Johannes Weiner [Thu, 25 Aug 2011 22:59:16 +0000 (15:59 -0700)]
memcg: fix hierarchical oom locking
Commit 79dfdaccd1d5 ("memcg: make oom_lock 0 and 1 based rather than
counter") tried to oom lock the hierarchy and roll back upon
encountering an already locked memcg.
The code is confused when it comes to detecting a locked memcg, though,
so it would fail and rollback after locking one memcg and encountering
an unlocked second one.
The result is that oom-locking hierarchies fails unconditionally and
that every oom killer invocation simply goes to sleep on the oom
waitqueue forever. The tasks practically hang forever without anyone
intervening, possibly holding locks that trip up unrelated tasks, too.
Axel Lin [Thu, 25 Aug 2011 22:59:14 +0000 (15:59 -0700)]
drivers/video/backlight/ep93xx_bl.c: add missing include of linux/module.h
ep93xx_bl.c uses interfaces from linux/module.h, so it should include
that file. This patch fixes build errors:
CC [M] drivers/video/backlight/ep93xx_bl.o
drivers/video/backlight/ep93xx_bl.c:138: error: 'THIS_MODULE' undeclared here (not in a function)
drivers/video/backlight/ep93xx_bl.c:158: error: expected declaration specifiers or '...' before string constant
drivers/video/backlight/ep93xx_bl.c:158: warning: data definition has no type or storage class
...
Replace/remove use of RIO v.1.2 registers/bits that are not
forward-compatible with newer versions of RapidIO specification.
RapidIO specification v.1.3 removed Write Port CSR, Doorbell CSR,
Mailbox CSR and Mailbox and Doorbell bits of the PEF CAR.
Use of removed (since RIO v.1.3) register bits affects users of
currently available 1.3 and 2.x compliant devices who may use not so
recent kernel versions.
Removing checks for unsupported bits makes corresponding routines
compatible with all versions of RapidIO specification. Therefore,
backporting makes stable kernel versions compliant with RIO v.1.3 and
later as well.
Shaohua Li [Thu, 25 Aug 2011 22:59:12 +0000 (15:59 -0700)]
vmscan: clear ZONE_CONGESTED for zone with good watermark
ZONE_CONGESTED is only cleared in kswapd, but pages can be freed in any
task. It's possible ZONE_CONGESTED isn't cleared in some cases:
1. the zone is already balanced just entering balance_pgdat() for
order-0 because concurrent tasks free memory. In this case, later
check will skip the zone as it's balanced so the flag isn't cleared.
2. high order balance fallbacks to order-0. quote from Mel: At the
end of balance_pgdat(), kswapd uses the following logic;
If reclaiming at high order {
for each zone {
if all_unreclaimable
skip
if watermark is not met
order = 0
loop again
/* watermark is met */
clear congested
}
}
i.e. it clears ZONE_CONGESTED if it the zone is balanced. if not,
it restarts balancing at order-0. However, if the higher zones are
balanced for order-0, kswapd will miss clearing ZONE_CONGESTED as
that only happens after a zone is shrunk. This can mean that
wait_iff_congested() stalls unnecessarily.
This patch makes kswapd clear ZONE_CONGESTED during its initial
highmem->dma scan for zones that are already balanced.
kernel/printk: do not turn off bootconsole in printk_late_init() if keep_bootcon
It seems that 7bf693951a8e ("console: allow to retain boot console via
boot option keep_bootcon") doesn't always achieve what it aims, as when
printk_late_init() runs it unconditionally turns off all boot consoles.
With this patch, I am able to see more messages on the boot console in
KVM guests than I can without, when keep_bootcon is specified.
I think it is appropriate for the relevant -stable trees. However, it's
more of an annoyance than a serious bug (ideally you don't need to keep
the boot console around as console handover should be working -- I was
encountering a situation where the console handover wasn't working and
not having the boot console available meant I couldn't see why).
Johannes Weiner [Thu, 25 Aug 2011 22:59:07 +0000 (15:59 -0700)]
memcg: pin execution to current cpu while draining stock
Commit d1a05b6973c7 ("memcg do not try to drain per-cpu caches without
pages") added a drain_local_stock() call to a preemptible section.
The draining task looks up the cpu-local stock twice to set the
draining-flag, then to drain the stock and clear the flag again. If the
task is migrated to a different CPU in between, noone will clear the
flag on the first stock and it will be forever undrainable. Its charge
can not be recovered and the cgroup can not be deleted anymore.
Properly pin the task to the executing CPU while draining stocks.
