| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
In https://bugzilla.kernel.org/show_bug.cgi?id=42976, a system with driver
rtl8192se used as an AP suffers from "Out of SW-IOMMU space" errors. These
are caused by the DMA buffers used for beacons never being unmapped.
This bug was also reported at
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/961618
Reported-and-Tested-by: Da Xue <[email protected]>
Signed-off-by: Larry Finger <[email protected]>
Cc: Stable <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
|
|
The current version of rtlwifi for USB operations uses kmalloc to
acquire a 32-bit buffer for each read of the device. When
_usb_read_sync() is called with the rcu_lock held, the result is
a "sleeping function called from invalid context" BUG. This is
reported for two cases in https://bugzilla.kernel.org/show_bug.cgi?id=42775.
The first case has the lock originating from within rtlwifi and could
be fixed by rearranging the locking; however, the second originates from
within mac80211. The kmalloc() call is removed from _usb_read_sync()
by creating a ring buffer pointer in the private area and
allocating the buffer data in the probe routine.
Signed-off-by: Larry Finger <[email protected]>
Cc: Stable <[email protected]> [This version good for 3.3+ - different patch for 3.2 - 2.6.39]
Signed-off-by: John W. Linville <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile
Pull arch/tile fixes from Chris Metcalf:
"This is one important change from Srivatsa Bhat that got dropped when
I put together my pull request for -rc2, plus a trivial change to
remove a compiler warning."
* 'stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
arch/tile: avoid unused variable warning in proc.c for tilegx
tile/CPU hotplug: Add missing call to notify_cpu_starting()
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
- A series of fixes for Conexant 20549 HD-audio codec chip
- A workaround for HDMI hotplug debug prints that annoyed people
- A fix for the new support of platform DAPM contexts
- Many driver-specific minor fixes
* tag 'sound-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: hda - hide HDMI/ELD printks unless snd.debug=2
ALSA: sound/isa/sscape.c: add missing resource-release code
sound: sound/oss/msnd_pinnacle.c: add vfrees
ALSA: hda - clean up CX20549 test mixer setup
ALSA: hda - CX20549 doesn't need pin_amp_workaround.
ALSA: hda - Remove CD control from model=benq for CX20549
ALSA: hda - fix record volume controls of CX20459 ("Venice")
ALSA: hda - Rename capture sources of CX20549 to match common conventions
ALSA: hda - Fix proc output for ADC amp values of CX20549
ASoC: tegra: fix i2s compilation when !CONFIG_DEBUG_FS
ASoC: set idle_bias_off=1 for all platform DAPM contexts
ASoC: imx-audmux: Check for NULL pointer
ASoC: imx-audmux: Fix ssi port numbers in sysfs
ASoC: ak4642: fixup: mute needs +1 step
MAINTAINERS: Don't list everyone working on Wolfson drivers
MAINTAINERS: Add missing ASoC OMAP co-maintainer
ASoC: pxa: pxa2xx-i2s: add io.h for IOMEM macro
ASoC: tegra: ensure clocks are enabled when touching registers
ASoC: sgtl5000: Enable VAG when DAC/ADC up
ALSA: asihpi - fix return value of hpios_locked_mem_alloc()
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
Pull media fixes from Mauro Carvalho Chehab:
- dvb core: there is a regression found when used with xine. For
whatever unknown reason, xine (and xine-lib clients) wants that the
frontend to tell what frequency he is using even before the PLL lock
(or at least, it expects a non-zero frequency).
On DVB, the frequency is only actually known after a frequency
zig-zag seek, done by the DVB core. Anyway, the fix was trivial.
That solves Fedora BZ#808871.
- ivtv: fix a regression when selecting the language channel
- uvc: fix a race-related crash
- it913x: fixes firmware loading
- two trivial patches (a dependency issue at a radio driver at sound
Kconfig, and a warning fix on dvb).
* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
[media] uvcvideo: Fix race-related crash in uvc_video_clock_update()
[media] Drivers/media/radio: Fix build error
[media] dvb_frontend: fix compiler warning
[media] it913x: fix firmware loading errors
[media] ivtv: Fix AUDIO_(BILINGUAL_)CHANNEL_SELECT regression
[media] dvb_frontend: regression fix: userspace ABI broken for xine
|
|
Pull GFS2 fixes from Steven Whitehouse
* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes:
GFS2: Allow caching of rindex glock
GFS2: Make sure rindex is uptodate before starting transactions
GFS2: use depends instead of select in kconfig
GFS2: put glock reference in error patch of read_rindex_entry
|
|
Until we push the unaligned access support for tilegx, it's silly
to have arch/tile/kernel/proc.c generate a warning about an unused
variable. Extend the #ifdef to cover all the code and data for now.
