10 years agokexec jump: remove duplication of kexec_restart_prepare()
Huang Ying [Fri, 15 Aug 2008 07:40:24 +0000]
kexec jump: remove duplication of kexec_restart_prepare()

Call kernel_restart_prepare() in kernel_kexec() instead of duplicating the
code.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Acked-by: Pavel Machek <pavel@suse.cz>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agokexec jump: check code size in control page
Huang Ying [Fri, 15 Aug 2008 07:40:23 +0000]
kexec jump: check code size in control page

Kexec/Kexec-jump require code size in control page is less than
PAGE_SIZE/2.  This patch add link-time checking for this.

ASSERT() of ld link script is used as the link-time checking mechanism.

[akpm@linux-foundation.org: build fix]
Signed-off-by: Huang Ying <ying.huang@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agokexec jump: rename KEXEC_CONTROL_CODE_SIZE to KEXEC_CONTROL_PAGE_SIZE
Huang Ying [Fri, 15 Aug 2008 07:40:22 +0000]
kexec jump: rename KEXEC_CONTROL_CODE_SIZE to KEXEC_CONTROL_PAGE_SIZE

Rename KEXEC_CONTROL_CODE_SIZE to KEXEC_CONTROL_PAGE_SIZE, because control
page is used for not only code on some platform.  For example in kexec
jump, it is used for data and stack too.

[akpm@linux-foundation.org: unbreak powerpc and arm, finish conversion]
Signed-off-by: Huang Ying <ying.huang@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agokexec jump: clean up #ifdef and comments
Huang Ying [Fri, 15 Aug 2008 07:40:21 +0000]
kexec jump: clean up #ifdef and comments

Move if (kexec_image->preserve_context) { ...  } into #ifdef
CONFIG_KEXEC_JUMP to make code looks cleaner.

Fix no longer correct comments of kernel_kexec().

Signed-off-by: Huang Ying <ying.huang@intel.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agokexec: fix compilation warning on xchg(&kexec_lock, 0) in kernel_kexec()
Huang Ying [Fri, 15 Aug 2008 07:40:20 +0000]
kexec: fix compilation warning on xchg(&kexec_lock, 0) in kernel_kexec()

kernel/kexec.c: In function 'kernel_kexec':
kernel/kexec.c:1506: warning: value computed is not used

Signed-off-by: Huang Ying <ying.huang@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agosuspend: fix section mismatch warning - register_nosave_region
Marcin Slusarz [Fri, 15 Aug 2008 07:40:19 +0000]
suspend: fix section mismatch warning - register_nosave_region

WARNING: vmlinux.o(.text+0xe684): Section mismatch in reference from the function register_nosave_region() to the function .init.text:__register_nosave_region()
  The function register_nosave_region() references
  the function __init __register_nosave_region().
  This is often because register_nosave_region lacks a __init
  annotation or the annotation of __register_nosave_region is wrong.

register_nosave_region calls __init function and is called only from
__init functions

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agoDocumentation/vm/page_migration: update reference to numa_maps + fix download URI
Michael Kerrisk [Fri, 15 Aug 2008 07:40:18 +0000]
Documentation/vm/page_migration: update reference to numa_maps + fix download URI

With man-pages-3.07, the numa_maps documentation home is now proc(5), so
the reference in Documentation/vm/page_migration needs updating.
(Cliff/Lee are removing numa_maps.5 from the numactl package.) Also, the
download location for the numactl package changed a while back.  This
patch fixes both things, as well as a typo (provided-->provides).

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Cliff Wickman <cpw@sgi.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Cc: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agobootmem allocator: alloc_bootmem_core(): page-align the end offset
Mikulas Patocka [Fri, 15 Aug 2008 07:40:17 +0000]
bootmem allocator: alloc_bootmem_core(): page-align the end offset

This is the minimal sequence that jams the allocator:

void *p, *q, *r;
p = alloc_bootmem(PAGE_SIZE);
q = alloc_bootmem(64);
free_bootmem(p, PAGE_SIZE);
p = alloc_bootmem(PAGE_SIZE);
r = alloc_bootmem(64);

after this sequence (assuming that the allocator was empty or page-aligned
before), pointer "q" will be equal to pointer "r".

What's hapenning inside the allocator:
p = alloc_bootmem(PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE, bitmap contains bits 10000...
q = alloc_bootmem(64);
in allocator: last_end_off == PAGE_SIZE + 64, bitmap contains 11000...
free_bootmem(p, PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE + 64, bitmap contains 01000...
p = alloc_bootmem(PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE, bitmap contains 11000...
r = alloc_bootmem(64);

and now:

it finds bit "2", as a place where to allocate (sidx)

it hits the condition

if (bdata->last_end_off && PFN_DOWN(bdata->last_end_off) + 1 == sidx))
start_off = ALIGN(bdata->last_end_off, align);

-you can see that the condition is true, so it assigns start_off =
ALIGN(bdata->last_end_off, align); (that is PAGE_SIZE) and allocates
over already allocated block.