Dan Carpenter [Thu, 25 Aug 2011 22:59:04 +0000 (15:59 -0700)]
w1: fix for loop in w1_f29_remove_slave()
The for loop was looking for i <= 0 instead of i >= 0 so this function
never did anything. Also we started with i = NB_SYSFS_BIN_FILES instead
of "NB_SYSFS_BIN_FILES - 1" which is an off by one bug.
drivers/misc/pti.c: In function 'get_id':
drivers/misc/pti.c:249: error: implicit declaration of function 'kmalloc'
drivers/misc/pti.c: In function 'pti_char_write':
drivers/misc/pti.c:658: error: implicit declaration of function 'copy_from_user'
Andy Lutomirski [Thu, 25 Aug 2011 20:10:33 +0000 (16:10 -0400)]
x86-32: Fix boot with CONFIG_X86_INVD_BUG
entry_32.S contained a hardcoded alternative instruction entry, and the
format changed in commit 59e97e4d6fbc ("x86: Make alternative
instruction pointers relative").
Replace the hardcoded entry with the altinstruction_entry macro. This
fixes the 32-bit boot with CONFIG_X86_INVD_BUG=y.
Tejun Heo [Thu, 25 Aug 2011 17:46:56 +0000 (19:46 +0200)]
mtrr: fix UP breakage caused during switch to stop_machine
While removing custom rendezvous code and switching to stop_machine,
commit 192d8857427d ("x86, mtrr: use stop_machine APIs for doing MTRR
rendezvous") completely dropped mtrr setting code on !CONFIG_SMP
breaking MTRR settting on UP.
Josh Boyer [Thu, 25 Aug 2011 11:48:12 +0000 (07:48 -0400)]
lockdep: Add helper function for dir vs file i_mutex annotation
Purely in-memory filesystems do not use the inode hash as the dcache
tells us if an entry already exists. As a result, they do not call
unlock_new_inode, and thus directory inodes do not get put into a
different lockdep class for i_sem.
We need the different lockdep classes, because the locking order for
i_mutex is different for directory inodes and regular inodes. Directory
inodes can do "readdir()", which takes i_mutex *before* possibly taking
mm->mmap_sem (due to a page fault while copying the directory entry to
user space).
In contrast, regular inodes can be mmap'ed, which takes mm->mmap_sem
before accessing i_mutex.
The two cases can never happen for the same inode, so no real deadlock
can occur, but without the different lockdep classes, lockdep cannot
understand that. As a result, if CONFIG_DEBUG_LOCK_ALLOC is set, this
can lead to false positives from lockdep like below:
find/645 is trying to acquire lock:
(&mm->mmap_sem){++++++}, at: [<ffffffff81109514>] might_fault+0x5c/0xac
but task is already holding lock:
(&sb->s_type->i_mutex_key#15){+.+.+.}, at: [<ffffffff81149f34>]
vfs_readdir+0x5b/0xb4
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
This patch moves the directory vs file lockdep annotation into a helper
function that can be called by in-memory filesystems and has hugetlbfs
call it.
* git://git.kernel.org/pub/scm/linux/kernel/git/brodo/cpupowerutils:
cpupower: use man(1) when calling "cpupower help subcommand"
cpupower: make NLS truly optional
cpupower: fix Makefile typo
cpupower: Make monitor command -c/--cpu aware
cpupower: Better detect offlined CPUs
cpupower: Do not show an empty Idle_Stats monitor if no idle driver is available
cpupower: mperf monitor - Use TSC to calculate max frequency if possible
cpupower: avoid using symlinks
Linus Torvalds [Thu, 25 Aug 2011 17:31:40 +0000 (10:31 -0700)]
Merge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging
* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging:
hwmon: (i5k_amb) Drop i5k_channel_pci_id
hwmon: (ntc_thermistor) Simplify if sequence
Linus Torvalds [Thu, 25 Aug 2011 17:30:51 +0000 (10:30 -0700)]
Merge branch '3.1-rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
* '3.1-rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (21 commits)
target: Convert acl_node_lock to be IRQ-disabling
target: Make locking in transport_deregister_session() IRQ safe
tcm_fc: init/exit functions should not be protected by "#ifdef MODULE"
target: Print subpage too for unhandled MODE SENSE pages
iscsi-target: Fix iscsit_allocate_se_cmd_for_tmr failure path bugs
iscsi-target: Implement iSCSI target IPv6 address printing.
target: Fix task SGL chaining breakage with transport_allocate_data_tasks
target: Fix task count > 1 handling breakage and use max_sector page alignment
target: Add missing DATA_SG_IO transport_cmd_get_valid_sectors check
target: Fix SYNCHRONIZE_CACHE zero LBA + range breakage
target: Remove duplicate task completions in transport_emulate_control_cdb
target: Fix WRITE_SAME usage with transport_get_size
target: Add WRITE_SAME (10) parsing and refactor passthrough checks
target: Fix write payload exception handling with ->new_cmd_map
iscsi-target: forever loop bug in iscsit_attach_ooo_cmdsn()
iscsi-target: remove duplicate return
target: Convert target_core_rd.c to use use BUG_ON
iscsi-target: Fix leak on failure in iscsi_copy_param_list()
target: Use ERR_CAST inlined function
target: Make standard INQUIRY return 'not connected' for tpg_virt_lun0
...
