10 Aug, 2010
29 commits
-
Greg prefers this to go through the trivial tree.
http://lkml.org/lkml/2010/6/24/1There are about 2500 void functions in drivers/usb
Only a few used return; at end of function.Standardize them a bit.
Moved a statement down a line in drivers/usb/host/u132-hcd.c
Signed-off-by: Joe Perches
Signed-off-by: Jiri Kosina -
Conflicts:
arch/arm/mach-omap1/board-nokia770.c -
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (59 commits)
igbvf.txt: Add igbvf Documentation
igb.txt: Add igb documentation
e100/e1000*/igb*/ixgb*: Add missing read memory barrier
ixgbe: fix build error with FCOE_CONFIG without DCB_CONFIG
netxen: protect tx timeout recovery by rtnl lock
isdn: gigaset: use after free
isdn: gigaset: add missing unlock
solos-pci: Fix race condition in tasklet RX handling
pkt_sched: Fix sch_sfq vs tcf_bind_filter oops
net: disable preemption before call smp_processor_id()
tcp: no md5sig option size check bug
iwlwifi: fix locking assertions
iwlwifi: fix TX tracer
isdn: fix information leak
net: Fix napi_gro_frags vs netpoll path
usbnet: remove noisy and hardly useful printk
rtl8180: avoid potential NULL deref in rtl8180_beacon_work
ath9k: Remove myself from the MAINTAINERS list
libertas: scan before assocation if no BSSID was given
libertas: fix association with some APs by using extended rates
... -
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide-2.6:
tx493xide: use min_t() macro instead of min()
drivers/ide: Use memdup_user
via82cxxx: fix typo for VT6415 PCIE PATA IDE Host Controller support.
ide-cd: Do not access completed requests in the irq handler -
* 'rs485fix' of git://www.jni.nu/cris:
CRIS: ioctl for getting RS485 information -
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc: fix build with make 3.82
Revert "Input: appletouch - fix integer overflow issue"
memblock: Fix memblock_is_region_reserved() to return a boolean
powerpc: Trim defconfigs
powerpc: fix i8042 module build error
sound/soc: mpc5200_psc_ac97: Use gpio pins for cold reset
powerpc/5200: add mpc5200_psc_ac97_gpio_reset -
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx: (30 commits)
DMAENGINE: at_hdmac: locking fixlet
DMAENGINE: pch_dma: kill another usage of __raw_{read|write}l
dma: dmatest: fix potential sign bug
ioat2: catch and recover from broken vtd configurations v6
DMAENGINE: add runtime slave control to COH 901 318 v3
DMAENGINE: add runtime slave config to DMA40 v3
DMAENGINE: generic slave channel control v3
dmaengine: Driver for Topcliff PCH DMA controller
intel_mid: Add Mrst & Mfld DMA Drivers
drivers/dma: Eliminate a NULL pointer dereference
dma/timb_dma: compile warning on 32 bit
DMAENGINE: ste_dma40: support older silicon
DMAENGINE: ste_dma40: support disabling physical channels
DMAENGINE: ste_dma40: no disabled phy channels on ux500
DMAENGINE: ste_dma40: fix suspend bug
DMAENGINE: ste_dma40: add DB8500 memcpy channels
DMAENGINE: ste_dma40: no flow control on memcpy
DMAENGINE: ste_dma40: arch updates for LCLA and LCPA
DMAENGINE: ste_dma40: allocate LCLA dynamically
DMAENGINE: ste_dma40: no premature stop
...Fix up trivial conflicts in arch/arm/mach-ux500/devices-db8500.c
-
According include/linux/console_struct.h,vc_scr_end is unsigned long.
struct vc_data {
unsigned short vc_num; /* Console number */
unsigned int vc_cols; /* [#] Console size */
unsigned int vc_rows;
unsigned int vc_size_row; /* Bytes per row */
unsigned int vc_scan_lines; /* # of scan lines */
unsigned long vc_origin; /* [!] Start of real screen */
unsigned long vc_scr_end; /* [!] End of real screen */
unsigned long vc_visible_origin; /* [!] Top of visible window */
unsigned int vc_top, vc_bottom; /* Scrolling region */
const struct consw *vc_sw;
unsigned short *vc_screenbuf;
...