With the patch it tries to continue at the end of previous allocation only
if the previous allocation ended in the middle of the page.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Johannes Weiner <hannes@saeurebad.de>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Thu, 14 Aug 2008 03:50:10 +0000]
Merge git://git./linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (47 commits)
  usb: musb: pass configuration specifics via pdata
  usb: musb: fix hanging when rmmod gadget driver
  USB: Add MUSB and TUSB support
  USB: serial: remove CONFIG_USB_DEBUG from sierra and option drivers
  USB: Add vendor/product id of ZTE MF628 to option
  USB: quirk PLL power down mode
  USB: omap_udc: fix compilation with debug enabled
  usb: cdc-acm: drain writes on close
  usb: cdc-acm: stop dropping tx buffers
  usb: cdc-acm: bugfix release()
  usb gadget: issue notifications from ACM function
  usb gadget: remove needless struct members
  USB: sh: r8a66597-hcd: fix disconnect regression
  USB: isp1301: fix compilation
  USB: fix compiler warning fix
  usb-storage: unusual_devs entry for Nokia 5300
  USB: cdc-acm.c: Fix compile warnings
  USB: BandRich BandLuxe C150/C250 HSPA Data Card Driver
  USB: ftdi_sio: add support for PHI Fisco data cable (FT232BM based, VID/PID 0403:e40b)
  usb: isp1760: don't be noisy about short packets.
  ...

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Thu, 14 Aug 2008 03:49:37 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  CRED: Introduce credential access wrappers

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Thu, 14 Aug 2008 03:48:46 +0000]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (56 commits)
  netns: Fix crash by making igmp per namespace
  bnx2x: Version update
  bnx2x: Checkpatch compliance
  bnx2x: Spelling mistakes
  bnx2x: Minor code improvements
  bnx2x: Driver info
  bnx2x: 1G LED does not turn off
  bnx2x: 8073 PHY changes
  bnx2x: Change GPIO for any port
  bnx2x: Pause settings
  bnx2x: Link order with external PHY
  bnx2x: No LRO without Rx checksum
  bnx2x: Wrong structure size
  bnx2x: WoL capability
  bnx2x: Clearing MAC addresses filters
  bnx2x: Delay in while loops
  bnx2x: PBA Table Page Alignment Workaround
  bnx2x: Self-test false positive
  bnx2x: Memory allocation
  bnx2x: HW attention lock
  ...

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Thu, 14 Aug 2008 03:48:25 +0000]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Handle stack trace attempts before irqstacks are setup.
  sparc64: Implement IRQ stacks.
  sparc: remove include of linux/of_device.h from asm/of_device.h
  sparc64: Fix recursion in stack overflow detection handling.
  sparc/drivers: use linux/of_device.h instead of asm/of_device.h
  sparc64: Don't MAGIC_SYSRQ ifdef smp_fetch_global_regs and support code.

10 years agousb: musb: pass configuration specifics via pdata
Felipe Balbi [Fri, 8 Aug 2008 09:40:54 +0000]
usb: musb: pass configuration specifics via pdata

Use platform_data to pass musb configuration-specific
details to musb driver.

This patch will prevent that other platforms selecting
HAVE_CLK and enabling musb won't break tree building.

The other parts of it will come when linux-omap merge
up more omap2/3 board-files.

Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: musb: fix hanging when rmmod gadget driver
Felipe Balbi [Mon, 4 Aug 2008 10:53:52 +0000]
usb: musb: fix hanging when rmmod gadget driver

If we try to modprobe a second gadget driver before
rmmoding the first one, the reference for the first
gadget driver would get NULLed avoiding usb to change
gadget drivers later.

Cc: David Brownell <david-b@pacbell.net>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: Add MUSB and TUSB support
Felipe Balbi [Thu, 24 Jul 2008 09:27:36 +0000]
USB: Add MUSB and TUSB support

This patch adds support for MUSB and TUSB controllers
integrated into omap2430 and davinci. It also adds support
for external tusb6010 controller.

Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: serial: remove CONFIG_USB_DEBUG from sierra and option drivers
Greg Kroah-Hartman [Tue, 12 Aug 2008 05:04:50 +0000]
USB: serial: remove CONFIG_USB_DEBUG from sierra and option drivers

These drivers should not be relying on CONFIG_USB_DEBUG.  By doing this,
it prevents users of kernels that do not enable this option from
enabling debugging in these drivers, unlike all other usb-serial
drivers.

Cc: Matthias Urlichs <smurf@smurf.noris.de>
Cc: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: Add vendor/product id of ZTE MF628 to option
Oliver Martin [Sat, 9 Aug 2008 02:49:26 +0000]
USB: Add vendor/product id of ZTE MF628 to option

This adds the vendor and product id (19d2:0015) of the ZTE MF628 HSDPA
modem to the option driver. It still needs a mode switch command issued
beforehand, this is currently handled by a userspace tool.