Signed-off-by: Chris Metcalf <[email protected]>
|
|
This merges the 32- and 64-bit versions of the x86 strncpy_from_user()
by just rewriting it in C rather than the ancient inline asm versions
that used lodsb/stosb and had been duplicated for (trivial) differences
between the 32-bit and 64-bit versions.
While doing that, it also speeds them up by doing the accesses a word at
a time. Finally, the new routines also properly handle the case of
hitting the end of the address space, which we have never done correctly
before (fs/namei.c has a hack around it for that reason).
Despite all these improvements, it actually removes more lines than it
adds, due to the de-duplication. Also, we no longer export (or define)
the legacy __strncpy_from_user() function (that was defined to not do
the user permission checks), since it's not actually used anywhere, and
the user address space checks are built in to the new code.
Other architecture maintainers have been notified that the old hack in
fs/namei.c will be going away in the 3.5 merge window, in case they
copied the x86 approach of being a bit cavalier about the end of the
address space.
Cc: [email protected]
Cc: Ingo Molnar <[email protected]>
Cc: Peter Anvin" <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Re-define XHCI_LEGACY_DISABLE_SMI and used it in right way. All SMI enable
bits will be cleared to zero and flag bits 29:31 are also cleared to zero.
Other bits should be presvered as Table 146.
This patch should be backported to kernels as old as 2.6.31.
Signed-off-by: Alex He <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
Cc: [email protected]
|
|
The caller is allowed to specify the GFP flags for these functions.
We should prefer their flags unless we have good reason. For
example, if we take a spin_lock ourselves we'd need to use
GFP_ATOMIC. But in this case it's safe to use the callers GFP
flags.
The callers all pass GFP_ATOMIC here, so this change doesn't affect
how the kernel behaves but we may add other callers later and this
is a cleanup.
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
|
|
The suspend operation of VIA xHCI host have some issues and
hibernate operation works fine, so The XHCI_RESET_ON_RESUME
quirk is added for it.
This patch should base on "xHCI: Don't write zeroed pointer
to xHC registers" that is released by Sarah. Otherwise, the
host system error will ocurr in the hibernate operation
process.
This should be backported to stable kernels as old as 2.6.37,
that contain the commit c877b3b2ad5cb9d4fe523c5496185cc328ff3ae9
"xhci: Add reset on resume quirk for asrock p67 host".
Signed-off-by: Elric Fu <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
Cc: [email protected]
|
|
When the Seagate Goflex USB3.0 device is attached to VIA xHCI
host, sometimes the device will downgrade mode to high speed.
By the USB analyzer, I found the device finished the link
training process and worked at superspeed mode. But the device
descriptor got from the device shows the device works at 2.1.
It is very strange and seems like the device controller of
Seagate Goflex has a little confusion.
The first 8 bytes of device descriptor should be:
12 01 00 03 00 00 00 09
But the first 8 bytes of wrong device descriptor are:
12 01 10 02 00 00 00 40
The wrong device descriptor caused the initialization of mass
storage failed. After a while, the device would be recognized
as a high speed device and works fine.
This patch will warm reset the device to fix the issue after
finding the bcdUSB field of device descriptor isn't 0x0300
but the speed mode of device is superspeed.
This patch should be backported to kernels as old as 3.2, or ones that
contain the commit 75d7cf72ab9fa01dc70877aa5c68e8ef477229dc "usbcore:
refine warm reset logic".
Signed-off-by: Elric Fu <[email protected]>
Acked-by: Andiry Xu <[email protected]>
Acked-by: Sergei Shtylyov <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
Cc: [email protected]
|
|
The xHCI 1.0 spec errata released on June 13, 2011, changes the ordering
that the xHCI registers are saved and restored in. It moves the
interrupt pending (IMAN) and interrupt control (IMOD) registers to be
saved and restored last. I believe that's because the host controller
may attempt to fetch the event ring table when interrupts are
re-enabled. Therefore we need to restore the event ring registers
before we re-enable interrupts.