}Signed-off-by: qiaochong
Cc: Greg KH
Cc: Alan Cox
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
kernel will die on some platform when switch from vga mode to framebuffer
mode. The reason of this bug is that bind_con_driver reset vc->vc_origin
to (unsigned long)vc->vc_screenbuf.On vgacon vc->vc_origin is not releated to vc->vc_screenbuf,if set
vc->vc_origin to vc->vc_screenbuf,kernel will die on vc_do_resize.static int vc_do_resize(struct tty_struct *tty, struct tty_struct *real_tty,
struct vc_data *vc, unsigned int cols, unsigned int lines)
{
unsigned long old_origin, new_origin, new_scr_end, rlth, rrem, err = 0;
unsigned int old_cols, old_rows, old_row_size, old_screen_size;
unsigned int new_cols, new_rows, new_row_size, new_screen_size;
unsigned int end, user;
...
end = (old_rows > new_rows) ? old_origin +
(old_row_size * new_rows) :
vc->vc_scr_end;...
/*
here for a test from vgacon to framebuffer:
old_origin=0x810814a0,end=0xb00b8fa0,vc->vc_origin=0x810814a0
the code bellow will copy memory from 0x810814a0 to 0xb00b8fa0,
this will cover kernel code,kernel died here.
*/while (old_origin < end) {
scr_memcpyw((unsigned short *) new_origin,
(unsigned short *) old_origin, rlth);
if (rrem)
scr_memsetw((void *)(new_origin + rlth),
vc->vc_video_erase_char, rrem);
old_origin += old_row_size;
new_origin += new_row_size;
}...
}[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: qiaochong
Cc: Greg KH
Cc: Alan Cox
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Allow device probing to recognise the Fintek F71808E.
Sysfs interface:
* Fan/pwm control is the same as for F71889FG
* Temperature and voltage sensor handling is largely the same as for
the F71889FG
- Has one temperature sensor less (doesn't have temp3)
- Misses one voltage sensor (doesn't have V6, thus in6_input refers to
what in7_input refers for F71889FG)For the purpose of the sysfs interface fxxxx_in_temp_attr[] is split up
such that it can largely be reused.Signed-off-by: Giel van Schijndel
Cc: Jean Delvare
Cc: Hans de Goede
Cc: Jonathan Cameron
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
register_hotcpu_notifier() is designed to make these ifdefs unnecessary.
Cc: Chen Gong
Cc: Rudolf Marek
Cc: Huaxu Wan
Cc: Jean Delvare
Cc: Guenter Roeck
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Update coretemp supported CPU TjMax lists and some cleanup work.
Signed-off-by: Chen Gong
Cc: Rudolf Marek
Cc: Huaxu Wan
Cc: Jean Delvare
Cc: Guenter Roeck
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If one coretemp device can't be added, it should allow subsequent adding
operation because every new-added device will create a new sysfs group,
not an additional sensor sys entry.Signed-off-by: Chen Gong
Cc: Rudolf Marek
Cc: Huaxu Wan
Cc: Jean Delvare
Cc: Guenter Roeck
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix two errors in hotplug. One is for hotplug notifier. The other is
unnecessary driver unregister. Because even none of online cpus supports
coretemp, we can't assume new onlined cpu doesn't support it either. If
related driver is unregistered there we have no chance to use coretemp
from then on.Signed-off-by: Chen Gong
Cc: Rudolf Marek
Cc: Huaxu Wan
Cc: Jean Delvare
Cc: Guenter Roeck
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Guenter Roeck
Cc: Jean Delvare
Cc: Randy Dunlap
Cc: Hans de Goede
Cc: Mark Brown
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This driver adds support for the monitoring features of the Summit
Microelectronics SMM665 Six-Channel Active DC Output Controller/Monitor.Signed-off-by: Guenter Roeck
Acked-by: Jonathan Cameron
Cc: Jean Delvare
Cc: Hans de Goede
Cc: Mark Brown
