linux.git
16 years agoMXC MMC: Remove custom error codes
Quinn Jensen [Thu, 25 Oct 2007 03:23:08 +0000 (21:23 -0600)]
MXC MMC: Remove custom error codes

Patch to remove custom error codes in MX MMC driver.
Replace MXC-specific error codes with generic kernel error codes
in the MMC driver.  Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-MXC-MMC-Remove-custom-error-codes.patch

16 years agoBugzilla 510: Update to code review fixes following L2618.2
Quinn Jensen [Thu, 25 Oct 2007 03:23:01 +0000 (21:23 -0600)]
Bugzilla 510: Update to code review fixes following L2618.2

Patch for Bugzilla 510: Update to code review fixes following L2618.2
Changes made to fix bugs and/or port to 2.6.19.2 after including GSO
L2618-2 patches as discovered and discussed in the code review.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-510-Update-to-code-review-fixes-f.patch

16 years agoUpdate drivers, such as dpm, dvfs, ide, and sir to use the new,
Quinn Jensen [Thu, 25 Oct 2007 03:22:56 +0000 (21:22 -0600)]
Update drivers, such as dpm, dvfs, ide, and sir to use the new,

Patch to update drivers, such as dpm, dvfs, ide, and sir to use the new,
generic clock API's.  Remove clock-gating calls from the gpio active and
inactive functions.  Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Update-drivers-such-as-dpm-dvfs-ide-and-si.patch

16 years agoBug fix of Bugzilla 363 to enable ATA subsystem clock before
Quinn Jensen [Thu, 25 Oct 2007 03:22:50 +0000 (21:22 -0600)]
Bug fix of Bugzilla 363 to enable ATA subsystem clock before

Bug fix of Bugzilla 363 to enable ATA subsystem clock before writing to
ATA subsystem on the linux 2.6.22 kernel for MX31/32 platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bug-fix-of-Bugzilla-363-to-enable-ATA-subs.patch

16 years agoWork-around compile breakage for mx27 on ov2640 camera driver
Quinn Jensen [Thu, 25 Oct 2007 03:22:45 +0000 (21:22 -0600)]
Work-around compile breakage for mx27 on ov2640 camera driver

Patch to work around compile breakage for mx27 on ov2640 camera driver.
Applies to linux 2.6.22 kernel on MX27 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Work-around-compile-breakage-for-mx27-on-o.patch

16 years agoBugzilla 533 - Fix "multi-times" message and prevent extra call to gpio_uart_active()
Quinn Jensen [Thu, 25 Oct 2007 03:22:40 +0000 (21:22 -0600)]
Bugzilla 533 - Fix "multi-times" message and prevent extra call to gpio_uart_active()

Patch for Bugzilla 533 - Fix "multi-times" message and prevent extra
call to gpio_uart_active() in the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-533-Fix-multi-times-message-and-p.patch

16 years agoWorkaround for IAPI sleep issue - spin in iapi_lowSynchChannel
Quinn Jensen [Thu, 25 Oct 2007 03:22:35 +0000 (21:22 -0600)]
Workaround for IAPI sleep issue - spin in iapi_lowSynchChannel

Patch to work around for IAPI sleep issue - spin in iapi_lowSynchChannel.
Change IAPI iapi_lowSynchChannel() to spin instead of sleep. This is
needed for the LIBATA pata_fsl driver, to avoid a "sheduling while
atomic" oops.  Implications on other MXC drivers and platforms is not
known.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Workaround-for-IAPI-sleep-issue-spin-in-ia.patch

16 years agoPrototype FSL PATA LIBATA driver
Quinn Jensen [Thu, 25 Oct 2007 03:22:29 +0000 (21:22 -0600)]
Prototype FSL PATA LIBATA driver

Add Prototype FSL PATA LIBATA driver to linux 2.6.22 kernel
for MX platforms. Working in PIO mode on MX31 and MX27.
Incompatible with current SDMA IAPI due to a sleep in
the DMA configuration setup path.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Prototype-FSL-PATA-LIBATA-driver.patch

16 years agoCommunity MMC/SD/SDIO support from Pierre Ossman's tree
Quinn Jensen [Thu, 25 Oct 2007 03:22:22 +0000 (21:22 -0600)]
Community MMC/SD/SDIO support from Pierre Ossman's tree

Patch to add Community MMC/SD/SDIO support from Pierre Ossman's tree
git://git.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc.git
Author: Pierre Ossman
Snapshot taken 20070910, and applied to linux 2.6.22.6 by Quinn Jensen.
Portions Copyright Pierre Ossman and Monta Vista.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Community-MMC-SD-SDIO-support-from-Pierre-.patch

16 years agoBackport to 2.6.22.6 of MMC/SD/SDIO updates in 2.6.23-rc3
Quinn Jensen [Thu, 25 Oct 2007 03:22:17 +0000 (21:22 -0600)]
Backport to 2.6.22.6 of MMC/SD/SDIO updates in 2.6.23-rc3

Backport of MMC/SD/SDIO updates from the linux 2.6.23-rc3 kernel to the
2.6.22.6 kernel.  This patch created from mmc diffs between linus git
2.6.22 and 2.6.23-rc3.  Sources are available from kernel.org.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Backport-to-2.6.22.6-of-MMC-SD-SDIO-update.patch

16 years agoCR ENGR00049086: fix scc fail fuse not blown
Quinn Jensen [Thu, 25 Oct 2007 03:22:11 +0000 (21:22 -0600)]
CR ENGR00049086: fix scc fail fuse not blown

Patch for CR ENGR00049086: fix scc fail fuse not blown.
Deregistering the driver properly to prevent boot crash if SCC fuse are
not blown.  Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00049086-fix-scc-fail-fuse-not-blow.patch

16 years agoCR ENGR00043063: platform device for scc added
Quinn Jensen [Thu, 25 Oct 2007 03:22:06 +0000 (21:22 -0600)]
CR ENGR00043063: platform device for scc added

Patch for CR ENGR00043063: add platform device support for scc.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00043063-platform-device-for-scc-ad.patch

16 years agoCR ENGR00047179: fix disable console on mxc uart breaks build
Quinn Jensen [Thu, 25 Oct 2007 03:22:00 +0000 (21:22 -0600)]
CR ENGR00047179: fix disable console on mxc uart breaks build

Patch for CR ENGR00047179: fix disable console on mxc uart breaks build
when console support on mxc uart is disabled.  Applies to linux 2.6.22
kernel on MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00047179-fix-disable-console-on-mxc.patch

16 years agoBugzilla 627. Prevent DPM from calling DVFS when DVFS is not enabled.
Quinn Jensen [Thu, 25 Oct 2007 03:21:55 +0000 (21:21 -0600)]
Bugzilla 627. Prevent DPM from calling DVFS when DVFS is not enabled.