Andi Kleen [Fri, 19 Aug 2011 23:15:10 +0000 (16:15 -0700)]
Add a personality to report 2.6.x version numbers
I ran into a couple of programs which broke with the new Linux 3.0
version. Some of those were binary only. I tried to use LD_PRELOAD to
work around it, but it was quite difficult and in one case impossible
because of a mix of 32bit and 64bit executables.
For example, all kind of management software from HP doesnt work, unless
we pretend to run a 2.6 kernel.
$ uname -a
Linux svivoipvnx001 3.0.0-08107-g97cd98f #1062 SMP Fri Aug 12 18:11:45 CEST 2011 i686 i686 i386 GNU/Linux
$ hpacucli ctrl all show
Error: No controllers detected.
$ rpm -qf /usr/sbin/hpacucli
hpacucli-8.75-12.0
Another notable case is that Python now reports "linux3" from
sys.platform(); which in turn can break things that were checking
sys.platform() == "linux2":
Kevin Hilman [Thu, 25 Aug 2011 13:31:14 +0000 (15:31 +0200)]
OMAP: omap_device: only override _noirq methods, not normal suspend/resume
Commit c03f007a8bf0e092caeb6856a5c8a850df10b974 (OMAP: PM:
omap_device: add system PM methods for PM domain handling) mistakenly
used SET_SYSTEM_SLEEP_PM_OPS() when trying to configure custom methods
for the PM domains noirq methods. Fix that by setting only the
suspend_noirq and resume_noirq methods with custom versions.
Note that all other PM domain methods (including the "normal"
suspend/resume methods) are populated using USE_PLATFORM_PM_SLEEP_OPS,
which configures them all to the default subsystem (platform_bus)
methods.
PM / Runtime: Correct documentation of pm_runtime_irq_safe()
The description of pm_runtime_irq_safe() has to be updated to follow
the code after commit 02b2677 (PM / Runtime: Allow _put_sync() from
interrupts-disabled context).
ALSA: hda: Conexant: Allow different output types to share DAC
Headphones has stopped working for the original reported (a regression
compared to 2.6.38). This is because Speaker and Headphones share the
same DAC, in which case no Headphones volume control was created.
This patch fixes so that both Speaker and Headphones volume
controls are created in such scenario.
Dereferencing a user pointer directly from kernel-space without going
through the copy_from_user family of functions is a bad idea. Two of
such usages can be found in the sendmsg code path called from sendmmsg,
added by
Usages are performed through memcmp() and memcpy() directly. Fix those
by using the already copied msg_sys structure instead of the __user *msg
structure. Note that msg_sys can be set to NULL by verify_compat_iovec()
or verify_iovec(), which requires additional NULL pointer checks.
Eric Dumazet [Mon, 22 Aug 2011 19:32:42 +0000 (19:32 +0000)]
arp: fix rcu lockdep splat in arp_process()
Dave Jones reported a lockdep splat triggered by an arp_process() call
from parp_redo().
Commit faa9dcf793be (arp: RCU changes) is the origin of the bug, since
it assumed arp_process() was called under rcu_read_lock(), which is not
true in this particular path.
Instead of adding rcu_read_lock() in parp_redo(), I chose to add it in
neigh_proxy_process() to take care of IPv6 side too.
Linus Torvalds [Wed, 24 Aug 2011 22:55:30 +0000 (15:55 -0700)]
firmware loader: allow builtin firmware load even if usermodehelper is disabled
In commit a144c6a6c924 ("PM: Print a warning if firmware is requested
when tasks are frozen") we not only printed a warning if somebody tried
to load the firmware when tasks are frozen - we also failed the load.
But that check was done before the check for built-in firmware, and then
when we disallowed usermode helpers during bootup (commit 288d5abec831:
"Boot up with usermodehelper disabled"), that actually means that
built-in modules can no longer load their firmware even if the firmware
is built in too. Which used to work, and some people depended on it for
the R100 driver.
So move the test for usermodehelper_is_disabled() down, to after
checking the built-in firmware.
Linus Torvalds [Wed, 24 Aug 2011 22:54:30 +0000 (15:54 -0700)]
Merge branch 'for-31-rc3/i2c-fixes' of git://git.fluff.org/bjdooks/linux
* 'for-31-rc3/i2c-fixes' of git://git.fluff.org/bjdooks/linux:
i2c-nomadik: fix kerneldoc warning
Revert "i2c-omap: fix static suspend vs. runtime suspend"
i2c-nomadik: Do not use _interruptible_ variant call
Magnus Damm [Wed, 24 Aug 2011 20:38:43 +0000 (22:38 +0200)]
ARM: mach-shmobile: sh7372 LCDC1 suspend fix
Associate the HDMI clock together with LCDC1 on sh7372.