Cc: Samuel Ortiz
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This driver will report the heading values in degrees to the sysfs
interface. The values returned are headings . e.g. 245.6Alan: Cleanups requested now all folded in and a sysfs description to keep
Andrew happy. The sysfs description now resembles hwmon.Signed-off-by: Kalhan Trisal
Reviewed-by: Jean Delvare
Signed-off-by: Alan Cox
Cc: Jonathan Cameron
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add Lenovo Thinkpad T400. I have done the testing on my laptop. The
hdaps module detects the device and the hdapsd daemon is able to [un]park
the disk.Signed-off-by: Ritesh Raj Sarraf
Cc: Frank Seidel
Cc: Dmitry Torokhov
Cc: Jean Delvare
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The error may happen at any iteration of the for loop, this patch properly
unregisters already registed edd_devices in error path.[akpm@linux-foundation.org: remove unneeded NULL test]
Signed-off-by: Axel Lin
Cc: Stephen Hemminger
Cc: Jens Axboe
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The previous change added WARN_ON() in misc_deregister(). So it is not
necessary to WARN_ON() misc_deregister() failure by callers.Signed-off-by: Akinobu Mita
Cc: Martin Schwidefsky
Cc: Heiko Carstens
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
misc_deregister() returns an error only when it attempts to unregister
the device that is not registered. This is the driver's bug.Most of the drivers don't check the return value of misc_deregister().
(It is not bad thing because most of kernel *_unregister() API always
succeed and do not return value)So it is better to indicate the error by WARN_ON() in misc_deregister().
Signed-off-by: Akinobu Mita
Cc: Heiko Carstens
Cc: Martin Schwidefsky
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This driver adds support for the BMP085 digital pressure sensor from Bosch
Sensortec. It exposes a sysfs api to userspace where pressure and
temperature measurement results can be read from the pressure0_input and
temp0_input file. The chip is able to calculate the average of up to
eight samples to increase the accuracy. This feature can be controlled by
writing to the oversampling file.The BMP085 digital pressure sensor can measure ambient air pressure and
temperature. Both values can be obtained from sysfs files. The pressure
is measured by reading from pressure0_input. Valid values range from
30000 to 110000 pascal with a resolution of 1 pascal (=0.01 millibar).temp0_input holds the current temperature in degree celsius, multiplied by
10. This results in a resolution of a tenth degree celsius. Values range
from -400 to 850.To increase the accuracy, this chip can calculate the average of 1, 2, 4
or 8 samples. This behavior is controlled through the oversampling sysfs
file. Two to the power of the value written to that file specifies how
many samples will be used. Valid values: 0..3.[akpm@linux-foundation.org: fix typo]
[shubhrajyoti@ti.com: optimize the wait time for the pressure sensor, definition of long is arch dependent so make it u32]
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Christoph Mair
Signed-off-by: Shubhrajyoti D
Acked-by: Jonathan Cameron
Cc: Stefan Schmidt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix i386 PAE compile warning:
drivers/misc/hpilo.c: In function `ilo_ccb_setup':
drivers/misc/hpilo.c:274: warning: cast to pointer from integer of different sizedma_addr_t is 64 on i386 PAE which causes a size mismatch.
Signed-off-by: Prarit Bhargava
Acked-by: David Altobelli
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add support for ROHM BH1780GLI Ambient light sensor.