Patch for Bugzilla 627. Prevent DPM from calling DVFS when DVFS is not
enabled on the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-627.-Prevent-DPM-from-calling-DVF.patch

16 years agoPort MXC FIR driver forward to 2.6.22
Quinn Jensen [Thu, 25 Oct 2007 03:21:49 +0000 (21:21 -0600)]
Port MXC FIR driver forward to 2.6.22

Port MXC FIR driver forward to 2.6.22 for MX3 platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Port-MXC-FIR-driver-forward-to-2.6.22.patch

16 years agoBugzilla 511: Redefine the min macro #defined in yaffs_qsort to yaffs_min
Quinn Jensen [Thu, 25 Oct 2007 03:21:44 +0000 (21:21 -0600)]
Bugzilla 511: Redefine the min macro #defined in yaffs_qsort to yaffs_min

Patch for Bugzilla 511: Redefine the min macro #defined in yaffs_qsort to yaffs_min
so that it doesn't conflict with the macro defined in linux/kernel.h of the linux
2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-511-Redefine-the-min-macro-define.patch

16 years agoAdd yaffs2 filesystem.
Quinn Jensen [Thu, 25 Oct 2007 03:21:39 +0000 (21:21 -0600)]
Add yaffs2 filesystem.

Add yaffs2 filesystem support to the linux 2.6.22 kernel.
Downloaded from the Aleph One Limited website at:
  http://www.aleph1.co.uk/cgi-bin/viewcvs.cgi/root.tar.gz?view=tar
Integrated into the kernel using patch-ker.sh.
Copyright (C) 2002-2007 Aleph One Ltd.
  for Toby Churchill Ltd and Brightstar Engineering

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Add-yaffs2-filesystem.patch

16 years agoBugzilla 630. Allow rng registers to be read and written for testing
Quinn Jensen [Thu, 25 Oct 2007 03:21:33 +0000 (21:21 -0600)]
Bugzilla 630. Allow rng registers to be read and written for testing

Patch for Bugzilla 630. Allow rng registers to be read and written for
testing purposes on the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-630.-Allow-rng-registers-to-be-re.patch

16 years agoMake the latency hiding option mx27 specific since it could
Quinn Jensen [Thu, 25 Oct 2007 03:21:28 +0000 (21:21 -0600)]
Make the latency hiding option mx27 specific since it could

Patch to mke the latency hiding option MX27 specific since it could
break other platforms that have a VPU.  Applies to linux 2.6.22
kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Make-the-latency-hiding-option-mx27-specif.patch

16 years agoFixed coding errors in the mxc v4l2 output driver.
Quinn Jensen [Thu, 25 Oct 2007 03:21:23 +0000 (21:21 -0600)]
Fixed coding errors in the mxc v4l2 output driver.

Patch to fix coding errors in the mxc v4l2 output driver.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Fixed-coding-errors-in-the-mxc-v4l2-output.patch

16 years agoCR ENGR00038129: Add vpu latency-hiding control. Fixes an mpeg4 pending
Quinn Jensen [Thu, 25 Oct 2007 03:21:18 +0000 (21:21 -0600)]
CR ENGR00038129: Add vpu latency-hiding control. Fixes an mpeg4 pending

Patch for CR ENGR00038129: Add vpu latency-hiding control. Fixes an
mpeg4 pending issue on MX27 TO2.  Applies to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00038129-Add-vpu-latency-hiding-con.patch

16 years agoCR 34439732: Camera and V4L2 improvements
Quinn Jensen [Thu, 25 Oct 2007 03:21:12 +0000 (21:21 -0600)]
CR 34439732: Camera and V4L2 improvements

Patch for CR 34439732: Camera and V4L2 improvements.
Add flicker control, white balance, automatic exposure support, and tear
prevention to the iMagic camera driver on the linux 2.6.22 kernel for MX
platforms.
Ported to 2.6.22 by Ross Wille

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-34439732-Camera-and-V4L2-improvements.patch

16 years agoCR ENGR00033520: Add a new algorithm to the PP and PRP drivers that
Quinn Jensen [Thu, 25 Oct 2007 03:21:07 +0000 (21:21 -0600)]
CR ENGR00033520: Add a new algorithm to the PP and PRP drivers that

Patch for CR ENGR00033520: Add a new algorithm to the PP and PRP drivers
that determines more accurate scaling factors.  Applies to linux 2.6.22
kernel for the MX27 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00033520-Add-a-new-algorithm-to-the.patch

16 years agoCR ENGR37631: Adds correct signal handling to the V4L2 input and output
Quinn Jensen [Thu, 25 Oct 2007 03:21:01 +0000 (21:21 -0600)]
CR ENGR37631: Adds correct signal handling to the V4L2 input and output

Patch for CR ENGR37631: Adds correct signal handling to the V4L2 input
and output devices. Adds handling of PRP overflow on channel2 under high
system load.  Applies to linux 2.6.22 kernel on MX27 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR37631-Adds-correct-signal-handling-.patch

16 years agoCR 34440439: Under heavy load v4l2 ioctls sometimes return error (EINTR
Quinn Jensen [Thu, 25 Oct 2007 03:20:56 +0000 (21:20 -0600)]
CR 34440439: Under heavy load v4l2 ioctls sometimes return error (EINTR

Patch for CR 34440439: Under heavy load v4l2 ioctls sometimes return
error (EINTR or ETIME). This patch fixes vpu v4l2 interrupt code to be
more robust.  Applies to linux 2.6.22 kernel for MX platforms.
Ported to 2.6.22 by Ross Wille

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-34440439-Under-heavy-load-v4l2-ioctls-s.patch

16 years agoCR ENGR00037631: Add handling of prp overflow on channel 2 under high system load.
Quinn Jensen [Thu, 25 Oct 2007 03:20:51 +0000 (21:20 -0600)]
CR ENGR00037631: Add handling of prp overflow on channel 2 under high system load.