Without this patch Suspend-to-RAM hangs on the boards
AP4EVB and Mackerel. The code hangs in the LCDC driver
where the software is waiting forever for the hardware to
power down. By explicitly associating the HDMI clock with
LCDC1 we can make sure the HDMI clock is enabled using
Runtime PM whenever the driver is accessing the hardware.
This HDMI and LCDC1 dependency is documented in the sh7372
data sheet. Older kernels did work as expected but the
recently merged (3.1-rc)
Since sci_port_enable() and sci_port_disable() may be run with
interrupts off and they execute pm_runtime_get_sync() and
pm_runtime_put_sync(), respectively, the SCI device's
power.irq_safe flag has to be set to indicate that it is safe
to execute runtime PM callbacks for this device with interrupts off.
PM: Use spinlock instead of mutex in clock management functions
The lock member of struct pm_clk_data is of type struct mutex,
which is a problem, because the suspend and resume routines
defined in drivers/base/power/clock_ops.c cannot be executed
with interrupts disabled for this reason. Modify
struct pm_clk_data so that its lock member is a spinlock.
Linus Torvalds [Wed, 24 Aug 2011 18:25:08 +0000 (11:25 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
HID: wiimote: Add status and return request handlers
HID: wiimote: Add drm request
HID: wiimote: Register led class devices
HID: wiimote: Correctly call HID open/close callbacks
HID: wiimote: Simplify synchronization
HID: usbhid: Add support for SiGma Micro chip
HID: add support for new revision of Apple aluminum keyboard
ath9k: Fix PS wrappers in ath9k_set_coverage_class
this callback is called during suspend/resume and also via iw command.
it configures parameters like sifs, slottime, acktimeout in
ath9k_hw_init_global_settings where few REG_READ, REG_RMW are also done
and hence the need for PS wrappers
carl9170: Fix mismatch in carl9170_op_set_key mutex lock-unlock
If is_main_vif(ar, vif) reports that we have to fall back
to software encryption, we goto err_softw; before locking ar->mutex.
As a result, we have unprotected call to carl9170_set_operating_mode
and unmatched mutex_unlock.
The patch fix the issue by adding mutex_lock before goto.
Found by Linux Driver Verification project (linuxtesting.org).
Linus Torvalds [Wed, 24 Aug 2011 16:19:03 +0000 (09:19 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: ad714x - read the interrupt status registers in a row
Input: ad714x - use DMA-safe buffers for spi_write()
Input: ad714x - fix endianness issues
Input: ad714xx-spi - force SPI bus into the default 8-bit mode
Input: ep93xx_keypad - add missing include of linux/module.h
Input: tnetv107x-ts - add missing include of linux/module.h
Input: max11801_ts - correct license statement
Input: atmel_mxt_ts - report pressure information from the driver
Input: bcm5974 - Add support for newer MacBookPro8,2
Input: wacom - report id 3 returns 4 bytes of data
Input: wacom - add WAC_MSG_RETRIES define
Input: wacom - add support for the Wacom Bamboo Pen (CTL-660/K)
Input: tegra-kbc - correct call to input_free_device
Input: mpu3050 - correct call to input_free_device
Input: bcm5974 - add support for touchpads found in MacBookAir4,2
Input: mma8450 - fix module device table type
Input: remove CLOCK_TICK_RATE from analog joystick driver
Linus Torvalds [Wed, 24 Aug 2011 16:14:42 +0000 (09:14 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: check size of FUSE_NOTIFY_INVAL_ENTRY message
fuse: mark pages accessed when written to
fuse: delete dead .write_begin and .write_end aops
fuse: fix flock
fuse: fix non-ANSI void function notation
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (29 commits)
bridge: fix a possible net_device leak
net: Documentation: RFC 2553bis is now RFC 3493
atm: br2684: Fix oops due to skb->dev being NULL
ipv6: Fix ipv6_getsockopt for IPV6_2292PKTOPTIONS
net: netdev-features.txt update to Documentation/networking/00-INDEX
vlan: reset headers on accel emulation path
forcedeth: call vlan_mode only if hw supports vlans
via-velocity: remove non-tagged packet filtering
bonding:reset backup and inactive flag of slave
net_sched: fix port mirror/redirect stats reporting
sit tunnels: propagate IPv6 transport class to IPv4 Type of Service
gianfar: reduce stack usage in gianfar_ethtool.c
net: minor update to Documentation/networking/scaling.txt
net: add missing entries to Documentation/networking/00-INDEX
gianfar: prevent buggy hw rx vlan tagging
net: sh_eth: Fix build by forgot including linux/interrupt.h
drivers/net/can/sja1000/plx_pci.c: eliminate double free
usbnet/cdc_ncm: Don't use stack variables for DMA
vmxnet3: Don't enable vlan filters in promiscuous mode.
iwlagn: sysfs couldn't find the priv pointer
...