BH1780 supports I2C interface. Driver supports read/update of power state
and read of lux value (through SYSFS). Writing value 3 to power_state
enables the sensor and current lux value could be read.Currently this driver follows the same sysfs convention as supported by
drivers/misc/isl29003.c.Signed-off-by: Hemanth V
Reviewed-by: Daniel Mack
Acked-by: Jonathan Cameron
Cc: Jean Delvare
Cc: Wolfram Sang
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
On some SoC chips, HW resources may be in use during any particular idle
period. As a consequence, the cpuidle states that the SoC is safe to
enter can change from idle period to idle period. In addition, the
latency and threshold of each cpuidle state can vary, depending on the
operating condition when the CPU becomes idle, e.g. the current cpu
frequency, the current state of the HW blocks, etc.cpuidle core and the menu governor, in the current form, are geared
towards cpuidle states that are static, i.e. the availabiltiy of the
states, their latencies, their thresholds are non-changing during run
time. cpuidle does not provide any hook that cpuidle drivers can use to
adjust those values on the fly for the current idle period before the menu
governor selects the target cpuidle state.This patch extends cpuidle core and the menu governor to handle states
that are dynamic. There are three additions in the patch and the patch
maintains backwards-compatibility with existing cpuidle drivers.1) add prepare() to struct cpuidle_device. A cpuidle driver can hook
into the callback and cpuidle will call prepare() before calling the
governor's select function. The callback gives the cpuidle driver a
chance to update the dynamic information of the cpuidle states for the
current idle period, e.g. state availability, latencies, thresholds,
power values, etc.2) add CPUIDLE_FLAG_IGNORE as one of the state flags. In the prepare()
function, a cpuidle driver can set/clear the flag to indicate to the
menu governor whether a cpuidle state should be ignored, i.e. not
available, during the current idle period.3) add power_specified bit to struct cpuidle_device. The menu governor
currently assumes that the cpuidle states are arranged in the order of
increasing latency, threshold, and power savings. This is true or can
be made true for static states. Once the state parameters are dynamic,
the latencies, thresholds, and power savings for the cpuidle states can
increase or decrease by different amounts from idle period to idle
period. So the assumption of increasing latency, threshold, and power
savings from Cn to C(n+1) can no longer be guaranteed.It can be straightforward to calculate the power consumption of each
available state and to specify it in power_usage for the idle period.
Using the power_usage fields, the menu governor then selects the state
that has the lowest power consumption and that still satisfies all other
critieria. The power_specified bit defaults to 0. For existing cpuidle
drivers, cpuidle detects that power_specified is 0 and fills in a dummy
set of power_usage values.Signed-off-by: Ai Li
Cc: Len Brown
Acked-by: Arjan van de Ven
Cc: Ingo Molnar
Cc: Venkatesh Pallipadi
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
drivers/base/node.c: In function 'node_read_meminfo':
drivers/base/node.c:139: warning: the frame size of 848 bytes is
larger than 512 bytesFix it by splitting the sprintf() into three parts. It has no functional
change.Signed-off-by: KOSAKI Motohiro
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix a build failure "error: void value not ignored as it ought to be"
by removing an assignment of a void return value. The functionality of
the code is not changed.Signed-off-by: Peter Huewe
Acked-by: Henrik Kretzschmar
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
After the commit that changed ipmi_si detecting sequence from SMBIOS/ACPI
to ACPI/SMBIOS,| commit 754d453185275951d39792865927ec494fa1ebd8
| Author: Matthew Garrett
| Date: Wed May 26 14:43:47 2010 -0700
|
| ipmi: change device discovery order
|
| The ipmi spec provides an ordering for si discovery. Change the driver to
| match, with the exception of preferring smbios to SPMI as HPs (at least)
| contain accurate information in the former but not the latter.ipmi_si can not be initialized.
[ 138.799739] calling init_ipmi_devintf+0x0/0x109 @ 1
[ 138.805050] ipmi device interface
[ 138.818131] initcall init_ipmi_devintf+0x0/0x109 returned 0 after 12797 usecs
[ 138.822998] calling init_ipmi_si+0x0/0xa90 @ 1
[ 138.840276] IPMI System Interface driver.