This should be backported to kernels as old as 2.6.37, that contain the
commit 5535b1d5f8885695c6ded783c692e3c0d0eda8ca "USB: xHCI: PCI power
management implementation"
Signed-off-by: Sarah Sharp <[email protected]>
Tested-by: Elric Fu <[email protected]>
Cc: Andiry Xu <[email protected]>
Cc: [email protected]
|
|
The xhci_save_registers() function saved the event ring dequeue pointer
in the s3 register structure, but xhci_restore_registers() never
restored it. No other code in the xHCI successful resume path would
ever restore it either. Fix that.
This should be backported to kernels as old as 2.6.37, that contain the
commit 5535b1d5f8885695c6ded783c692e3c0d0eda8ca "USB: xHCI: PCI power
management implementation".
Signed-off-by: Sarah Sharp <[email protected]>
Tested-by: Elric Fu <[email protected]>
Cc: Andiry Xu <[email protected]>
Cc: [email protected]
|
|
When xhci_mem_cleanup() is called, we can't be sure if the xHC is
actually halted. We can ask the xHC to halt by writing to the RUN bit
in the command register, but that might timeout due to a HW hang.
If the host controller is still running, we should not write zeroed
values to the event ring dequeue pointers or base tables, the DCBAA
pointers, or the command ring pointers. Eric Fu reports his VIA VL800
host accesses the event ring pointers after a failed register restore on
resume from suspend. The hypothesis is that the host never actually
halted before the register write to change the event ring pointer to
zero.
Remove all writes of zeroed values to pointer registers in
xhci_mem_cleanup(). Instead, make all callers of the function reset the
host controller first, which will reset those registers to zero.
xhci_mem_init() is the only caller that doesn't first halt and reset the
host controller before calling xhci_mem_cleanup().
This should be backported to kernels as old as 2.6.32.
Signed-off-by: Sarah Sharp <[email protected]>
Tested-by: Elric Fu <[email protected]>
Cc: [email protected]
|
|
Eric Fu reports a problem with his VIA host controller fetching a zeroed
event ring pointer on resume from suspend. The host should have been
halted, but we can't be sure because that code ignores the return value
from xhci_halt(). Print a warning when the host controller refuses to
halt within XHCI_MAX_HALT_USEC (currently 16 seconds).
(Update: it turns out that the VIA host controller is reporting a halted
state when it fetches the zeroed event ring pointer. However, we still
need this warning for other host controllers.)
Signed-off-by: Sarah Sharp <[email protected]>
|
|
While we're at that, define IMAN bitfield to aid readability.
The interrupt enable bit should be set once on driver init, and we
shouldn't need to continually re-enable it. Commit c21599a3 introduced
a read of the irq_pending register, and that allows us to preserve the
state of the IE bit. Before that commit, we were blindly writing 0x3 to
the register.
This patch should be backported to kernels as old as 2.6.36, or ones
that contain the commit c21599a36165dbc78b380846b254017a548b9de5 "USB:
xhci: Reduce reads and writes of interrupter registers".
Signed-off-by: Felipe Balbi <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
Cc: [email protected]
|
|
keyctl_session_to_parent(task) sets ->replacement_session_keyring,
it should be processed and cleared by key_replace_session_keyring().
However, this task can fork before it notices TIF_NOTIFY_RESUME and
the new child gets the bogus ->replacement_session_keyring copied by
dup_task_struct(). This is obviously wrong and, if nothing else, this
leads to put_cred(already_freed_cred).
change copy_creds() to clear this member. If copy_process() fails
before this point the wrong ->replacement_session_keyring doesn't
matter, exit_creds() won't be called.
Cc: <[email protected]>
Signed-off-by: Oleg Nesterov <[email protected]>
Acked-by: David Howells <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Running 'perf kvm --host --guest --guestmount /tmp/guestmount record -a -g -- sleep 2'
Was resulting in a segfault. For event type PERF_RECORD_MMAP,
event->ip.pid is being used in perf_session__find_machine_for_cpumode,
which is not correct.
The event->ip.pid field happens to be 0 in this case and results in
returning a NULL machine object. Finally, access to self->pid in
machine__mmap_name, results in a segfault later.
For PERF_RECORD_MMAP type, pass event->mmap.pid.
