not changed, which permits
both CPU 2 and CPU 0 to try updating it concurrently. If CPU 0's
update prevails, later calls to force_quiescent_state() can
count old quiescent states against the new grace period, which
can in turn result in premature ending of grace periods.
Not good.
This patch adds an RCU_GP_IDLE state for rsp->signaled that is
set initially at boot time and any time a grace period ends.
This prevents CPU 0 from getting into the workings of
force_quiescent_state() in step 4. Additional locking and
checks prevent the concurrent update of rsp->signaled in step 6.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: mathieu.desnoyers@polymtl.ca
Cc: josh@joshtriplett.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
Cc: Valdis.Kletnieks@vt.edu
Cc: dhowells@redhat.com
LKML-Reference: <1256742889199-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
|
|
Ingo triggered the following warning:
WARNING: at lib/debugobjects.c:255 debug_print_object+0x42/0x50()
Hardware name: System Product Name
ODEBUG: init active object type: timer_list
Modules linked in:
Pid: 2619, comm: dmesg Tainted: G W 2.6.32-rc5-tip+ #5298
Call Trace:
[<81035443>] warn_slowpath_common+0x6a/0x81
[<8120e483>] ? debug_print_object+0x42/0x50
[<81035498>] warn_slowpath_fmt+0x29/0x2c
[<8120e483>] debug_print_object+0x42/0x50
[<8120ec2a>] __debug_object_init+0x279/0x2d7
[<8120ecb3>] debug_object_init+0x13/0x18
[<810409d2>] init_timer_key+0x17/0x6f
[<81041526>] free_uid+0x50/0x6c
[<8104ed2d>] put_cred_rcu+0x61/0x72
[<81067fac>] rcu_do_batch+0x70/0x121
debugobjects warns about an enqueued timer being initialized. If
CONFIG_USER_SCHED=y the user management code uses delayed work to
remove the user from the hash table and tear down the sysfs objects.
free_uid is called from RCU and initializes/schedules delayed work if
the usage count of the user_struct is 0. The init/schedule happens
outside of the uidhash_lock protected region which allows a concurrent
caller of find_user() to reference the about to be destroyed
user_struct w/o preventing the work from being scheduled. If the next
free_uid call happens before the work timer expired then the active
timer is initialized and the work scheduled again.
The race was introduced in commit 5cb350ba (sched: group scheduling,
sysfs tunables) and made more prominent by commit 3959214f (sched:
delayed cleanup of user_struct)
Move the init/schedule_delayed_work inside of the uidhash_lock
protected region to prevent the race.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Cc: Paul E. McKenney <paulmck@us.ibm.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: stable@kernel.org
|
|
I got a boot crash when forcing cpumasks offstack on 32 bit,
because find_new_ilb() returned 3 on my UP system (nohz.cpu_mask
wasn't zeroed).
AFAICT the others need to be zeroed too: only
nohz.ilb_grp_nohz_mask is initialized before use.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <200911022037.21282.rusty@rustcorp.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
|
|
Reboot does not work out of the box on my "Early 2009" Mac mini
(3,1). Detect this machine via DMI as we do for recent MacBooks.
Signed-off-by: Gottfried Haider <gottfried.haider@gmail.com>
Cc: Ozan Çağlayan <ozan@pardus.org.tr>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
|
|
The patch below also addresses a couple of other corner cases in readdir
seen with a large (e.g. 64k) msize. I'm not sure what people think of
my co-opting of fid->aux here. I'd be happy to rework if there's a better
way.
When the size of the user supplied buffer passed to readdir is smaller
than the data returned in one go by the 9P read request, v9fs_dir_readdir()
currently discards extra data so that, on the next call, a 9P read
request will be issued with offset < previous offset + bytes returned,
which voilates the constraint described in paragraph 3 of read(5) description.
This patch preseves the leftover data in fid->aux for use in the next call.
Signed-off-by: Jim Garlick <garlick@llnl.gov>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
|
|
I do not know if you've looked on the patch, but unfortunately it is
incorrect. A suggested better version is in this email (the old
version didn't work in case the user provided buffer was not long
enough - it incorrectly appended null byte on a position of last char,
and thus broke the contract of the readlink method). However, I'm
still not sure this is 100% correct thing to do, I think readlink is
supposed to return buffer without last null byte in all cases, but we
do return last null byte (even the old version).. on the other hand it
is likely unspecified what is in the remaining part of the buffer, so
null character may be fine there ;):
Signed-off-by: Martin Stava <martin.stava@gmail.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
|
|
Here is a proposed patch for bug in readdir. Listing of dirs with
many files fails without this patch.
Signed-off-by: Martin Stava <martin.stava@gmail.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
|
|
alc_automute_pin() might be called even if any HP pin is defined, and
it will result in verbs with NID=0.
This patch adds a check for the validity of HP widget before issuing
any verbs.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
Now up to 128 substreams are supported.
Reported-by: Adrian Bridgett <adrian@smop.co.uk>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
|
sys_ppoll syscall needs to use a compat handler on 64bit kernels with o32
user-space.
Signed-off-by: Arnaud Patard <apatard@mandriva.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Unify the naming method between kernel and the user-space oprofile tool.
Because loongson is used instead of godson in most of the places, we agreed
to use loongson instead, which will simplify future maintenance.
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Acked-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
The handle_edge_irq() flowhandler disables edge int sources which occur
too fast (i.e. another edge comes in before the irq handler function
had a chance to finish). Currently, the mask_ack() callback does not
ack the edges in hardware, leading to an endless loop in the flowhandler
where it tries to shut up the irq source.