[ 138.846137] ipmi_si: probing via ACPI
[ 138.849225] ipmi_si 00:09: [io 0x0ca2] regsize 1 spacing 1 irq 0
[ 138.864438] ipmi_si: Adding ACPI-specified kcs state machine
[ 138.870893] ipmi_si: probing via SMBIOS
[ 138.880945] ipmi_si: Adding SMBIOS-specified kcs state machineipmi_si: duplicate interface
[ 138.896511] ipmi_si: probing via SPMI
[ 138.899861] ipmi_si: Adding SPMI-specified kcs state machineipmi_si: duplicate interface
[ 138.917095] ipmi_si: Trying ACPI-specified kcs state machine at i/o address 0xca2, slave address 0x0, irq 0
[ 138.928658] ipmi_si: Interface detection failed
[ 138.953411] initcall init_ipmi_si+0x0/0xa90 returned 0 after 110847 usecsin smbios has
DMI/SMBIOS
Handle 0x00C5, DMI type 38, 18 bytes
IPMI Device Information
Interface Type: KCS (Keyboard Control Style)
Specification Version: 2.0
I2C Slave Address: 0x00
NV Storage Device: Not Present
Base Address: 0x0000000000000CA2 (I/O)
Register Spacing: 32-bit Boundaries
in DSDT has
Device (BMC)
{Name (_HID, EisaId ("IPI0001"))
Method (_STA, 0, NotSerialized)
{
If (LEqual (OSN, Zero))
{
Return (Zero)
}Return (0x0F)
}Name (_STR, Unicode ("IPMI_KCS"))
Name (_UID, Zero)
Name (_CRS, ResourceTemplate ()
{
IO (Decode16,
0x0CA2, // Range Minimum
0x0CA2, // Range Maximum
0x00, // Alignment
0x01, // Length
)
IO (Decode16,
0x0CA6, // Range Minimum
0x0CA6, // Range Maximum
0x00, // Alignment
0x01, // Length
)
})
Method (_IFT, 0, NotSerialized)
{
Return (One)
}Method (_SRV, 0, NotSerialized)
{
Return (0x0200)
}
}so the reg spacing should be 4 instead of 1.
Try to calculate regspacing for this kind of system.
Observed on a Sun Fire X4800. Other OSes work and pass certification.
Signed-off-by: Yinghai Lu
Cc: Bjorn Helgaas
Acked-by: Matthew Garrett
Cc: Len Brown
Cc: Myron Stowe
Cc: Corey Minyard
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
drm: fix fallouts from slow-work -> wq conversion
workqueue: workqueue_cpu_callback() should be cpu_notifier instead of hotcpu_notifier
workqueue: add missing __percpu markup in kernel/workqueue.c
09 Aug, 2010
11 commits
-
Add ioctl to CRIS serial driver to get RS485 data from user-space.
Signed-off-by: Claudio Scordino
Signed-off-by: Jesper Nilsson -
Commit 991ea75c (drm: use workqueue instead of slow-work), which made
drm to use wq instead of slow-work, didn't account for the return
value difference between delayed_slow_work_enqueue() and
queue_delayed_work(). The former returns 0 on success and -errno on
failures while the latter never fails and only uses the return value
to indicate whether the work was already pending or not.This misconversion triggered spurious error messages. Remove the now
unnecessary return value check and error message.Markus: caught another incorrect conversion in drm_kms_helper_poll_enable()
Signed-off-by: Tejun Heo
Reported-by: Markus Trippelsdorf
Tested-by: Markus Trippelsdorf
Cc: David Airlie
Cc: dri-devel@lists.freedesktop.org -
This fixes a warning ("comparison of distinct pointer types lacks a
cast") introduced by the commit
040f6b4f14adb2ca5babb84e9fb2ebc6661e0be2 ("tx493xide: use ->pio_mode
value to determine pair device speed").Signed-off-by: Atsushi Nemoto
Signed-off-by: David S. Miller -
Use memdup_user when user data is immediately copied into the
allocated region.The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)//
@@
expression from,to,size,flag;
position p;
identifier l1,l2;
@@- to = \(kmalloc@p\|kzalloc@p\)(size,flag);
+ to = memdup_user(from,size);
if (
- to==NULL
+ IS_ERR(to)
|| ...) {
}
- if (copy_from_user(to, from, size) != 0) {
-
- }
//Signed-off-by: Julia Lawall
Signed-off-by: David S. Miller -
Without this fix, init of the via82cxxx driver causes a oops with a
stack resembling the one below, and the boot blocks between init of
USB devices and launch of init (was easy to bisect by booting with