Signed-off-by: Nikunj A. Dadhania <[email protected]>
Reviewed-by: David Ahern <[email protected]>
Tested-by: David Ahern <[email protected]>
Cc: David Ahern <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Stephane Eranian <[email protected]>
Cc: Nikunj A. Dadhania <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Marc Merlin reported many order-1 allocations failures in TX path on its
wireless setup, that dont make any sense with MTU=1500 network, and non
SG capable hardware.
After investigation, it turns out TCP uses sk_stream_alloc_skb() and
used as a convention skb_tailroom(skb) to know how many bytes of data
payload could be put in this skb (for non SG capable devices)
Note : these skb used kmalloc-4096 (MTU=1500 + MAX_HEADER +
sizeof(struct skb_shared_info) being above 2048)
Later, mac80211 layer need to add some bytes at the tail of skb
(IEEE80211_ENCRYPT_TAILROOM = 18 bytes) and since no more tailroom is
available has to call pskb_expand_head() and request order-1
allocations.
This patch changes sk_stream_alloc_skb() so that only
sk->sk_prot->max_header bytes of headroom are reserved, and use a new
skb field, avail_size to hold the data payload limit.
This way, order-0 allocations done by TCP stack can leave more than 2 KB
of tailroom and no more allocation is performed in mac80211 layer (or
any layer needing some tailroom)
avail_size is unioned with mark/dropcount, since mark will be set later
in IP stack for output packets. Therefore, skb size is unchanged.
Reported-by: Marc MERLIN <[email protected]>
Tested-by: Marc MERLIN <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Marc Merlin reported many order-1 allocations failures in TX path on its
wireless setup, that dont make any sense with MTU=1500 network, and non
SG capable hardware.
Turns out part of the problem comes from pskb_expand_head() not using
ksize() to get exact head size given by kmalloc(). Doing the same thing
than __alloc_skb() allows more tailroom in skb and can prevent future
reallocations.
As a bonus, struct skb_shared_info becomes cache line aligned.
Reported-by: Marc MERLIN <[email protected]>
Tested-by: Marc MERLIN <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
As it stands the bridge IGMP snooping system will respond to
group leave messages with queries for remaining membership.
This is both unnecessary and undesirable. First of all any
multicast routers present should be doing this rather than us.
What's more the queries that we send may end up upsetting other
multicast snooping swithces in the system that are buggy.
In fact, we can simply remove the code that send these queries
because the existing membership expiry mechanism doesn't rely
on them anyway.
So this patch simply removes all code associated with group
queries in response to group leave messages.
Signed-off-by: Herbert Xu <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Since there are still many Acer models that might not be covered by
the current fixup table, let's add back a few typical model names so
that user can test the fixup without recompiling.
Signed-off-by: Takashi Iwai <[email protected]>
|
|
According to an internal workaround master list, we need to set bit 5
of register 9400 to avoid issues with color blits.
Testing shows that this seems to fix the blitter hangs when fbc is
enabled on snb, thanks to Chris Wilson for figuring this out.
Tested-by: Chris Wilson <[email protected]>
Tested-by: Michael "brot" Groh <[email protected]>
Acked-by: Ben Widawsky <[email protected]>
Signed-Off-by: Daniel Vetter <[email protected]>
|
|
The 845g shares the errata with i830 whereby executing a command
within 2 cachelines of the end of the ringbuffer may cause a GPU hang.
Signed-off-by: Chris Wilson <[email protected]>
Cc: [email protected]
Signed-off-by: Daniel Vetter <[email protected]>
|
|
This reverts commit c3dfefa0a6d235bd465309e12f4c56ea16e71111.
gmbus in 3.4 has simply too many known issues:
- gmbus is too noisy, we need to rework the logging:
https://bugs.freedesktop.org/show_bug.cgi?id=48248
- zero-length writes cause an OOPS, and they are
userspace-triggerable:
https://lkml.org/lkml/2012/3/30/176
- same for zero-length reads:
https://bugs.freedesktop.org/show_bug.cgi?id=48269
We can try again for 3.5.