Michael Holzheu [Wed, 24 Aug 2011 15:15:12 +0000 (17:15 +0200)]
[S390] Change default action from reipl to stop for on_restart
The main purpose for PSW restart will be kdump. Therefore customers will
issue "system restart" for creating a dump. If kdump is not enabled,
currently "PSW restart" will reboot the system and then no dump can
be created any more. In order to still allow a manual stand-alone dump in
the case a user issues "PSW restart" on a system that has not enabled
kdump we now stop the system.
x = \(kmalloc\|kzalloc\|kcalloc\)(...);
...
if (x == NULL) S
<... when != x
when != if (...) { <+...kfree(x)...+> }
when any
when != true x == NULL
x->fl
...>
(
if (x == NULL) S1
|
if (...) { ... when != x
when forall
(
return \(0\|<+...x...+>\|ptr\);
|
* return ...;
)
}
)
// </smpl>
Heiko Carstens [Wed, 24 Aug 2011 15:15:09 +0000 (17:15 +0200)]
[S390] nss,initrd: kernel image and initrd must be in different segments
When IPL'ing from a block device and an NSS should be created we must
make sure that the kernel image and the initrd are in different 1MB
segments. Otherwise creating the NSS will fail.
So we make sure the initrd is 4MB behind the end of the kernel image
like we do already when IPL via the VM reader is performed.
Stephen Warren [Tue, 23 Aug 2011 17:16:28 +0000 (11:16 -0600)]
ASoC: Tegra: wm8903 machine driver: Drop Ventana support
Board file support for Ventana is not yet mainlined, and probably won't
ever be given the move to Device-Tree. Consequently, the Ventana entry
is being removed from arch/arm/tools/mach-types in the next merge window,
since it was registered over a year ago.
This will also remove function machine_is_ventana(), which is used by
the ASoC Tegra WM8903 machine driver. This will cause compilation
failures. Drop Ventana support to resolve this.
Hopefully, in the not-too-distant future, tegra_wm8903.c will be able to
configure itself from Device-Tree, and hence we'll be able to re-instate
Ventana support just by creating a .dts file for the board.
Also note that Aebl support is in a similar boat. However, that board
isn't scheduled for deprecation for at least another 5 months, and
perhaps we will have completely removed non-Device-Tree support from
tegra_wm8903.c by then and/or adjusted mach-types policy.
Miklos Szeredi [Wed, 24 Aug 2011 08:20:17 +0000 (10:20 +0200)]
fuse: check size of FUSE_NOTIFY_INVAL_ENTRY message
FUSE_NOTIFY_INVAL_ENTRY didn't check the length of the write so the
message processing could overrun and result in a "kernel BUG at
fs/fuse/dev.c:629!"
Jean Delvare [Wed, 24 Aug 2011 04:04:28 +0000 (21:04 -0700)]
hwmon: (i5k_amb) Drop i5k_channel_pci_id
Function i5k_channel_pci_id looks like it can fail, while a better
code design would make it more obvious that it can't. We can even get
rid of the function.
Linus Torvalds [Wed, 24 Aug 2011 01:09:08 +0000 (18:09 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86-32, vdso: On system call restart after SYSENTER, use int $0x80
x86, UV: Remove UV delay in starting slave cpus
x86, olpc: Wait for last byte of EC command to be accepted
H. Peter Anvin [Mon, 22 Aug 2011 20:27:06 +0000 (13:27 -0700)]
x86-32, vdso: On system call restart after SYSENTER, use int $0x80
When we enter a 32-bit system call via SYSENTER or SYSCALL, we shuffle
the arguments to match the int $0x80 calling convention. This was
probably a design mistake, but it's what it is now. This causes
errors if the system call as to be restarted.
For SYSENTER, we have to invoke the instruction from the vdso as the
return address is hardcoded. Accordingly, we can simply replace the
jump in the vdso with an int $0x80 instruction and use the slower
entry point for a post-restart.
Seth Jennings [Wed, 10 Aug 2011 17:56:49 +0000 (12:56 -0500)]
staging: zcache: fix crash on high memory swap
zcache_put_page() was modified to pass page_address(page) instead of the
actual page structure. In combination with the function signature changes
to tmem_put() and zcache_pampd_create(), zcache_pampd_create() tries to
(re)derive the page structure from the virtual address. However, if the
original page is a high memory page (or any unmapped page), this
virt_to_page() fails because the page_address() in zcache_put_page()
returned NULL.