Patch for CR ENGR00037631: Add handling of prp overflow on channel 2
under high system load.  Applies to linux 2.6.22 kernel for the MX27
platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-ENGR00037631-Add-handling-of-prp-overfl.patch

16 years agoBugzilla 562: Suspend resumes on eth activity
Quinn Jensen [Thu, 25 Oct 2007 03:20:46 +0000 (21:20 -0600)]
Bugzilla 562: Suspend resumes on eth activity

Patch for Bugzilla 562: Suspend resumes on eth activity
Patch to linux-2.6.19.2 for MX31.  When the MX31 is suspended it
was resuming on ethernet activity due to a CPLD interrupt.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-562-Suspend-resumes-on-eth-activi.patch

16 years agoBugzilla 43: Fix FIR on MX31
Quinn Jensen [Thu, 25 Oct 2007 03:20:39 +0000 (21:20 -0600)]
Bugzilla 43: Fix FIR on MX31

Patch for Bugzilla 43: Fix FIR on linux 2.6.22 kernel for MX31.
After adding FIR support, loading mxc_ir.ko fails with ENOMEM and
"Failed to request UART memory region".  This patch makes sure UART2
is left available for the FIR driver.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-43-Fix-FIR-on-MX31.patch

16 years agoBugzilla 422: Enable second MMC/SD slot for MX31.
Quinn Jensen [Thu, 25 Oct 2007 03:20:34 +0000 (21:20 -0600)]
Bugzilla 422: Enable second MMC/SD slot for MX31.

Patch for Bugzilla 422: Enable second MMC/SD slot for MX31 on the
linux 2.6.22 kernel for MX31/32.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bugzilla-422-Enable-second-MMC-SD-slot-for.patch

16 years agoCR SDCaa01608: I2C uses wrong adapter name
Quinn Jensen [Thu, 25 Oct 2007 03:20:28 +0000 (21:20 -0600)]
CR SDCaa01608: I2C uses wrong adapter name

Patch for CR SDCaa01608: I2C uses wrong adapter name.
In the mxc_i2c driver, set the adapter name to be the same as the
platform_device name.  Otherwise, drivers that use i2c may fail
when checking the mxc_i2c adapter name, such as the USB driver.
Applies to linux 2.6.22 kernel on MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-CR-SDCaa01608-I2C-uses-wrong-adapter-name.patch

16 years agoAdd interrogation of V4L2 or Linux udev to determine if no camera is present.
Quinn Jensen [Thu, 25 Oct 2007 03:20:23 +0000 (21:20 -0600)]
Add interrogation of V4L2 or Linux udev to determine if no camera is present.

Patch to add interrogation of V4L2 or Linux udev to determine if a
camera is present. Add status and detection functionality for mt9v111
camera to the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Add-interrogation-of-V4L2-or-Linux-udev-to.patch

16 years agoPatch to enable mxc framebuffer driver to be loaded/unloaded as a
Quinn Jensen [Thu, 25 Oct 2007 03:20:18 +0000 (21:20 -0600)]
Patch to enable mxc framebuffer driver to be loaded/unloaded as a

Patch to enable MXC framebuffer driver to be loaded/unloaded as a
module.  Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Patch-to-enable-mxc-framebuffer-driver-to-.patch

16 years agoPatch to fix irda mode bug.
Quinn Jensen [Thu, 25 Oct 2007 03:20:12 +0000 (21:20 -0600)]
Patch to fix irda mode bug.

Patch to fix irda mode bug.
Applies to linux 2.6.22 kernel on MX27 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Patch-to-fix-irda-mode-bug.patch

16 years agoPatch to properly reset the Focus 453 chip.
Quinn Jensen [Thu, 25 Oct 2007 03:20:06 +0000 (21:20 -0600)]
Patch to properly reset the Focus 453 chip.

Patch to properly reset the Focus 453 chip.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Patch-to-properly-reset-the-Focus-453-chip.patch

16 years agoPatch to fix dma errors when accessing IDE on MX27.
Quinn Jensen [Thu, 25 Oct 2007 03:20:00 +0000 (21:20 -0600)]
Patch to fix dma errors when accessing IDE on MX27.

Patch to fix dma errors when accessing IDE on MX27.
Applies to linux 2.6.22 kernel on MX27 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Patch-to-fix-dma-errors-when-accessing-IDE.patch

16 years agoARM926 (MX27) D-cache writethrough bug fix.
Quinn Jensen [Thu, 25 Oct 2007 03:19:54 +0000 (21:19 -0600)]
ARM926 (MX27) D-cache writethrough bug fix.

Patch for ARM926 (MX27) D-cache writethrough bug fix.
Source: linux-arm-kernel list, posted by Gilles Maigne on 11/17/06.

Text from e-mail posting:

arm926_dma_flush_range write the content of dcache in memory when linux
is configured with CONFIG_CPU_DCACHE_WRITETHROUGH ( cache in
write-through) and does not invalidate the virtual address range.
This is wrong for two reasons :
1/ the cache line does not need to be written because the cache is
write-through
2/ the cache line is not invalidated, so next read at this address may
come from the cache, which is wrong because main memory may contain
valid data resulting of a DMA transaction.
This patch applies to linux 2.6.18

http://www.bitshrine.org/gpp/linux-2.6.19.2-mx-arm926_dcache_writethrough_flush_fix.patch

http://www.bitshrine.org/gpp/linux-2.6.22-mx-ARM926-MX27-D-cache-writethrough-bug-fix.patch

16 years agoBug fix for when processor enters state-retention mode after camera
Quinn Jensen [Thu, 25 Oct 2007 03:19:49 +0000 (21:19 -0600)]
Bug fix for when processor enters state-retention mode after camera

Bug fix for when processor enters state-retention mode after camera
activity on the linux 2.6.22 kernel for the MX31 platform.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bug-fix-for-when-processor-enters-state-re.patch

16 years agoBug fix of Bugzilla 217 to ensure I2C stop timing is met, but
Quinn Jensen [Thu, 25 Oct 2007 03:19:44 +0000 (21:19 -0600)]
Bug fix of Bugzilla 217 to ensure I2C stop timing is met, but

Bug fix of Bugzilla 217 to ensure I2C stop timing is met, but also with
deadlock prevention on linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Bug-fix-of-Bugzilla-217-to-ensure-I2C-stop.patch

16 years agoAdd generic clksource and clkevent support from GSO.
Quinn Jensen [Thu, 25 Oct 2007 03:19:38 +0000 (21:19 -0600)]
Add generic clksource and clkevent support from GSO.

Patch to add generic clksource and clkevent support from GSO to the
linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Add-generic-clksource-and-clkevent-support.patch

16 years agoFix new-line problems on printk statements.
Quinn Jensen [Thu, 25 Oct 2007 03:19:31 +0000 (21:19 -0600)]
Fix new-line problems on printk statements.