Acked-by: Chris Wilson <[email protected]>
Signed-Off-by: Daniel Vetter <[email protected]>
|
|
Wake secondary cpu via resched instead of
"Unknown IPI message 0x1"
Signed-off-by: Jonas Aaberg <[email protected]>
Reviewed-by: Rickard Andersson <[email protected]>
Signed-off-by: Linus Walleij <[email protected]>
|
|
It seems it can corrupt the monitor EDID in certain cases on certain
boards when running sensors detect. It's rarely used anyway outside
of AIW boards.
http://lists.lm-sensors.org/pipermail/lm-sensors/2012-April/035847.html
http://lists.freedesktop.org/archives/xorg/2011-January/052239.html
Signed-off-by: Alex Deucher <[email protected]>
Cc: [email protected]
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
Fix build warnings by providing a struct stub since no fields of
the struct are used:
include/linux/vgaarb.h:66:9: warning: 'struct pci_dev' declared inside parameter list
include/linux/vgaarb.h:66:9: warning: its scope is only this definition or declaration, which is probably not what you want
include/linux/vgaarb.h:99:34: warning: 'struct pci_dev' declared inside parameter list
include/linux/vgaarb.h:109:6: warning: 'struct pci_dev' declared inside parameter list
include/linux/vgaarb.h:121:8: warning: 'struct pci_dev' declared inside parameter list
include/linux/vgaarb.h:140:37: warning: 'struct pci_dev' declared inside parameter list
Signed-off-by: Randy Dunlap <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
This patch fixes the irq_domain_mapping debugfs output to pad pointer
values with leading zeros so that pointer values are displayed
correctly. Otherwise you get output similar to "0x 5e0000000000000".
Also, when the irq_domain is set to 'null'
Signed-off-by: Grant Likely <[email protected]>
Cc: David Daney <[email protected]>
Cc: Mika Westerberg <[email protected]>
|
|
The actual name of the irq_domain mapping debugfs file is
"irq_domain_mapping" not "virq_mapping".
Signed-off-by: Mika Westerberg <[email protected]>
Signed-off-by: Grant Likely <[email protected]>
|
|
It makes no sense to export this trivial function. Make it a static inline
instead.
This patch also drops virq_to_hw from arch/c6x since it is unused by that
architecture.
v2: Move irq_hw_number_t into types.h to fix ARM build failure
Signed-off-by: Grant Likely <[email protected]>
Acked-by: Thomas Gleixner <[email protected]>
Cc: Benjamin Herrenschmidt <[email protected]>
|
|
In commit 4bbdd45a (irq_domain/powerpc: eliminate irq_map; use
irq_alloc_desc() instead) code was added that ignores error returns
from irq_alloc_desc_from() by (silently) casting the return value to
unsigned. The negitive value error return now suddenly looks like a
valid irq number.
Commits cc79ca69 (irq_domain: Move irq_domain code from powerpc to
kernel/irq) and 1bc04f2c (irq_domain: Add support for base irq and
hwirq in legacy mappings) move this code to its current location in
irqdomain.c
The result of all of this is a null pointer dereference OOPS if one of
the error cases is hit.
The fix: Don't cast away the negativeness of the return value and then
check for errors.
Signed-off-by: David Daney <[email protected]>
Acked-by: Rob Herring <[email protected]>
[grant.likely: dropped addition of new 'irq' variable]
Signed-off-by: Grant Likely <[email protected]>
|
|
When no platform_data is present and either 'spi-num-chipselects' is
not defined in the DT or 'cs-gpios' has less entries than
'spi-num-chipselects' specifies, the NULL platform_data pointer is
being dereferenced.
Signed-off-by: Lothar Waßmann <[email protected]>
Signed-off-by: Grant Likely <[email protected]>
|
|
Fixes the following warning when "SAMSUNG EXYNOS5" is not selected:
warning: ‘exynos5_gpios_1’ defined but not used [-Wunused-variable]
warning: ‘exynos5_gpios_2’ defined but not used [-Wunused-variable]
warning: ‘exynos5_gpios_3’ defined but not used [-Wunused-variable]
warning: ‘exynos5_gpios_4’ defined but not used [-Wunused-variable]
Signed-off-by: Sachin Kamat <[email protected]>
Signed-off-by: Grant Likely <[email protected]>
|
|
After discussion with Tim Hockin.
Signed-off-by: David S. Miller <[email protected]>
|
|
Commit fe1952fc0afb9a2e4c79f103c08aef5d13db1873
"powerpc: Rework runlatch code" has a nasty typo
where it uses "TLF_RUNLATCH" instead of "_TLF_RUNLATCH"
(bit number instead of bit mask), causing some flags to
be potentially lost such as _TLF_RESTORE_SIGMASK
(Brown paper bag for me ! We should be able to make
that break at compile time with a bit of magic, any
volunteer ?)
