16 May, 2012
5 commits
-
Due to an errata, the PA7300LC generates a TLB miss interruption even on the
prefetch instruction. This means that prefetch(NULL), which is supposed to be
a nop on linux actually generates a NULL deref fault. Fix this by testing the
address of prefetch against NULL before doing the prefetch.Cc: stable@vger.kernel.org
Signed-off-by: James Bottomley -
As pointed out by serveral people, PA1.1 only has a type 26 instruction
meaning that the space register must be explicitly encoded. Not giving an
explicit space means that the compiler uses the type 24 version which is PA2.0
only resulting in an illegal instruction crash.This regression was caused by
commit f311847c2fcebd81912e2f0caf8a461dec28db41
Author: James Bottomley
Date: Wed Dec 22 10:22:11 2010 -0600parisc: flush pages through tmpalias space
Reported-by: Helge Deller
Signed-off-by: John David Anglin
Cc: stable@vger.kernel.org #2.6.39+
Signed-off-by: James Bottomley -
All PA1.1 systems have been oopsing on boot since
commit f311847c2fcebd81912e2f0caf8a461dec28db41
Author: James Bottomley
Date: Wed Dec 22 10:22:11 2010 -0600parisc: flush pages through tmpalias space
because a PA2.0 instruction was accidentally introduced into the PA1.1 TLB
insertion interruption path when it was consolidated with the do_alias macro.
Fix the do_alias macro only to use PA2.0 instructions if compiled for 64 bit.
Cc: stable@vger.kernel.org #2.6.39+
Signed-off-by: James Bottomley -
The scheduler depends on receiving the CPU_STARTING notification, without
which we end up into a lot of trouble. So add the missing call to
notify_cpu_starting() in the bringup code.Signed-off-by: Srivatsa S. Bhat
Signed-off-by: Linus Torvalds -
The scheduler depends on receiving the CPU_STARTING notification, without
which we end up into a lot of trouble. So add the missing call to
notify_cpu_starting() in the bringup code.Signed-off-by: Srivatsa S. Bhat
Acked-and-Tested-by: Mikulas Patocka
Acked-and-Tested-by: Tobias Ulmer
Tested-by: John David Anglin
Signed-off-by: Linus Torvalds
15 May, 2012
4 commits
-
Export handle_edge_irq() and irq_to_desc() to modules to allow them to
do things such as__irq_set_handler_locked(...., handle_edge_irq);
This fixes
ERROR: "handle_edge_irq" [drivers/gpio/gpio-pch.ko] undefined!
ERROR: "irq_to_desc" [drivers/gpio/gpio-pch.ko] undefined!when gpio-pch is being built as a module.
This was introduced by commit df9541a60af0 ("gpio: pch9: Use proper flow
type handlers") that added__irq_set_handler_locked(d->irq, handle_edge_irq);
but handle_edge_irq() was not exported for modules (and inlined
__irq_set_handler_locked() requires irq_to_desc() exported as well)Signed-off-by: Jiri Kosina
Signed-off-by: Linus Torvalds -
Pull media fixes from Mauro Carvalho Chehab:
"For a some fix patches for v3.4, including a regression fix at DVB core"Fix up trivial conflicts in Documentation/feature-removal-schedule.txt
* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
[media] gspca - sonixj: Fix a zero divide in isoc interrupt
[media] media: videobuf2-dma-contig: include header for exported symbols
[media] media: videobuf2-dma-contig: quiet sparse noise about plain integer as NULL pointer
[media] media: vb2-memops: Export vb2_get_vma symbol
[media] s5p-fimc: Correct memory allocation for VIDIOC_CREATE_BUFS
[media] s5p-fimc: Fix locking in subdev set_crop op
[media] dvb_frontend: fix a regression with DVB-S zig-zag
[media] fintek-cir: change || to &&
[media] V4L: Schedule V4L2_CID_HCENTER, V4L2_CID_VCENTER controls for removal
[media] rc: Postpone ISR registration
[media] marvell-cam: fix an ARM build error
[media] V4L: soc-camera: protect hosts during probing from overzealous user-space -
Pull networking fixes from David Miller:
"The main purpose of this pull request is to fix up the erroneous
bonding patch I applied last round. I meant to apply v4 of the patch
from Jiri but I applied v3 by accident. Mea culpa.Also, eagle eyed Dan Carpenter noticed that openvswitch has one of
those "X = alloc(); if (!Y)" mistakes, test the proper pointer
instead."* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
openvswitch: checking wrong variable in queue_userspace_packet()
bonding: Fix LACPDU rx_dropped commit. -
3.4-rc introduced a regression when setting the LEDS. We do the right thing
but then return an error code.Resolves-bug: https://bugzilla.kernel.org/show_bug.cgi?id=43144
Reported-by: Christian Casteyde
Signed-off-by: Alan Cox
Signed-off-by: Linus Torvalds
14 May, 2012
3 commits
-
"skb" is non-NULL here, for example we dereference it in skb_clone().