This patch changes zcache_put_page() and zcache_get_page() to pass
the page structure instead of the page's virtual address, which
may or may not exist.
staging: brcm80211: fix compile error on non-x86 archs since 3.0 kernel
Since the arrival of kernel version 3.0 in the staging tree it
turns out compile error occurs for sparc64, powerpc, and arm
platforms. This patch fixes that issue.
Luciano Coelho [Tue, 23 Aug 2011 08:42:25 +0000 (11:42 +0300)]
wl12xx: add max_sched_scan_ssids value to the hw description
After commit 5a865ba, we require a separate value to indicate the
number of supported SSIDs in scheduled scans. This patch adds a
proper value to the wl12xx driver.
This fixes a regression in 3.1-rc3 where scheduled scans were not
working properly with the wl12xx driver.
Ido Yariv [Mon, 22 Aug 2011 20:19:49 +0000 (23:19 +0300)]
wl12xx: Fix validation of pm_runtime_get_sync return value
wl1271_sdio_power_on checks if the return value of pm_runtime_get_sync
is non-zero, and if so bails out.
However, pm_runtime_get_sync can return a positive number which does not
suggest an error has occurred. This is problematic for two reasons:
1. The function will needlessly bail out without decrementing back the
runtime PM reference counter.
2. wl1271_power_on only checks if wl1271_power_on return value is
negative. This means that wl1271_power_on will continue even if
wl1271_sdio_power_on bailed out. As a result, sdio transactions will
be initiated without properly enabling the sdio function and claiming
the host. This could even lead to a kernel panic.
Fix this by only checking that the return value of pm_runtime_get_sync
is non-negative.
Ido Yariv [Mon, 22 Aug 2011 20:19:48 +0000 (23:19 +0300)]
wl12xx: Remove obsolete testmode NVS push command
The testmode NVS push command is no longer in use. In addition, it has
several implementation issues that prevent it from working correctly:
1. wl1271_tm_cmd_configure relies on wl->chip.id being set. However,
since the device was not necessarily booted by the time the function
is called, wl->chip.id will be initialized to 0.
2. The NVS file is fetched by calling request_firmware() before it is
possible to push an NVS file.
3. The maximum allowed size of nl binary payloads is not sufficient for
pushing NVS files.
4. Pushing 128x NVS files will always fail due to a bug in the
validation code.
5. In case the pushed NVS file is found invalid, the mutex will be kept
locked and the nvs member will become a dangling pointer.
Since this feature is not being used, remove it completely instead of
fixing it.
Remove system PM methods which can race with runtime PM methods.
Also, as of v3.1, the PM domain level code for OMAP handles device
power state transistions automatically for devices, so drivers no
longer need to specifically call the bus/pm_domain methods themselves.
i2c-nomadik: Do not use _interruptible_ variant call
If there is a signal pending and wait_for_completion_interruptible_timeout
exited because of the -ERESTARTSYS error we are unable to send any more
i2c messages.
Linus Torvalds [Tue, 23 Aug 2011 18:41:44 +0000 (11:41 -0700)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
* 'for-linus' of git://oss.sgi.com/xfs/xfs:
xfs: fix tracing builds inside the source tree
xfs: remove subdirectories
xfs: don't expect xfs headers to be in subdirectories
Nick Pelly [Fri, 15 Jul 2011 20:53:08 +0000 (13:53 -0700)]
omap-serial: Allow IXON and IXOFF to be disabled.
Fixes logic bug that software flow control cannot be disabled, because
serial_omap_configure_xonxoff() is not called if both IXON and IXOFF bits
are cleared.
Linus Torvalds [Tue, 23 Aug 2011 17:46:21 +0000 (10:46 -0700)]
Merge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6
* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
drm/radeon: Extended DDC Probing for Toshiba L300D Radeon Mobility X1100 HDMI-A Connector
drm/ttm: ensure ttm for new node is bound before calling move_notify()
drm/ttm: unbind ttm before destroying node in accel move cleanup
drm/ttm: fix ttm_bo_add_ttm(user) failure path
drm/radeon: Make vramlimit parameter actually work.
drm/radeon: Explicitly print GTT/VRAM offsets on test failure.
drm/radeon: Take IH ring into account for test size calculation.
drm/radeon/alpha: Add Alpha support to Radeon DRM code
It turns out that this breaks several drivers, one example being OMAP
boards which use the on-board OMAP UARTs and the omap-serial driver that
will not boot to userspace after the commit.