Patch to fix new-line problems on printk statements.
Applies to linux 2.6.22 kernel on MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Fix-new-line-problems-on-printk-statements.patch

16 years agoUSB: Miscellaneous cleanup/porting.
Quinn Jensen [Thu, 25 Oct 2007 03:19:26 +0000 (21:19 -0600)]
USB: Miscellaneous cleanup/porting.

Patch to apply miscellaneous cleanup/porting fixes.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-USB-Miscellaneous-cleanup-porting.patch

16 years agoUSB: fix OTGSC_IE_1ms_TIMER bit definition.
Quinn Jensen [Thu, 25 Oct 2007 03:19:20 +0000 (21:19 -0600)]
USB: fix OTGSC_IE_1ms_TIMER bit definition.

Patch to fix USB OTGSC_IE_1ms_TIMER bit definition.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-USB-fix-OTGSC_IE_1ms_TIMER-bit-definition.patch

16 years agoUSB: port to 2.6.22.
Quinn Jensen [Thu, 25 Oct 2007 03:19:15 +0000 (21:19 -0600)]
USB: port to 2.6.22.

Port Freescale USB to linux kernel 2.6.22.
Switch to delayed_work() and fix irq flags.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-USB-port-to-2.6.22.patch

16 years agoUSB: disable autosuspend.
Quinn Jensen [Thu, 25 Oct 2007 03:19:10 +0000 (21:19 -0600)]
USB: disable autosuspend.

Patch to disable USB autosuspend.
Autosuspend tries to suspend the host controller even
after it's been switched over to being a gadget controller,
which doesn't do much for gadget operations.  Disable it.
Applies to linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-USB-disable-autosuspend.patch

16 years agoSony's implementation of __mxc_ioremap, ported to our kernel.
Quinn Jensen [Thu, 25 Oct 2007 03:19:04 +0000 (21:19 -0600)]
Sony's implementation of __mxc_ioremap, ported to our kernel.

Patch of Sony's implementation of __mxc_ioremap, ported to the
linux 2.6.22 MX kernel.
Copyright Sony Corporation.
Note: The contents of this patch ends up being completely
replaced by code written by Freescale.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Sony-s-implementation-of-__mxc_ioremap-por.patch

16 years agoAdd OV2640 camera support.
Quinn Jensen [Thu, 25 Oct 2007 03:18:59 +0000 (21:18 -0600)]
Add OV2640 camera support.

Add OV2640 camera support to the linux 2.6.22 kernel for MX platforms.
Ported to 2.6.22 by Ross Wille.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Add-OV2640-camera-support.patch

16 years agoAdd MX27, MX31, and MX32 splash screens to kernel.
Quinn Jensen [Thu, 25 Oct 2007 03:18:52 +0000 (21:18 -0600)]
Add MX27, MX31, and MX32 splash screens to kernel.

Add MX27, MX31, and MX32 splash screens to the linux 2.6.22 kernel.
Graphics images created by Ross Wille
Tux logo created by Larry Ewing (http://www.isc.tamu.edu/~lewing/linux/)

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Add-MX27-MX31-and-MX32-splash-screens-to-2.patch

16 years agoAdditional porting changes as a result of code review.
Quinn Jensen [Thu, 25 Oct 2007 03:18:46 +0000 (21:18 -0600)]
Additional porting changes as a result of code review.

Minor porting fixes resulting from code review. Does not affect functionality.
Patch applies to the linux 2.6.22 kernel for MX platforms.
1. Make SAHARA IRQ comment inclusive of all MX2 chips.
2. Remove BREAKPT_FREQ definition, since it is unused.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-Additional-porting-changes-as-a-result-of-.patch

16 years agoThis patch adds ALSA sound to the linux 2.6.22 kernel for MX platforms.
Quinn Jensen [Thu, 25 Oct 2007 03:18:41 +0000 (21:18 -0600)]
This patch adds ALSA sound to the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-sound.patch

16 years agoThis patch adds MX platform-specific files to the linux 2.6.22 kernel.
Quinn Jensen [Thu, 25 Oct 2007 03:18:35 +0000 (21:18 -0600)]
This patch adds MX platform-specific files to the linux 2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-plat_mxc.patch

16 years agoThis patch adds MX3-specific files to the linux 2.6.22 kernel.
Quinn Jensen [Thu, 25 Oct 2007 03:18:30 +0000 (21:18 -0600)]
This patch adds MX3-specific files to the linux 2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-mach_mx3.patch

16 years agoThis patch adds MX27-specific files to the linux 2.6.22 kernel.
Quinn Jensen [Thu, 25 Oct 2007 03:18:22 +0000 (21:18 -0600)]
This patch adds MX27-specific files to the linux 2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-mach_mx27.patch

16 years agoThis patch adds include files that are specific to the MX platform and
Quinn Jensen [Thu, 25 Oct 2007 03:18:16 +0000 (21:18 -0600)]
This patch adds include files that are specific to the MX platform and
related drivers to the linux 2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-include_mxc.patch

16 years agoThis patch adds Dallas 1-wire to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:18:11 +0000 (21:18 -0600)]
This patch adds Dallas 1-wire to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_w1.patch

16 years agoThis patch adds frame-buffer support to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:18:05 +0000 (21:18 -0600)]
This patch adds frame-buffer support to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_video.patch

16 years agoThis patch adds USB Host, Gadget and OTG functionality to the linux
Quinn Jensen [Thu, 25 Oct 2007 03:18:00 +0000 (21:18 -0600)]
This patch adds USB Host, Gadget and OTG functionality to the linux
2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_usb.patch

16 years agoThis patch adds the SPI driver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:55 +0000 (21:17 -0600)]
This patch adds the SPI driver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_spi.patch

16 years agoThis patch adds external UART (8250) and internal UART (mxc_uart)
Quinn Jensen [Thu, 25 Oct 2007 03:17:49 +0000 (21:17 -0600)]
This patch adds external UART (8250) and internal UART (mxc_uart)
support to the linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_serial.patch

16 years agoThis patch adds the RTC driver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:44 +0000 (21:17 -0600)]
This patch adds the RTC driver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_rtc.patch

16 years agoThis patch adds PCMCIA functionality to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:38 +0000 (21:17 -0600)]
This patch adds PCMCIA functionality to the linux 2.6.22 kernel
for MX platforms.  (Note that there may be hardware issues with
PCMCIA on the various MX platforms.)