The intent was to test "nskb" which was just set.Signed-off-by: Dan Carpenter
Acked-by: Jesse Gross
Signed-off-by: David S. Miller -
I applied the wrong version of Jiri's bonding fix in commit
13a8e0c8cdb43982372bd6c65fb26839c8fd8ce9 ("bonding: don't increase
rx_dropped after processing LACPDUs")I applied v3, which introduces warnings I asked him to fix,
instead of v4 which properly takes care of those issues.This inter-diffs such that the warnings are now gone.
Signed-off-by: David S. Miller
-
Pull three MTD fixes from David Woodhouse:
- Fix a lock ordering deadlock in JFFS2
- Fix an oops in the dataflash driver, triggered by a dummy call to test
whether it has OTP functionality.
- Fix request_mem_region() failure on amsdelta NAND driver.* tag 'for-linus-3.4-20120513' of git://git.infradead.org/linux-mtd:
mtd: ams-delta: fix request_mem_region() failure
jffs2: Fix lock acquisition order bug in gc path
mtd: fix oops in dataflash driver
13 May, 2012
16 commits
-
.. and this should hopefully be the last -rc before final 3.4 release.
-
Pull ARM: SoC fixes from Olof Johansson:
"I was hoping to be done with fixes for 3.4 but we got two branches
from subarch maintainers the last couple of days. So here is one
last(?) pull request for arm-soc containing 7 patches:- Five of them are for shmobile dealing with SMP setup and compile
failures
- The remaining two are for regressions on the Samsung platforms"* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
ARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1
ARM: EXYNOS: use s5p-timer for UniversalC210 board
ARM / mach-shmobile: Invalidate caches when booting secondary cores
ARM / mach-shmobile: sh73a0 SMP TWD boot regression fix
ARM / mach-shmobile: r8a7779 SMP TWD boot regression fix
ARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper
ARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper -
Pull a few more GPIO bug fixes from Grant Likely:
"Oops, missed a couple. Here's an updated pull req for GPIO"A set of PCH bug fixes, and one patch to fix up compile warnings
* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
gpio/exynos: Fix compiler warnings when non-exynos machines are selected
gpio: pch9: Use proper flow type handlers -
…/git/kgene/linux-samsung into fixes
* 'v3.4-samsung-fixes-5' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
ARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1
ARM: EXYNOS: use s5p-timer for UniversalC210 board -
It should be (1 << 2) for ctrlbit of exynos5_clk_pdma1.
Signed-off-by: Kukjin Kim
-
Commit 069d4e743 ("ARM: EXYNOS4: Remove clock event timers using
ARM private timers") removed support for local timers and forced
to use MCT as event source. However MCT is not operating properly
on early revision of EXYNOS4 SoCs. All UniversalC210 boards are
based on it, so that commit broke support for it. This patch
provides a workaround that enables UniversalC210 boards to boot
again. s5p-timer is used as an event source, it works only for
non-SMP builds.Signed-off-by: Marek Szyprowski
Signed-off-by: Kyungmin Park
Signed-off-by: Kukjin Kim -
By Guennadi Liakhovetski (2) and others via Rafael J. Wysocki:
"[...] urgent fixes for Renesas ARM-based platforms. Four of these
commits are fixes of regressions new in 3.4-rc and the last one is
necessary for SMP to work on those systems in general."* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas:
ARM / mach-shmobile: Invalidate caches when booting secondary cores
ARM / mach-shmobile: sh73a0 SMP TWD boot regression fix
ARM / mach-shmobile: r8a7779 SMP TWD boot regression fix
ARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper
ARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper -
Make sure L1 caches are invalidated when booting secondary
cores. Needed to boot all mach-shmobile SMP systems that
are using Cortex-A9 including sh73a0, r8a7779 and EMEV2.Thanks to imx and tegra guys for actual code.