Paul Walmsley reports that enabling CONFIG_DEBUG_SHIRQ reveals 'IRQ
handler type mismatch' errors:
IRQ handler type mismatch for IRQ 74
current handler: serial idle
...
and the reason is that setting IRQF_ONESHOT will now result in those
interrupt handlers having different IRQF flags, and thus being
unsharable. So the commit log in the reverted commit:
"Since it is required for those users and
there is no difference for others it makes sense to add this flag
unconditionally."
is simply not true: there may not be any difference from a "actions at
irq time", but there is a *big* difference wrt this flag testing irq
management (see __setup_irq() in kernel/irq/manage.c).
One solution may be to stop verifying IRQF_ONESHOT in __setup_irq(), but
right now the safe course of action is to revert the change. Let's
revisit this in a later merge window.
Jiri Slaby [Wed, 17 Aug 2011 11:48:15 +0000 (13:48 +0200)]
TTY: serial, document ignoring of uart->ops->startup error
When a user has SYS_ADMIN capabilities and uart->ops->startup returns
an error in uart_startup, we silently drop the error. We then return 0
and behave as if it didn't fail. (Not quite, since we set TTY_IO_ERROR
bit and leave ASYNC_INITIALIZED bit cleared.)
This all is to allow setserial to work with improperly configured or
unconfigured ports. User can thus set port properties and reconfigure
properly.
Jiri Slaby [Wed, 10 Aug 2011 12:59:28 +0000 (14:59 +0200)]
TTY: pty, fix pty counting
tty_operations->remove is normally called like:
queue_release_one_tty
->tty_shutdown
->tty_driver_remove_tty
->tty_operations->remove
However tty_shutdown() is called from queue_release_one_tty() only if
tty_operations->shutdown is NULL. But for pty, it is not.
pty_unix98_shutdown() is used there as ->shutdown.
So tty_operations->remove of pty (i.e. pty_unix98_remove()) is never
called. This results in invalid pty_count. I.e. what can be seen in
/proc/sys/kernel/pty/nr.
I see this was already reported at:
https://lkml.org/lkml/2009/11/5/370
But it was not fixed since then.
This patch is kind of a hackish way. The problem lies in ->install. We
allocate there another tty (so-called tty->link). So ->install is
called once, but ->remove twice, for both tty and tty->link. The fix
here is to count both tty and tty->link and divide the count by 2 for
user.
And to have ->remove called, let's make tty_driver_remove_tty() global
and call that from pty_unix98_shutdown() (tty_operations->shutdown).
While at it, let's document that when ->shutdown is defined,
tty_shutdown() is not called.
Al Cooper [Mon, 25 Jul 2011 20:19:52 +0000 (16:19 -0400)]
8250: Fix race condition in serial8250_backup_timeout().
This is to fix an issue where output will suddenly become very slow.
The problem occurs on 8250 UARTS with the hardware bug UART_BUG_THRE.
BACKGROUND
For normal UARTs (without UART_BUG_THRE): When the serial core layer
gets new transmit data and the transmitter is idle, it buffers the
data and calls the 8250s' serial8250_start_tx() routine which will
simply enable the TX interrupt in the IER register and return. This
should immediately fire a THRE interrupt and begin transmitting the
data.
For buggy UARTs (with UART_BUG_THRE): merely enabling the TX interrupt
in IER does not necessarily generate a new THRE interrupt.
Therefore, a background timer periodically checks to see if there is
pending data, and starts transmission if that is the case.
The bug happens on SMP systems when the system has nothing to transmit,
the transmit interrupt is disabled and the following sequence occurs:
- CPU0: The background timer routine serial8250_backup_timeout()
starts and saves the state of the interrupt enable register (IER)
and then disables all interrupts in IER. NOTE: The transmit interrupt
(TI) bit is saved as disabled.
- CPU1: The serial core gets data to transmit, grabs the port lock and
calls serial8250_start_tx() which enables the TI in IER.
- CPU0: serial8250_backup_timeout() waits for the port lock.
- CPU1: finishes (with TI enabled) and releases the port lock.
- CPU0: serial8250_backup_timeout() calls the interrupt routine which
will transmit the next fifo's worth of data and then restores the
IER from the previously saved value (TI disabled).
At this point, as long as the serial core has more transmit data
buffered, it will not call serial8250_start_tx() again and the
background timer routine will slowly transmit the data.
The fix is to have serial8250_start_tx() get the port lock before
it saves the IER state and release it after restoring IER. This will
prevent serial8250_start_tx() from running in parallel.
Eric Smith [Tue, 12 Jul 2011 04:53:13 +0000 (22:53 -0600)]
8250_pci: add support for Rosewill RC-305 4x serial port card
This patch adds support for the Rosewill RC-305 four-port PCI serial
card, and probably any other four-port serial cards based on the
Moschip MCS9865 chip, assuming that the EEPROM on the card was
programmed in accordance with Table 6 of the MCS9865 EEPROM
Application Note version 0.3 dated 16-May-2008, available from the
Moschip web site (registration required).