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_pcmcia.patch

16 years agoThis patch adds network drivers (CS8900, FEC, IrDA) to the linux 2.6.22
Quinn Jensen [Thu, 25 Oct 2007 03:17:33 +0000 (21:17 -0600)]
This patch adds network drivers (CS8900, FEC, IrDA) to the linux 2.6.22
kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_net.patch

16 years agoThis patch adds security drivers to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:27 +0000 (21:17 -0600)]
This patch adds security drivers to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_mxc_security.patch

16 years agoThis patch adds PMIC (Power Management IC) drivers to the linux 2.6.22
Quinn Jensen [Thu, 25 Oct 2007 03:17:22 +0000 (21:17 -0600)]
This patch adds PMIC (Power Management IC) drivers to the linux 2.6.22
kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_mxc_pmic.patch

16 years agoThis patch adds MX-specific drivers to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:16 +0000 (21:17 -0600)]
This patch adds MX-specific drivers to the linux 2.6.22 kernel
for MX platforms.  Drivers include: Digital Audio Mux (DAM),
MPEG4, Image Processing Unit (IPU), Power Management, SSI, and
VPU (Video Processing Unit) drivers.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_mxc.patch

16 years agoThis patch adds mtd NOR and NAND capability to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:08 +0000 (21:17 -0600)]
This patch adds mtd NOR and NAND capability to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_mtd.patch

16 years agoThis patch adds basic mmc driver capability to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:17:03 +0000 (21:17 -0600)]
This patch adds basic mmc driver capability to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_mmc.patch

16 years agoThis patch adds the multimedia capture and output functionality to the
Quinn Jensen [Thu, 25 Oct 2007 03:16:58 +0000 (21:16 -0600)]
This patch adds the multimedia capture and output functionality to the
linux 2.6.22 kernel for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_media.patch

16 years agoThis patch adds keypad and touchscreendriver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:16:53 +0000 (21:16 -0600)]
This patch adds keypad and touchscreendriver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_input.patch

16 years agoThis patch adds ATA/IDE driver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:16:47 +0000 (21:16 -0600)]
This patch adds ATA/IDE driver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_ide.patch

16 years agoThis patch adds the i2c driver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:16:42 +0000 (21:16 -0600)]
This patch adds the i2c driver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_i2c.patch

16 years agoThis patch adds the watchdog-timer driver to the linux 2.6.22 kernel
Quinn Jensen [Thu, 25 Oct 2007 03:16:36 +0000 (21:16 -0600)]
This patch adds the watchdog-timer driver to the linux 2.6.22 kernel
for MX platforms.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-drivers_char.patch

16 years agoAdd CodeTEST functionality to the linux 2.6.22 kernel for ARM platforms.
Quinn Jensen [Thu, 25 Oct 2007 03:16:31 +0000 (21:16 -0600)]
Add CodeTEST functionality to the linux 2.6.22 kernel for ARM platforms.
Ported to 2.6.22 by Ross Wille

http://www.bitshrine.org/gpp/linux-2.6.22-mx-codetest.patch

16 years agoPatch to add oprofile and event-monitoring support for ARM-11 to the
Quinn Jensen [Thu, 25 Oct 2007 03:16:26 +0000 (21:16 -0600)]
Patch to add oprofile and event-monitoring support for ARM-11 to the
linux 2.6.22 kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-arm_oprofile.patch

16 years agoPatch to add basic MX-platform support to arch/arm of the linux 2.6.22
Quinn Jensen [Thu, 25 Oct 2007 03:16:21 +0000 (21:16 -0600)]
Patch to add basic MX-platform support to arch/arm of the linux 2.6.22
kernel.

http://www.bitshrine.org/gpp/linux-2.6.22-mx-arch_arm.patch

16 years agoLinux 2.6.22.6 v2.6.22.6
Greg Kroah-Hartman [Fri, 31 Aug 2007 06:21:01 +0000 (23:21 -0700)]
Linux 2.6.22.6

16 years agousb: add PRODUCT, TYPE to usb-interface events
Kay Sievers [Wed, 22 Aug 2007 22:15:51 +0000 (15:15 -0700)]
usb: add PRODUCT, TYPE to usb-interface events

This fixes a regression for userspace programs that were relying on these events.

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Cc: Andreas Jellinghaus <aj@ciphirelabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix DoS in pwc USB video driver
Oliver Neukum [Wed, 22 Aug 2007 22:15:43 +0000 (15:15 -0700)]
USB: fix DoS in pwc USB video driver

the pwc driver has a disconnect method that waits for user space to
close the device. This opens up an opportunity for a DoS attack,
blocking the USB subsystem and making khubd's task busy wait in
kernel space. This patch shifts freeing resources to close if an opened
device is disconnected.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: allow retry on descriptor fetch errors
Alan Stern [Wed, 22 Aug 2007 22:15:42 +0000 (15:15 -0700)]
USB: allow retry on descriptor fetch errors

This patch (as964) was suggested by Steffen Koepf.  It makes
usb_get_descriptor() retry on all errors other than ETIMEDOUT, instead
of only on EPIPE.  This helps with some devices.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: disable MSI on RX790
Tejun Heo [Wed, 22 Aug 2007 22:12:55 +0000 (15:12 -0700)]
PCI: disable MSI on RX790

RX790 can't do MSI like its predecessors.  Disable MSI on RX790.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: disable MSI on RD580
Tejun Heo [Wed, 22 Aug 2007 22:12:54 +0000 (15:12 -0700)]
PCI: disable MSI on RD580

RD580 can't do MSI like its predecessors.  Disable MSI on RD580.

Signed-off-by: Tejun Heo <teheo@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: disable MSI on RS690
Tejun Heo [Wed, 22 Aug 2007 22:12:53 +0000 (15:12 -0700)]
PCI: disable MSI on RS690

RS690 can't do MSI like its predecessors.  Disable MSI on RS690.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Henry Su <henry.su@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPCI: lets kill the 'PCI hidden behind bridge' message
Bernhard Kaindl [Wed, 22 Aug 2007 22:12:51 +0000 (15:12 -0700)]
PCI: lets kill the 'PCI hidden behind bridge' message

Adrian Bunk wrote:
> Alois NeÅ¡por wrote
>> PCI: Bus #0b (-#0e) is hidden behind transparent bridge #0a (-#0b) (try 'pci=assign-busses')
>> Please report the result to linux-kernel to fix this permanently"
>>
>> dmesg:
>> "Yenta: Raising subordinate bus# of parent bus (#0a) from #0b to #0e"
>> without pci=assign-busses and nothing with pci=assign-busses.
>
> Bernhard?