Signed-off-by: Magnus Damm
Tested-by: Kuninori Morimoto
Signed-off-by: Rafael J. Wysocki -
Fix SMP TWD boot regression on sh73a0 based platforms caused by:
4200b16 ARM: shmobile: convert to twd_local_timer_register() interface
After the merge of the above commit it has been impossible to boot
sh73a0 based SoCs with SMP enabled and CONFIG_HAVE_ARM_TWD=y. The
kernel crashes at smp_init_cpus() timing which is before the console
has been initialized, so to the user this looks like a kernel lock up
without any particular error message.This patch fixes the regression on sh73a0 by moving the TWD
registration code from smp_init_cpus() to sys_timer->init() time.This patch removed shmobile_twd_init() which is no longer needed
Signed-off-by: Kuninori Morimoto
Signed-off-by: Magnus Damm
Signed-off-by: Rafael J. Wysocki -
Fix SMP TWD boot regression on r8a7779 based platforms caused by:
4200b16 ARM: shmobile: convert to twd_local_timer_register() interface
After the merge of the above commit it has been impossible to boot
r8a7779 based SoCs with SMP enabled and CONFIG_HAVE_ARM_TWD=y. The
kernel crashes at smp_init_cpus() timing which is before the console
has been initialized, so to the user this looks like a kernel lock up
without any particular error message.This patch fixes the regression on r8a7779 by moving the TWD
registration code from smp_init_cpus() to sys_timer->init() time.Signed-off-by: Magnus Damm
Acked-by: Marc Zyngier
Signed-off-by: Rafael J. Wysocki -
This also fixes the following modular mmc build failure:
arch/arm/mach-shmobile/built-in.o: In function `mackerel_sdhi0_gpio_cd':
pfc-sh7372.c:(.text+0x1138): undefined reference to `mmc_detect_change'on this platform by eliminating the use of an inline function, which
calls into the mmc core.Signed-off-by: Guennadi Liakhovetski
Reviewed-by: Simon Horman
Acked-by: Magnus Damm
Signed-off-by: Rafael J. Wysocki -
This also fixes the following modular mmc build failure:
arch/arm/mach-shmobile/built-in.o: In function `ag5evm_sdhi0_gpio_cd':
pfc-sh73a0.c:(.text+0x7c0): undefined reference to `mmc_detect_change'on this platform by eliminating the use of an inline function, which
calls into the mmc core.Signed-off-by: Guennadi Liakhovetski
Tested-by: Simon Horman
Acked-by: Magnus Damm
Signed-off-by: Rafael J. Wysocki -
Pull SCSI fixes from James Bottomley:
"This is a set of minor qla and virto fixes plus one major regression
fix (oops in all legacy host drivers)."* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
[SCSI] virtio_scsi: fix TMF use-after-free
[SCSI] fix oops in all legacy host adapters caused by 6f381fa
[SCSI] qla2xxx: Update version number to 8.04.00.03-k.
[SCSI] qla2xxx: Properly check for current state after the fabric-login request.
[SCSI] qla2xxx: Proper completion to scsi-ml for scsi status task_set_full and busy.
[SCSI] qla2xxx: Block flash access from application when device is initialized for ISP82xx.
[SCSI] qla2xxx: Fix reset time out as qla2xxx not ack to reset request. -
Pull networking fixes from David S. Miller:
1) Since we do RCU lookups on ipv4 FIB entries, we have to test if the
entry is dead before returning it to our caller.2) openvswitch locking and packet validation fixes from Ansis Atteka,
Jesse Gross, and Pravin B Shelar.3) Fix PM resume locking in IGB driver, from Benjamin Poirier.
4) Fix VLAN header handling in vhost-net and macvtap, from Basil Gor.
5) Revert a bogus network namespace isolation change that was causing
regressions on S390 networking devices.6) If bonding decides to process and handle a LACPDU frame, we
shouldn't bump the rx_dropped counter. From Jiri Bohac.7) Fix mis-calculation of available TX space in r8169 driver when doing
TSO, which can lead to crashes and/or hung device. From Julien
Ducourthial.8) SCTP does not validate cached routes properly in all cases, from
Nicolas Dichtel.9) Link status interrupt needs to be handled in ks8851 driver, from
Stephen Boyd.10) Use capable(), not cap_raised(), in connector/userns netlink code.
From Eric W. Biederman via Andrew Morton.11) Fix pktgen OOPS on module unload, from Eric Dumazet.
12) iwlwifi under-estimates SKB truesizes, also from Eric Dumazet.