init=/bin/sh).Pid: 279, comm: work_for_cpu Not tainted 2.6.34.1-00003-ga42ea77 #2
Call Trace:
[] ? warn_slowpath_common+0x76/0x8c
[] ? warn_slowpath_fmt+0x40/0x45
[] ? printk+0x40/0x47
[] ? enable_irq+0x3e/0x64
[] ? ide_probe_port+0x55c/0x589 [ide_core]
[] ? ide_host_register+0x273/0x628 [ide_core]
[] ? ide_pci_init_two+0x4da/0x5c5 [ide_core]
[] ? up+0xe/0x36
[] ? release_console_sem+0x17e/0x1ae
[] ? klist_iter_exit+0x14/0x1e
[] ? bus_find_device+0x75/0x83
[] ? via_init_one+0x269/0x28a [via82cxxx]
[] ? init_chipset_via82cxxx+0x0/0x1ea [via82cxxx]
[] ? do_work_for_cpu+0x0/0x1b
[] ? local_pci_probe+0x12/0x16
[] ? do_work_for_cpu+0xb/0x1b
[] ? kthread+0x75/0x7d
[] ? kernel_thread_helper+0x4/0x10
[] ? kthread+0x0/0x7d
[] ? kernel_thread_helper+0x0/0x10
---[ end trace 89c8cb70379b5bda ]---The typo was introduced in a354ae8747d0687093ce244e76b15b6174d2f098,
and affects 2.6.33-rc4 and later.Signed-off-by: Yann Dirson
Signed-off-by: David S. Miller -
ide_cd_error_cmd() can complete an erroneous request with leftover
buffers. Signal this with its return value so that the request is not
accessed after its completion in the irq handler and we oops.Cc: # 32.x 33.x 34.x
Signed-off-by: Borislav Petkov
Signed-off-by: David S. Miller -
Signed-off-by: Paul Bolle
Signed-off-by: Jiri Kosina -
To avoid more patches, I also fixed other spelling
and grammar bugs when they were in the same or
following line:successfull -> successful
parse -> parses
controler -> controller
controlers -> controllersCc: Jiri Kosina
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Stefan Weil
Signed-off-by: Jiri Kosina -
This reverts commit 04b4b88cca0ebe3813b4b6f014fb6a0db380b137.
While the original problem only caused a slight disturbance on the
edge of the touchpad, the commit above to "fix" it completely breaks
operation on some other models such as mine.We'll sort this out separately, revert the patch for now.
Signed-off-by: Benjamin Herrenschmidt
-
Based on patches from Sonny Rao and Milton Miller...
Combined the patches to fix up clean_tx_irq and clean_rx_irq.
The PowerPC architecture does not require loads to independent bytes
to be ordered without adding an explicit barrier.In ixgbe_clean_rx_irq we load the status bit then load the packet data.
With packet split disabled if these loads go out of order we get a
stale packet, but we will notice the bad sequence numbers and drop it.The problem occurs with packet split enabled where the TCP/IP header
and data are in different descriptors. If the reads go out of order
we may have data that doesn't match the TCP/IP header. Since we use
hardware checksumming this bad data is never verified and it makes it
all the way to the application.This bug was found during stress testing and adding this barrier has
been shown to fix it. The bug can manifest as a data integrity issue
(bad payload data) or as a BUG in skb_pull().This was a nasty bug to hunt down, if people agree with the fix I think
it's a candidate for stable.Previously Submitted to e1000-devel only for ixgbe
http://marc.info/?l=e1000-devel&m=126593062701537&w=3
We've now seen this problem hit with other device drivers (e1000e mostly)
So I'm resubmitting with fixes for other Intel Device Drivers with
similar issues.CC: Milton Miller
CC: Anton Blanchard
CC: Sonny Rao
CC: stable
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller -
Building ixgbe without DCB_CONFIG and FCOE_CONFIG will cause
a build error. This resolves the build error by wrapping
the fcoe.up in CONFIG_IXGBE_DCB ifdefs.Also frames were being priority VLAN tagged even without DCB
enabled. This fixes this so that 8021Q priority tags are
only added with DCB actually enabled.Reported-by: divya
Reported-by: Jon Mason
Signed-off-by: John Fastabend
Tested-by: Stephen Ko
Tested-by: Ross Brattain
Signed-off-by: Jeff Kirsher
Signed-off-by: David S. Miller