Signed-off-by: Benjamin Herrenschmidt <[email protected]>
|
|
This fixes builds where CONFIG_AUDIT is not defined and
CONFIG_SECURITY_SMACK=y.
This got introduced by the stack-usage reducation commit 48c62af68a40
("LSM: shrink the common_audit_data data union").
Signed-off-by: Kees Cook <[email protected]>
Acked-by: Eric Paris <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/djbw/dmaengine
Pull dmaengine fixes from Dan Williams:
1/ regression fix for Xen as it now trips over a broken assumption
about the dma address size on 32-bit builds
2/ new quirk for netdma to ignore dma channels that cannot meet
netdma alignment requirements
3/ fixes for two long standing issues in ioatdma (ring size overflow)
and iop-adma (potential stack corruption)
* tag 'dmaengine-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/dmaengine:
netdma: adding alignment check for NETDMA ops
ioatdma: DMA copy alignment needed to address IOAT DMA silicon errata
ioat: ring size variables need to be 32bit to avoid overflow
iop-adma: Corrected array overflow in RAID6 Xscale(R) test.
ioat: fix size of 'completion' for Xen
|
|
xhci_unregister_pci() is called in xhci_hcd_init().
Signed-off-by: Gerard Snitselaar <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
|
|
Correct the print of HSEE of USBCMD in xhci-dbg.c.
Signed-off-by: Alex He <[email protected]>
Signed-off-by: Sarah Sharp <[email protected]>
|
|
1/ convert open-coded KERN_ERR+dump_stack() to WARN(), so that automated
tools pick up this warning.
2/ include the 'child' and 'parent' kobject names. This information was
useful for tracking down the case where scsi invoked device_del() on a
parent object and subsequently invoked device_add() on a child. Now the
warning looks like:
kobject_add_internal failed for target8:0:16 (error: -2 parent: end_device-8:0:24)
Pid: 2942, comm: scsi_scan_8 Not tainted 3.3.0-rc7-isci+ #2
Call Trace:
[<ffffffff8125e551>] kobject_add_internal+0x1c1/0x1f3
[<ffffffff81075149>] ? trace_hardirqs_on+0xd/0xf
[<ffffffff8125e659>] kobject_add_varg+0x41/0x50
[<ffffffff8125e723>] kobject_add+0x64/0x66
[<ffffffff8131124b>] device_add+0x12d/0x63a
[<ffffffff8125e0ef>] ? kobject_put+0x4c/0x50
[<ffffffff8132f370>] scsi_sysfs_add_sdev+0x4e/0x28a
[<ffffffff8132dce3>] do_scan_async+0x9c/0x145
Cc: Greg Kroah-Hartman <[email protected]>
Cc: James Bottomley <[email protected]>
Signed-off-by: Dan Williams <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
In scsi at least two cases of the parent device being deleted before the
child is added have been observed.
1/ scsi is performing async scans and the device is removed prior to the
async can thread running (can happen with an in-opportune / unlikely
unplug during initial scan).
2/ libsas discovery event running after the parent port has been torn
down (this is a bug in libsas).
Result in crash signatures like:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000098
IP: [<ffffffff8115e100>] sysfs_create_dir+0x32/0xb6
...
Process scsi_scan_8 (pid: 5417, threadinfo ffff88080bd16000, task ffff880801b8a0b0)
Stack:
00000000fffffffe ffff880813470628 ffff88080bd17cd0 ffff88080614b7e8
ffff88080b45c108 00000000fffffffe ffff88080bd17d20 ffffffff8125e4a8
ffff88080bd17cf0 ffffffff81075149 ffff88080bd17d30 ffff88080614b7e8
Call Trace:
[<ffffffff8125e4a8>] kobject_add_internal+0x120/0x1e3
[<ffffffff81075149>] ? trace_hardirqs_on+0xd/0xf
[<ffffffff8125e641>] kobject_add_varg+0x41/0x50
[<ffffffff8125e70b>] kobject_add+0x64/0x66
[<ffffffff8131122b>] device_add+0x12d/0x63a
In this scenario the parent is still valid (because we have a
reference), but it has been device_del()'d which means its kobj->sd
pointer is NULL'd via:
device_del()->kobject_del()->sysfs_remove_dir()
...and then sysfs_create_dir() (without this fix) goes ahead and
de-references parent_sd via sysfs_ns_type():
return (sd->s_flags & SYSFS_NS_TYPE_MASK) >> SYSFS_NS_TYPE_SHIFT;
This scenario is being fixed in scsi/libsas, but if other subsystems
present the same ordering the system need not immediately crash.