13) Cure division by zero in SFC driver, from Ben Hutchings.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
ks8851: Update link status during link change interrupt
macvtap: restore vlan header on user read
vhost-net: fix handle_rx buffer size
bonding: don't increase rx_dropped after processing LACPDUs
connector/userns: replace netlink uses of cap_raised() with capable()
sctp: check cached dst before using it
pktgen: fix crash at module unload
Revert "net: maintain namespace isolation between vlan and real device"
ehea: fix losing of NEQ events when one event occurred early
igb: fix rtnl race in PM resume path
ipv4: Do not use dead fib_info entries.
r8169: fix unsigned int wraparound with TSO
sfc: Fix division by zero when using one RX channel and no SR-IOV
openvswitch: Validation of IPv6 set port action uses IPv4 header
net: compare_ether_addr[_64bits]() has no ordering
cdc_ether: Ignore bogus union descriptor for RNDIS devices
bnx2x: bug fix when loading after SAN boot
e1000: Silence sparse warnings by correcting type
igb, ixgbe: netdev_tx_reset_queue incorrectly called from tx init path
openvswitch: Release rtnl_lock if ovs_vport_cmd_build_info() failed.
... -
Pull device-mapper fixes from Alasdair G Kergon:
"Fix a couple of serious memory leaks in device-mapper thin
provisioning and tidy its MODULE_DESCRIPTION.Mitigate occasional reported hangs associated with multipath scsi_dh
module loading."* tag 'dm-3.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm:
dm mpath: check if scsi_dh module already loaded before trying to load
dm thin: correct module description
dm thin: fix unprotected use of prepared_discards list
dm thin: reinstate missing mempool_free in cell_release_singleton -
Since cpufreq has no official maintainer at the moment, I'm willing
to maintain it along some other power management core code I've been
maintaining already.Signed-off-by: Rafael J. Wysocki
Signed-off-by: Linus Torvalds
12 May, 2012
12 commits
-
If the requested scsi_dh module is already loaded then skip
request_module().Multipath table loads can hang in an unnecessary __request_module.
Reported-by: Ben Marzinski
Cc: stable@kernel.org
Signed-off-by: Mike Snitzer
Signed-off-by: Alasdair G Kergon -
Remove duplicate copy of string "device-mapper" (DM_NAME) from
MODULE_DESCRIPTION.Signed-off-by: Alasdair G Kergon
-
Fix two places in commit 104655fd4dce ("dm thin: support discards") that
didn't use pool->lock to protect against concurrent changes to the
prepared_discards list.Without this fix, thin_endio() can race with process_discard(), leading
to concurrent list_add()s that result in the processes locking up with
an error like the following:WARNING: at lib/list_debug.c:32 __list_add+0x8f/0xa0()
...
list_add corruption. next->prev should be prev (ffff880323b96140), but was ffff8801d2c48440. (next=ffff8801d2c485c0).
...
Pid: 17205, comm: kworker/u:1 Tainted: G W O 3.4.0-rc3.snitm+ #1
Call Trace:
[] warn_slowpath_common+0x7f/0xc0
[] warn_slowpath_fmt+0x46/0x50
[] ? bio_detain+0xc6/0x210 [dm_thin_pool]
[] __list_add+0x8f/0xa0
[] process_discard+0x2a2/0x2d0 [dm_thin_pool]
[] ? remap_and_issue+0x38/0x50 [dm_thin_pool]
[] process_deferred_bios+0x7b/0x230 [dm_thin_pool]
[] ? process_deferred_bios+0x230/0x230 [dm_thin_pool]
[] do_worker+0x52/0x60 [dm_thin_pool]
[] process_one_work+0x129/0x450
[] worker_thread+0x17c/0x3c0
[] ? manage_workers+0x120/0x120
[] kthread+0x9e/0xb0
[] kernel_thread_helper+0x4/0x10
[] ? kthread_freezable_should_stop+0x70/0x70
[] ? gs_change+0x13/0x13
---[ end trace 7e0a523bc5e52692 ]---Signed-off-by: Mike Snitzer
Signed-off-by: Alasdair G Kergon -
Fix a significant memory leak inadvertently introduced during
simplification of cell_release_singleton() in commit
6f94a4c45a6f744383f9f695dde019998db3df55 ("dm thin: fix stacked bi_next
usage").A cell's hlist_del() must be accompanied by a mempool_free().