Ok, lets kill the message. As Alois NeÅ¡por also saw, that's fixed up by Yenta,
so PCI does not have to warn about it. PCI could still warn about it if
is_cardbus is 0 in that instance of pci_scan_bridge(), but so far I have
not seen a report where this would have been the case so I think we can
spare the kernel of that check (removes ~300 lines of asm) unless debugging
is done.

History: The whole check was added in the days before we had the fixup
for this in Yenta and pci=assign-busses was the only way to get CardBus
cards detected on many (not all) of the machines which give this warning.

In theory, there could be cases when this warning would be triggered and
it's not cardbus, then the warning should still apply, but I think this
should only be the case when working on a completely broken PCI setup,
but one may have already enabled the debug code in drivers/pci and the
patched check would then trigger.

I do not sign this off yet because it's completely untested so far, but
everyone is free to test it (with the #ifdef DEBUG replaced by #if 1 and
pr_debug( changed to printk(.

We may also dump the whole check (remove everything within the #ifdef from
the source) if that's perferred.

On Alois NeÅ¡por's machine this would then (only when debugging) this message:

"PCI: Bus #0b (-#0e) is partially hidden behind transparent bridge #0a (-#0b)"

"partially" should be in the message on his machine because #0b of #0b-#0e
is reachable behind #0a-#0b, but not #0c-#0e.

But that differentiation is now moot anyway because the fixup in Yenta takes
care of it as far as I could see so far, which means that unless somebody
is debugging a totally broken PCI setup, this message is not needed anymore,
not even for debugging PCI.

Ok, here the patch with the following changes:

* Refined to say that the bus is only partially hidden when the parent
  bus numbers are not totally way off (outside of) the child bus range
* remove the reference to pci=assign-busses and the plea to report it

We could add a pure source code-only comment to keep a reference to
pci=assign-busses the in case when this is triggered by someone who
is debugging the cause of this message and looking the way to solve it.

From: Bernhard Kaindl <bk@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoPPP: Fix PPP buffer sizing.
Konstantin Sharlaimov [Wed, 22 Aug 2007 04:20:25 +0000 (21:20 -0700)]
PPP: Fix PPP buffer sizing.

This patch addresses the issue with "osize too small" errors in mppe
encryption.  The patch fixes the issue with wrong output buffer size
being passed to ppp decompression routine.

--------------------
As pointed out by Suresh Mahalingam, the issue addressed by
ppp-fix-osize-too-small-errors-when-decoding patch is not fully resolved yet.
The size of allocated output buffer is correct, however it size passed to
ppp->rcomp->decompress in ppp_generic.c if wrong. The patch fixes that.
--------------------

Signed-off-by: Konstantin Sharlaimov <konstantin.sharlaimov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoTCP: Fix TCP handling of SACK in bidirectional flows.
Ilpo Järvinen [Wed, 22 Aug 2007 04:18:07 +0000 (21:18 -0700)]
TCP: Fix TCP handling of SACK in bidirectional flows.

It's possible that new SACK blocks that should trigger new LOST
markings arrive with new data (which previously made is_dupack
false). In addition, I think this fixes a case where we get
a cumulative ACK with enough SACK blocks to trigger the fast
recovery (is_dupack would be false there too).

I'm not completely pleased with this solution because readability
of the code is somewhat questionable as 'is_dupack' in SACK case
is no longer about dupacks only but would mean something like
'lost_marker_work_todo' too... But because of Eifel stuff done
in CA_Recovery, the FLAG_DATA_SACKED check cannot be placed to
the if statement which seems attractive solution. Nevertheless,
I didn't like adding another variable just for that either... :-)

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoTCP: Fix TCP rate-halving on bidirectional flows.
Ilpo Järvinen [Wed, 22 Aug 2007 04:17:03 +0000 (21:17 -0700)]
TCP: Fix TCP rate-halving on bidirectional flows.

Actually, the ratehalving seems to work too well, as cwnd is
reduced on every second ACK even though the packets in flight
remains unchanged. Recoveries in a bidirectional flows suffer
quite badly because of this, both NewReno and SACK are affected.

After this patch, rate halving is performed for ACK only if
packets in flight was supposedly changed too.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoTCP: Do not autobind ports for TCP sockets
David Miller [Wed, 22 Aug 2007 04:14:45 +0000 (21:14 -0700)]
TCP: Do not autobind ports for TCP sockets

[TCP]: Invoke tcp_sendmsg() directly, do not use inet_sendmsg().

As discovered by Evegniy Polyakov, if we try to sendmsg after
a connection reset, we can do incredibly stupid things.

The core issue is that inet_sendmsg() tries to autobind the
socket, but we should never do that for TCP.  Instead we should
just go straight into TCP's sendmsg() code which will do all
of the necessary state and pending socket error checks.

TCP's sendpage already directly vectors to tcp_sendpage(), so this
merely brings sendmsg() in line with that.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoSPARC64: Fix sparc64 PCI config accesses on sun4u
David Miller [Wed, 22 Aug 2007 04:12:32 +0000 (21:12 -0700)]
SPARC64: Fix sparc64 PCI config accesses on sun4u

[SPARC64]: Fix sun4u PCI config space accesses on sun4u.

Don't provide fake PCI config space for sun4u.

Also, put back the funny host controller space handling that
at least Sabre needs.  You have to read PCI host controller
registers at their nature size otherwise you get zeros instead
of correct values.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoSPARC64: Fix sparc64 task stack traces.
David Miller [Wed, 22 Aug 2007 04:11:14 +0000 (21:11 -0700)]
SPARC64: Fix sparc64 task stack traces.

It didn't handle that case at all, and now dump_stack()
can be implemented directly as show_stack(current, NULL)

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoNET: Fix missing rcu unlock in __sock_create()
Herbert Xu [Wed, 22 Aug 2007 04:09:15 +0000 (21:09 -0700)]
NET: Fix missing rcu unlock in __sock_create()

[NET]: Fix unbalanced rcu_read_unlock in __sock_create

The recent RCU work created an unbalanced rcu_read_unlock
in __sock_create.  This patch fixes that.  Reported by
oleg 123.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoSNAP: Fix SNAP protocol header accesses.
Herbert Xu [Wed, 22 Aug 2007 04:07:30 +0000 (21:07 -0700)]
SNAP: Fix SNAP protocol header accesses.

The snap_rcv code reads 5 bytes so we should make sure that
we have 5 bytes in the head before proceeding.

Based on diagnosis and fix by Evgeniy Polyakov, reported by
Alan J. Wylie.