Cc: Greg Kroah-Hartman <[email protected]>
Cc: James Bottomley <[email protected]>
Signed-off-by: Dan Williams <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Do not let the kernel crash when a device is registered with
sysfs while group attributes are not set (aka NULL).
Warn about the offender with some information about the offending
device.
This would warn instead of trying NULL pointer deref like:
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<ffffffff81152673>] internal_create_group+0x83/0x1a0
PGD 0
Oops: 0000 [#1] SMP
CPU 0
Modules linked in:
Pid: 1, comm: swapper/0 Not tainted 3.4.0-rc1-x86_64 #3 HP ProLiant DL360 G4
RIP: 0010:[<ffffffff81152673>] [<ffffffff81152673>] internal_create_group+0x83/0x1a0
RSP: 0018:ffff88019485fd70 EFLAGS: 00010202
RAX: 0000000000000001 RBX: 0000000000000000 RCX: 0000000000000001
RDX: ffff880192e99908 RSI: ffff880192e99630 RDI: ffffffff81a26c60
RBP: ffff88019485fdc0 R08: 0000000000000000 R09: 0000000000000000
R10: ffff880192e99908 R11: 0000000000000000 R12: ffffffff81a16a00
R13: ffff880192e99908 R14: ffffffff81a16900 R15: 0000000000000000
FS: 0000000000000000(0000) GS:ffff88019bc00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 0000000001a0c000 CR4: 00000000000007f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process swapper/0 (pid: 1, threadinfo ffff88019485e000, task ffff880194878000)
Stack:
ffff88019485fdd0 ffff880192da9d60 0000000000000000 ffff880192e99908
ffff880192e995d8 0000000000000001 ffffffff81a16a00 ffff880192da9d60
0000000000000000 0000000000000000 ffff88019485fdd0 ffffffff811527be
Call Trace:
[<ffffffff811527be>] sysfs_create_group+0xe/0x10
[<ffffffff81376ca6>] device_add_groups+0x46/0x80
[<ffffffff81377d3d>] device_add+0x46d/0x6a0
...
Signed-off-by: Bruno Prémont <[email protected]>
Acked-by: Ingo Molnar <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
Fix race between probe and open by making sure that the disconnected
flag is not cleared until all ports have been registered.
A call to tty_open while probe is running may get a reference to the
serial structure in serial_install before its ports have been
registered. This may lead to usb_serial_core calling driver open before
port is fully initialised.
With ftdi_sio this result in the following NULL-pointer dereference as
the private data has not been initialised at open:
[ 199.698286] IP: [<f811a089>] ftdi_open+0x59/0xe0 [ftdi_sio]
[ 199.698297] *pde = 00000000
[ 199.698303] Oops: 0000 [#1] PREEMPT SMP
[ 199.698313] Modules linked in: ftdi_sio usbserial
[ 199.698323]
[ 199.698327] Pid: 1146, comm: ftdi_open Not tainted 3.2.11 #70 Dell Inc. Vostro 1520/0T816J
[ 199.698339] EIP: 0060:[<f811a089>] EFLAGS: 00010286 CPU: 0
[ 199.698344] EIP is at ftdi_open+0x59/0xe0 [ftdi_sio]
[ 199.698348] EAX: 0000003e EBX: f5067000 ECX: 00000000 EDX: 80000600
[ 199.698352] ESI: f48d8800 EDI: 00000001 EBP: f515dd54 ESP: f515dcfc
[ 199.698356] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
[ 199.698361] Process ftdi_open (pid: 1146, ti=f515c000 task=f481e040 task.ti=f515c000)
[ 199.698364] Stack:
[ 199.698368] f811a9fe f811a9e0 f811b3ef 00000000 00000000 00001388 00000000 f4a86800
[ 199.698387] 00000002 00000000 f806e68e 00000000 f532765c f481e040 00000246 22222222
[ 199.698479] 22222222 22222222 22222222 f5067004 f5327600 f5327638 f515dd74 f806e6ab
[ 199.698496] Call Trace:
[ 199.698504] [<f806e68e>] ? serial_activate+0x2e/0x70 [usbserial]
[ 199.698511] [<f806e6ab>] serial_activate+0x4b/0x70 [usbserial]
[ 199.698521] [<c126380c>] tty_port_open+0x7c/0xd0