When I rewrote the alchemy IRQ code I wrongly assumed the mask_ack()
callback was only used by the level flowhandler, hence it omitted the
(at the time pointless) edge acks. Turned out I was wrong; so here
is a complete mask_ack implementation for Alchemy IC, which fixes
the above mentioned problem.
Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
TXx9 SPI bit rate is calculated by:
fBR = fSPI / 2 / (n + 1)
(fSPI is SPI master clock freq, i.e. imbusclk freq.)
So use imbus_clk / 2 as a spi-baseclk.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
For consistency with other BCM63xx SoC set the CPU name to "Broadcom
BCM6338" when actually running on that system.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Loongson 2 does not have dcache aliases when is using 16k pages. and the
And because Loongson 2 doesn't do SMP , cpu_icache_snoops_remote_store does
not matter here.
Signed-off-by: Zhang Le <r0bertz@gentoo.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
set_saved_sp reads Context register. Avoid reading stale value from
earlier incomplete write.
Issue found and fixed for head.S by Chris Dearman <chris@mips.com>.
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Some newer Octeon chips have registers that allow lockless operation of
the interrupt controller. Take advantage of them.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Since the locks are used from interrupt context we need the
irqsave/irqrestore versions of the locking functions.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Chris Dearman (chris@mips.com)
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch ensures that the sign bit is always updated for NaN operands.
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Based on original patch by Chris Dearman <chris@mips.com>.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This fixes the remaining problems introduced by
f197465384bf7ef1af184c2ed1a4e268911a91e3 (incorrect access length &
byteswapping in bigendian mode)
Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
On an SMP system with cache aliases, the following sequence of events may
happen:
1) copy_user_highpage() runs on CPU0, invoking kmap_coherent() to create a
temporary mapping in the fixmap region
2) copy_page() starts on CPU0
3) CPU1 sends CPU0 an IPI asking CPU0 to run local_r4k_flush_cache_page()
4) CPU0 takes the interrupt, interrupting copy_page()
5) local_r4k_flush_cache_page() on CPU0 calls kmap_coherent() again
6) The second invocation of kmap_coherent() on CPU0 tries to use the
same fixmap virtual address that was being used by copy_user_highpage()
7) CPU0 throws a machine check exception for the TLB address conflict
Fixed by creating an extra set of fixmap entries for use in interrupt
handlers. This prevents fixmap VA conflicts between copy_user_highpage()
running in user context, and local_r4k_flush_cache_page() invoked from an
SMP IPI.
Signed-off-by: Kevin Cernekee <cernekee@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch checks if the watchdog enable bit is set in the DCL register
meaning that the hardware watchdog actually works and if so, register the
ar7_wdt platform_device.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch prepares the board code to register a bcm63xx_wdt
platform_device that we are going to use in a subsequent patch.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch removes the calls to bcm63xx_uart_register in board_bcm963xx.c
and make bcm63xx_uart_register an initfunc. Allows us to remove
bcm63xx_dev_uart.h which was there to make checkpatch.pl happy.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch fixes the following warning, which becomes an error due to
-Werror to be turned on:
CC arch/mips/alchemy/common/gpiolib-au1000.o
cc1: warnings being treated as errors
arch/mips/alchemy/common/gpiolib-au1000.c: In function 'au1100_gpio2_to_irq':
/home/florian/dev/kernel/linux-queue/arch/mips/include/asm/mach-au1x00/gpio-au1000.h:107: warning: control reaches end of non-void function
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
This patch fixes a lockup on BCM6345 where setting the PLL soft reset bit
will also lock the other blocks including UART. Instead of setting only
the PLL soft reset bit in the software reset register, set this bit but do
not touch the others.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Error was introduced by commit 0385d1f3d394c6814be0b165c153fc3fc254469a.
Signed-off-by: Ralf Roesch <ralf.roesch@rw-gmbh.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
As the commit 3ee4c147 shows, we need to "Add IRQF_TIMER flag for timer
interrupts", Atsushi Nemoto have reported that some other timer interrupts
should be considered, Here it is.
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
An o32 aplication passes a 64-bit value in a pair of registers; a 64-bit
kernel expects a 64-bit argument in a single register.
Signed-off-by: Chen Jie <chenj@lemote.com>
Signed-off-by: Hu Hongbing <huhb@lemote.com>
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Reported-by: Josip Rodin <joy@entuzijast.net>
|
|
Along the lines of d6c585a4342a2ff627a29f9aea77c5ed4cd76023, add IRQF_TIMER
flag for all timer interrupts This ensures that timer interrupts won't be
disabled on suspend and not threaded for PREEMPT_RT.
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
That code is executed with irq disabled already, so, remove the redundant
local_irq_disable() here.
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Broken by 182a85f8a119c789610a9d464f4129ded9f3c107.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
We've silently been relying on the hardware chopping off excess, reserved
ASID bits for no better reason that it saving an instruction. Because we
already have:
#define cpu_asid(cpu, mm) (cpu_context((cpu), (mm)) & ASID_MASK)
in <asm/mmu_context.h>.
We can use a cleanup to avoid writing non-zero bits into the reserved
entryhi bits. This avoid triggering some debugging assertion in the
Cavium simulator.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Some firmware may pass well over 256 bytes these days.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of
requiring manual settings of PCMCIA_DEBUG.
Also, remove all usages of the CS_CHECK macro and replace them with proper
Linux style calling and return value checking. The extra error reporting may
be dropped, as the PCMCIA core already complains about any (non-driver-author)
errors.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
|