Patch also kills the skb->sk assignment before kfree_skb
since it's redundant.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoNetfilter: Missing Kbuild entry for netfilter
Chuck Ebbert [Wed, 22 Aug 2007 04:05:14 +0000 (21:05 -0700)]
Netfilter: Missing Kbuild entry for netfilter

Author: Chuck Ebbert <cebbert@redhat.com>

Add xt_statistic.h to the list of headers to install.

Apparently needed to build newer versions of iptables.

Signed-off-by: Chuck Ebbert <cebbert@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoFix soft-fp underflow handling.
David Miller [Wed, 22 Aug 2007 04:04:07 +0000 (21:04 -0700)]
Fix soft-fp underflow handling.

The underflow exception cases were wrong.

This is one weird area of ieee1754 handling in that the underflow
behavior changes based upon whether underflow is enabled in the trap
enable mask of the FPU control register.  As a specific case the Sparc
V9 manual gives us the following description:

--------------------
If UFM = 0:     Underflow occurs if a nonzero result is tiny and a
                loss of accuracy occurs.  Tininess may be detected
                before or after rounding.  Loss of accuracy may be
                either a denormalization loss or an inexact result.

If UFM = 1:     Underflow occurs if a nonzero result is tiny.
                Tininess may be detected before or after rounding.
--------------------

What this amounts to in the packing case is if we go subnormal,
we set underflow if any of the following are true:

1) rounding sets inexact
2) we ended up rounding back up to normal (this is the case where
   we set the exponent to 1 and set the fraction to zero), this
   should set inexact too
3) underflow is set in FPU control register trap-enable mask

The initially discovered example was "DBL_MIN / 16.0" which
incorrectly generated an underflow.  It should not, unless underflow
is set in the trap-enable mask of the FPU csr.

Another example, "0x0.0000000000001p-1022 / 16.0", should signal both
inexact and underflow.  The cpu implementations and ieee1754
literature is very clear about this.  This is case #2 above.

However, if underflow is set in the trap enable mask, only underflow
should be set and reported as a trap.  That is handled properly by the
prioritization logic in

arch/sparc{,64}/math-emu/math.c:record_exception().

Based upon a report and test case from Jakub Jelinek.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoIPv6: Invalid semicolon after if statement
Ilpo Jarvinen [Wed, 22 Aug 2007 04:02:27 +0000 (21:02 -0700)]
IPv6: Invalid semicolon after if statement

Author: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>

A similar fix to netfilter from Eric Dumazet inspired me to
look around a bit by using some grep/sed stuff as looking for
this kind of bugs seemed easy to automate. This is one of them
I found where it looks like this semicolon is not valid.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoIPV6: Fix kernel panic while send SCTP data with IP fragments
Wei Yongjun [Wed, 22 Aug 2007 04:01:01 +0000 (21:01 -0700)]
IPV6: Fix kernel panic while send SCTP data with IP fragments

If ICMP6 message with "Packet Too Big" is received after send SCTP DATA,
kernel panic will occur when SCTP DATA is send again.

This is because of a bad dest address when call to skb_copy_bits().

The messages sequence is like this:

Endpoint A                             Endpoint B
                               <-------  SCTP DATA (size=1432)
ICMP6 message ------->
(Packet Too Big pmtu=1280)
                               <-------  Resend SCTP DATA (size=1432)
------------kernel panic---------------

 printing eip:
c05be62a
*pde = 00000000
Oops: 0002 [#1]
SMP
Modules linked in: scomm l2cap bluetooth ipv6 dm_mirror dm_mod video output sbs battery lp floppy sg i2c_piix4 i2c_core pcnet32 mii button ac parport_pc parport ide_cd cdrom serio_raw mptspi mptscsih mptbase scsi_transport_spi sd_mod scsi_mod ext3 jbd ehci_hcd ohci_hcd uhci_hcd
CPU:    0
EIP:    0060:[<c05be62a>]    Not tainted VLI
EFLAGS: 00010282   (2.6.23-rc2 #1)
EIP is at skb_copy_bits+0x4f/0x1ef
eax: 000004d0   ebx: ce12a980   ecx: 00000134   edx: cfd5a880
esi: c8246858   edi: 00000000   ebp: c0759b14   esp: c0759adc
ds: 007b   es: 007b   fs: 00d8  gs: 0000  ss: 0068
Process swapper (pid: 0, ti=c0759000 task=c06d0340 task.ti=c0713000)
Stack: c0759b88 c0405867 ce12a980 c8bff838 c789c084 00000000 00000028 cfd5a880
       d09f1890 000005dc 0000007b ce12a980 cfd5a880 c8bff838 c0759b88 d09bc521
       000004d0 fffff96c 00000200 00000100 c0759b50 cfd5a880 00000246 c0759bd4
Call Trace:
 [<c0405e1d>] show_trace_log_lvl+0x1a/0x2f
 [<c0405ecd>] show_stack_log_lvl+0x9b/0xa3
 [<c040608d>] show_registers+0x1b8/0x289
 [<c0406271>] die+0x113/0x246
 [<c0625dbc>] do_page_fault+0x4ad/0x57e
 [<c0624642>] error_code+0x72/0x78
 [<d09bc521>] ip6_output+0x8e5/0xab2 [ipv6]
 [<d09bcec1>] ip6_xmit+0x2ea/0x3a3 [ipv6]
 [<d0a3f2ca>] sctp_v6_xmit+0x248/0x253 [sctp]
 [<d0a3c934>] sctp_packet_transmit+0x53f/0x5ae [sctp]
 [<d0a34bf8>] sctp_outq_flush+0x555/0x587 [sctp]
 [<d0a34d3c>] sctp_retransmit+0xf8/0x10f [sctp]
 [<d0a3d183>] sctp_icmp_frag_needed+0x57/0x5b [sctp]
 [<d0a3ece2>] sctp_v6_err+0xcd/0x148 [sctp]
 [<d09cf1ce>] icmpv6_notify+0xe6/0x167 [ipv6]
 [<d09d009a>] icmpv6_rcv+0x7d7/0x849 [ipv6]
 [<d09be240>] ip6_input+0x1dc/0x310 [ipv6]
 [<d09be965>] ipv6_rcv+0x294/0x2df [ipv6]
 [<c05c3789>] netif_receive_skb+0x2d2/0x335
 [<c05c5733>] process_backlog+0x7f/0xd0
 [<c05c58f6>] net_rx_action+0x96/0x17e
 [<c042e722>] __do_softirq+0x64/0xcd
 [<c0406f37>] do_softirq+0x5c/0xac
 =======================
Code: 00 00 29 ca 89 d0 2b 45 e0 89 55 ec 85 c0 7e 35 39 45 08 8b 55 e4 0f 4e 45 08 8b 75 e0 8b 7d dc 89 c1 c1 e9 02 03 b2 a0 00 00 00 <f3> a5 89 c1 83 e1 03 74 02 f3 a4 29 45 08 0f 84 7b 01 00 00 01
EIP: [<c05be62a>] skb_copy_bits+0x4f/0x1ef SS:ESP 0068:c0759adc
Kernel panic - not syncing: Fatal exception in interrupt