[ 199.698527] [<f806e660>] ? serial_set_termios+0xa0/0xa0 [usbserial]
[ 199.698534] [<f806e76f>] serial_open+0x2f/0x70 [usbserial]
[ 199.698540] [<c125d07c>] tty_open+0x20c/0x510
[ 199.698546] [<c10e9eb7>] chrdev_open+0xe7/0x230
[ 199.698553] [<c10e48f2>] __dentry_open+0x1f2/0x390
[ 199.698559] [<c144bfec>] ? _raw_spin_unlock+0x2c/0x50
[ 199.698565] [<c10e4b76>] nameidata_to_filp+0x66/0x80
[ 199.698570] [<c10e9dd0>] ? cdev_put+0x20/0x20
[ 199.698576] [<c10f3e08>] do_last+0x198/0x730
[ 199.698581] [<c10f4440>] path_openat+0xa0/0x350
[ 199.698587] [<c10f47d5>] do_filp_open+0x35/0x80
[ 199.698593] [<c144bfec>] ? _raw_spin_unlock+0x2c/0x50
[ 199.698599] [<c10ff110>] ? alloc_fd+0xc0/0x100
[ 199.698605] [<c10f0b72>] ? getname_flags+0x72/0x120
[ 199.698611] [<c10e4450>] do_sys_open+0xf0/0x1c0
[ 199.698617] [<c11fcc08>] ? trace_hardirqs_on_thunk+0xc/0x10
[ 199.698623] [<c10e458e>] sys_open+0x2e/0x40
[ 199.698628] [<c144c990>] sysenter_do_call+0x12/0x36
[ 199.698632] Code: 85 89 00 00 00 8b 16 8b 4d c0 c1 e2 08 c7 44 24 14 88 13 00 00 81 ca 00 00 00 80 c7 44 24 10 00 00 00 00 c7 44 24 0c 00 00 00 00 <0f> b7 41 78 31 c9 89 44 24 08 c7 44 24 04 00 00 00 00 c7 04 24
[ 199.698884] EIP: [<f811a089>] ftdi_open+0x59/0xe0 [ftdi_sio] SS:ESP 0068:f515dcfc
[ 199.698893] CR2: 0000000000000078
[ 199.698925] ---[ end trace 77c43ec023940cff ]---
Reported-and-tested-by: Ken Huang <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Johan Hovold <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
of_gpio_simple_xlate() has an off-by-one bug where it checks to see if
args[0] is > ngpio instead of >=. args[0] must always be less than
ngpio because it is a zero-based enumeration.
Signed-off-by: Roland Stigge <[email protected]>
[grant.likely: beef up commit text]
Signed-off-by: Grant Likely <[email protected]>
|
|
The variable 'bit' is uninitialized in the first iteration of for
loop. Fix it.
Signed-off-by: Axel Lin <[email protected]>
Signed-off-by: Grant Likely <[email protected]>
|
|
The irqdomain api changed significantly in v3.4 which caused a build
failure for this driver.
Signed-off-by: Grant Likely <[email protected]>
Acked-by: Sebastian Andrzej Siewior <[email protected]>
Cc: Hans J. Koch <[email protected]>
Cc: Torben Hohn <[email protected]>
|
|
Fix a code path in tcp_rcv_rtt_update() that was comparing scaled and
unscaled RTT samples.
The intent in the code was to only use the 'm' measurement if it was a
new minimum. However, since 'm' had not yet been shifted left 3 bits
but 'new_sample' had, this comparison would nearly always succeed,
leading us to erroneously set our receive-side RTT estimate to the 'm'
sample when that sample could be nearly 8x too high to use.
The overall effect is to often cause the receive-side RTT estimate to
be significantly too large (up to 40% too large for brief periods in
my tests).
Signed-off-by: Neal Cardwell <[email protected]>
Acked-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Commit c43b874d5d714f (tcp: properly initialize tcp memory limits) tried
to fix a regression added in commits 4acb4190 & 3dc43e3,
but still get it wrong.
Result is machines with low amount of memory have too small tcp_rmem[2]
value and slow tcp receives : Per socket limit being 1/1024 of memory
instead of 1/128 in old kernels, so rcv window is capped to small
values.
Fix this to match comment and previous behavior.
Signed-off-by: Eric Dumazet <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Glauber Costa <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|