Use __cell_release() to do this, like before.Signed-off-by: Mike Snitzer
Signed-off-by: Alasdair G Kergon -
Fixes the following compiler warnings:
drivers/gpio/gpio-samsung.c: In function ‘samsung_gpiolib_init’:
drivers/gpio/gpio-samsung.c:2980:1: warning: label ‘err_ioremap1’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2978:1: warning: label ‘err_ioremap2’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2976:1: warning: label ‘err_ioremap3’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2974:1: warning: label ‘err_ioremap4’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2722:55: warning: unused variable ‘gpio_base4’ [-Wunused-variable]drivers/gpio/gpio-samsung.c:455:32: warning: ‘exynos_gpio_cfg’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2126:33: warning: ‘exynos4_gpios_1’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2228:33: warning: ‘exynos4_gpios_2’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2373:33: warning: ‘exynos4_gpios_3’ defined but not used [-Wunused-variable]Signed-off-by: Sachin Kamat
Signed-off-by: Grant Likely -
Jean-Francois Dagenais reported:
Configuring a gpio pin with the gpio-pch driver with
"IRQF_TRIGGER_LOW | IRQF_ONESHOT" generates an interrupt storm for
threaded ISR until the ISR thread actually gets to physically clear
the interrupt on the triggering chip!! The immediate observable
symptom is the high CPU usage for my ISR thread task and the
interrupt count in /proc/interrupts incrementing radically.The driver is wrong in several ways:
1) Using handle_simple_irq() does not provide proper flow control
handling. In the case of oneshot threaded handlers for the
demultiplexed interrupts this results in an interrupt storm because
the simple handler does not deal with masking/unmasking. Even
without threaded oneshot handlers an interrupt storm for level type
interrupts can easily be triggered when the interrupt is disabled
and the interrupt line is activated from the device.2) Acknowlegding the demultiplexed interrupt before calling the
handler is wrong for level type interrupts.3) The set_type function unconditionally enables the interrupt. It's
supposed to set the type and nothing else. The unmasking is done by
the core code.Move the acknowledge code into a separate function and add it to the
demux irqchip callbacks.Remove the unconditional enabling from the set_type() callback and set
the proper flow handlers depending on the selected type (level/edge).Reported-and-tested-by: Jean-Francois Dagenais
Signed-off-by: Thomas Gleixner
Signed-off-by: Grant Likely -
Pull GPIO omap bug fix from Grant Likely.
* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
gpio/omap: fix incorrect initialization of omap_gpio_mod_init -
Pull another powerpc irq fix from Benjamin Herrenschmidt:
"It looks like my previous fix for the lazy irq masking problem wasn't
quite enough. There was another problem related to performance
monitor interrupts acting as NMIs leaving the flags in an incorrect
state. Here's a fix that finally seems to make perf solid again."* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc/irq: Fix another case of lazy IRQ state getting out of sync -
Pull target fix from Nicholas Bellinger:
"This patch removes some incorrect legacy code to free se_lun_acl
memory in the NodeACL release path that could potentially trigger an
OOPS during shutdown once dynamic -> explicit initiator NodeACL
conversion has occurred.That said, we've been able to trigger an OOPS in v4.0 code for this
special case when the associated MappedLUNs had not also been made
explicit based on active TPG LUN layout during the conversion, so it
really makes senses to go ahead and drop this extra cruft to avoid any
possible issues here.This ends up only effecting iscsi-target module code (it's the only
user) and is CC'ed to stable."* '3.4-urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
target: Drop incorrect se_lun_acl release for dynamic -> explict ACL conversion -
So we have another case of paca->irq_happened getting out of
sync with the HW irq state. This can happen when a perfmon
interrupt occurs while soft disabled, as it will return to a
soft disabled but hard enabled context while leaving a stale
PACA_IRQ_HARD_DIS flag set.This patch fixes it, and also adds a test for the condition
of those flags being out of sync in arch_local_irq_restore()
when CONFIG_TRACE_IRQFLAGS is enabled.This helps catching those gremlins faster (and so far I
can't seem see any anymore, so that's good news).Signed-off-by: Benjamin Herrenschmidt
-
If a link change interrupt comes in we just clear the interrupt
and continue along without notifying the upper networking layers
that the link has changed. Use the mii_check_link() function to
update the link status whenever a link change interrupt occurs.Cc: Ben Dooks
Signed-off-by: Stephen Boyd
Signed-off-by: David S. Miller -
Ethernet vlan header is not on the packet and kept in the skb->vlan_tci
when it comes from lower dev. This patch inserts vlan header in user
buffer during skb copy on user read.Signed-off-by: Basil Gor
Acked-by: Michael S. Tsirkin
Signed-off-by: David S. Miller