Arnaldo says:
====================
Thanks! I'm to blame for this one, problem was introduced in:

b0e380b1d8a8e0aca215df97702f99815f05c094

                /*
                 *      Copy a block of the IP datagram.
                 */
-               if (skb_copy_bits(skb, ptr, frag->h.raw, len))
+               if (skb_copy_bits(skb, ptr, skb_transport_header(skb),
len))
                        BUG();
                left -= len;
====================

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoDCCP: Fix DCCP GFP_KERNEL allocation in atomic context
Gerrit Renker [Wed, 22 Aug 2007 03:53:27 +0000 (20:53 -0700)]
DCCP: Fix DCCP GFP_KERNEL allocation in atomic context

This fixes the following bug reported in syslog:

[ 4039.051658] BUG: sleeping function called from invalid context at /usr/src/davem-2.6/mm/slab.c:3032
[ 4039.051668] in_atomic():1, irqs_disabled():0
[ 4039.051670] INFO: lockdep is turned off.
[ 4039.051674]  [<c0104c0f>] show_trace_log_lvl+0x1a/0x30
[ 4039.051687]  [<c0104d4d>] show_trace+0x12/0x14
[ 4039.051691]  [<c0104d65>] dump_stack+0x16/0x18
[ 4039.051695]  [<c011371e>] __might_sleep+0xaf/0xbe
[ 4039.051700]  [<c0157b66>] __kmalloc+0xb1/0xd0
[ 4039.051706]  [<f090416f>] ccid2_hc_tx_alloc_seq+0x35/0xc3 [dccp_ccid2]
[ 4039.051717]  [<f09048d6>] ccid2_hc_tx_packet_sent+0x27f/0x2d9 [dccp_ccid2]
[ 4039.051723]  [<f085486b>] dccp_write_xmit+0x1eb/0x338 [dccp]
[ 4039.051741]  [<f085603d>] dccp_sendmsg+0x113/0x18f [dccp]
[ 4039.051750]  [<c03907fc>] inet_sendmsg+0x2e/0x4c
[ 4039.051758]  [<c033a47d>] sock_aio_write+0xd5/0x107
[ 4039.051766]  [<c015abc1>] do_sync_write+0xcd/0x11c
[ 4039.051772]  [<c015b296>] vfs_write+0x118/0x11f
[ 4039.051840]  [<c015b932>] sys_write+0x3d/0x64
[ 4039.051845]  [<c0103e7c>] syscall_call+0x7/0xb
[ 4039.051848]  =======================

The problem was that GFP_KERNEL was used; fixed by using gfp_any().

Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agosignalfd: make it group-wide, fix posix-timers scheduling
Oleg Nesterov [Wed, 22 Aug 2007 21:01:48 +0000 (14:01 -0700)]
signalfd: make it group-wide, fix posix-timers scheduling

With this patch any thread can dequeue its own private signals via signalfd,
even if it was created by another sub-thread.

To do so, we pass "current" to dequeue_signal() if the caller is from the same
thread group. This also fixes the scheduling of posix timers broken by the
previous patch.

If the caller doesn't belong to this thread group, we can't handle __SI_TIMER
case properly anyway. Perhaps we should forbid the cross-process signalfd usage
and convert ctx->tsk to ctx->sighand.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Cc: Roland McGrath <roland@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agosignalfd: fix interaction with posix-timers
Oleg Nesterov [Wed, 22 Aug 2007 21:01:42 +0000 (14:01 -0700)]
signalfd: fix interaction with posix-timers

dequeue_signal:

if (__SI_TIMER) {
spin_unlock(&tsk->sighand->siglock);
do_schedule_next_timer(info);
spin_lock(&tsk->sighand->siglock);
}

Unless tsk == curent, this is absolutely unsafe: nothing prevents tsk from
exiting. If signalfd was passed to another process, do_schedule_next_timer()
is just wrong.

Add yet another "tsk == current" check into dequeue_signal().

This patch fixes an oopsable bug, but breaks the scheduling of posix timers
if the shared __SI_TIMER signal was fetched via signalfd attached to another
sub-thread. Mostly fixed by the next patch.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Cc: Roland McGrath <roland@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoi386: fix lazy mode vmalloc synchronization for paravirt
Zachary Amsden [Wed, 22 Aug 2007 21:02:02 +0000 (14:02 -0700)]
i386: fix lazy mode vmalloc synchronization for paravirt

Found this looping Ubuntu installs with VMI.

If unlucky enough to hit a vmalloc sync fault during a lazy mode
operation (from an IRQ handler for a module which was not yet populated
in current page directory, or from inside copy_one_pte, which touches
swap_map, and hit in an unused 4M region), the required PDE update would
never get flushed, causing an infinite page fault loop.

This bug affects any paravirt-ops backend which uses lazy updates, I
believe that makes it a bug in Xen, VMI and lguest.  It only happens on
LOWMEM kernels.

Touching vmalloc memory in the middle of a lazy mode update can generate a
kernel PDE update, which must be flushed immediately.  The fix is to leave
lazy mode when doing a vmalloc sync.

Signed-off-by: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agouml: fix previous request size limit fix
Jeff Dike [Wed, 22 Aug 2007 21:01:53 +0000 (14:01 -0700)]
uml: fix previous request size limit fix

The previous patch which limited the number of sectors in a single request
to a COWed device was correct in concept, but the limit was implemented in
the wrong place.

By putting it in ubd_add, it covered the cases where the COWing was
specified on the command line.  However, when the command line only has the
COW file specified, the fact that it's a COW file isn't known until it's
opened, so the limit is missed in these cases.

This patch moves the sector limit from ubd_add to ubd_open_dev.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agosky2: don't clear phy power bits
Stephen Hemminger [Tue, 21 Aug 2007 18:10:22 +0000 (11:10 -0700)]
sky2: don't clear phy power bits

There are special PHY settings available on Yukon EC-U chip that
should not get cleared. This should solve mysterious errors on some
motherboards (like Gigabyte DS-3).

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>