This patch is based on an earlier patch [1] for the SYBA 6x serial
port card by Ira W. Snyder.
Takashi Iwai [Tue, 23 Aug 2011 12:57:08 +0000 (14:57 +0200)]
ALSA: hda - Fix double-headphone/speaker paths for Cxt auto-parser
When multiple headphones or speakers are assigned but no individual
DACs are available, the driver should take the first HP/SPK DAC instead
of another primary output. The patch adds a bit-flag to dac field of
struct pin_dac_pair indicating that it's a slave DAC.
Toshiba Satellite L300D with ATI Mobility Radeon X1100 sends data
to i2c bus for a HDMI connector that is not implemented/existent
on the notebook's board.
Fix by applying extented DDC probing for this connector.
Requires [PATCH] drm/radeon: Extended DDC Probing for Connectors
with Improperly Wired DDC Lines
Tested for kernel 2.6.38 on Toshiba Satellite L300D notebook
Takashi Iwai [Tue, 23 Aug 2011 10:41:09 +0000 (12:41 +0200)]
ALSA: hda - Update jack-sense info even when no automute is set
The internal states, jack_present and line_jack_present should be
updated upon unsolicited events even if no automute is set.
Otherwise the wrong state is referred when the automute behavior is
changed by the mixer control.
Takashi Iwai [Tue, 23 Aug 2011 10:36:28 +0000 (12:36 +0200)]
ALSA: hda - Fix output-path initialization for Realtek auto-parser
When the headphone or speaker output has no own DAC, initialize the path
using the primary DAC. Otherwise the path won't be set properly and
can result in the silence.
David Herrmann [Wed, 17 Aug 2011 09:43:24 +0000 (11:43 +0200)]
HID: wiimote: Add status and return request handlers
The wiimote resets the current drm when an extension is plugged in.
Fortunately, it also sends a status report in this situation so we just
reset the drm on every status report to keep the drm consistent.
Also handle return reports from the wiimote which indicate success and
failure of requests that we've sent.
David Herrmann [Wed, 17 Aug 2011 09:43:23 +0000 (11:43 +0200)]
HID: wiimote: Add drm request
The wiimote reports data in several data reporting modes (DRM). The DRM
request makes the wiimote send data in the requested drm.
The DRM mode can be set explicitely or can be chosen by the driver. To let
the driver choose the DRM mode, pass WIIPROTO_REQ_NULL placeholder to it. This
is no valid request and is replaced with an appropriate DRM.
Currently, the driver always sets the basic DRM_K mode, but this will be
extended when further peripherals like accelerometer and IR are supported.
David Herrmann [Wed, 17 Aug 2011 09:43:22 +0000 (11:43 +0200)]
HID: wiimote: Register led class devices
This registers 4 led devices to allow controlling the wiimote leds via standard
LED sysfs API. It removes the four sysfs attributes so we don't have two APIs
for one device.
David Herrmann [Wed, 17 Aug 2011 09:43:20 +0000 (11:43 +0200)]
HID: wiimote: Simplify synchronization
The new locking scheme in HID core allows us to remove a bit of synchronization.
Since the HID layer acts synchronously we simply register input core last and
there are no synchonization issues anymore.
Also register sysfs files after that to simplify the code.
Ben Skeggs [Mon, 22 Aug 2011 03:15:05 +0000 (03:15 +0000)]
drm/ttm: ensure ttm for new node is bound before calling move_notify()
This was true for new TTM_PL_SYSTEM and new TTM_PL_TT cases, but wasn't
the case on TTM_PL_SYSTEM<->TTM_PL_TT moves, which causes trouble on some
paths as nouveau's move_notify() hook requires that the dma addresses be
valid at this point.
Ben Skeggs [Mon, 22 Aug 2011 03:15:04 +0000 (03:15 +0000)]
drm/ttm: unbind ttm before destroying node in accel move cleanup
Nouveau makes the assumption that if a TTM is bound there will be a mm_node
around for it and the backwards ordering here resulted in a use-after-free
on some eviction paths.
Jeremiah Matthey [Tue, 23 Aug 2011 07:44:30 +0000 (09:44 +0200)]
HID: usbhid: Add support for SiGma Micro chip
Patch to add SiGma Micro-based keyboards (1c4f:0002) to hid-quirks.
These keyboards dont seem to allow the records to be initialized, and hence a
timeout occurs when the usbhid driver attempts to initialize them. The patch
just adds the signature for these keyboards to the hid-quirks list with the
setting HID_QUIRK_NO_INIT_REPORTS. This removes the 5-10 second wait for the
timeout to occur.
projects / linux.git / log