Signed-off-by: Oliver Martin <oliver.martin@student.tuwien.ac.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: quirk PLL power down mode
Libin Yang [Fri, 8 Aug 2008 07:03:31 +0000]
USB: quirk PLL power down mode

On some AMD 700 series southbridges, ISO OUT transfers (such as audio
playback through speakers) on the USB OHCI controller may be corrupted
when an A-Link express power saving feature is active.

PLL power down mode in conjunction with link power management feature
L1 being enabled is the bad combination ... this patch prevents them
from being enabled when ISO transfers are pending.

Signed-off-by: Crane Cai <crane.cai@amd.com>
Signed-off-by: Libin Yang <libin.yang@amd.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: omap_udc: fix compilation with debug enabled
Dmitry Baryshkov [Thu, 7 Aug 2008 12:29:25 +0000]
USB: omap_udc: fix compilation with debug enabled

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: cdc-acm: drain writes on close
David Brownell [Thu, 7 Aug 2008 01:46:10 +0000]
usb: cdc-acm: drain writes on close

Add a mechanism to let the write queue drain naturally before
closing the TTY, rather than always losing that data.  There
is a timeout, so it can't wait too long.

Provide missing locking inside acm_wb_is_avail(); it matters
more now.  Note, this presumes an earlier patch was applied,
removing a call to this routine where the lock was held.

Slightly improved diagnostics on write URB completion, so we
can tell when a write URB gets killed and, if so, how much
data it wrote first ... and so that I/O path is normally
silent (and can't much change timings).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: cdc-acm: stop dropping tx buffers
David Brownell [Thu, 7 Aug 2008 01:44:12 +0000]
usb: cdc-acm: stop dropping tx buffers

The "increase cdc-acm write throughput" patch left in place two
now-obsolete mechanisms, either of which can make the cdc-acm
driver drop TX data (nasty!).  This patch removes them:

  - The write_ready flag ... if an URB and buffer were found,
    they can (and should!) always be used.

  - TX path acm_wb_is_used() ... used when the buffer was just
    allocated, so that check is pointless.

Also fix a won't-yet-matter leak of a write buffer on a disconnect path.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: David Engraf <david.engraf@netcom.eu>
Acked-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: cdc-acm: bugfix release()
David Brownell [Thu, 7 Aug 2008 01:41:12 +0000]
usb: cdc-acm: bugfix release()

Bugfixes to the usb_driver_release_interface() usage;

  (a) make sure releasing *either* interface first will release
      the other, instead of insisting it be the control interface;

  (b) remove the recently-added self-deadlock.

(The "fix disconnect bug in cdc-acm" patch was incomplete and incorrect.)

Plus a small "sparse" fix:  rename a local variable so it doesn't
shadow a function parameter.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: issue notifications from ACM function
David Brownell [Thu, 7 Aug 2008 01:49:57 +0000]
usb gadget: issue notifications from ACM function

Update the CDC-ACM gadget code to support the peripheral-to-host
notifications when the tty is opened or closed, or issues a BREAK.
The serial framework code calls new generic hooks; right now only
CDC-ACM uses those hooks.  This resolves several REVISIT comments
in the code.  (Based on a patch from Felipe Balbi.)

Note that this doesn't expose USB_CDC_CAP_BRK to the host, since
this code still rejects USB_CDC_REQ_SEND_BREAK control requests
for host-to-peripheral BREAK signaling (received via /dev/ttyGS*).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: remove needless struct members
David Brownell [Thu, 7 Aug 2008 01:48:52 +0000]
usb gadget: remove needless struct members

This removes some unused members from the various USB functions.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: sh: r8a66597-hcd: fix disconnect regression
Yoshihiro Shimoda [Fri, 11 Jul 2008 09:53:45 +0000]
USB: sh: r8a66597-hcd: fix disconnect regression

fix the regression in commit 29fab0cd897519be9009ba8c898410ab83b378e9
that this driver executed reconnection processing when disconnected
some devices.

Signed-off-by: Yoshihiro Shimoda <shimoda.yoshihiro@renesas.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: isp1301: fix compilation
Dmitry Baryshkov [Thu, 7 Aug 2008 12:29:24 +0000]
USB: isp1301: fix compilation

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: fix compiler warning fix
Alan Stern [Thu, 7 Aug 2008 17:04:51 +0000]
USB: fix compiler warning fix

This patch (as1123b) fixes a compiler warning: do_unbind_rebind() is
defined but not used if CONFIG_PM=n.

Problem originally found and initial patch submitted by Alexander
Beregalov <a.beregalov@gmail.com>.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb-storage: unusual_devs entry for Nokia 5300
Alan Stern [Thu, 7 Aug 2008 17:02:40 +0000]
usb-storage: unusual_devs entry for Nokia 5300

This patch (as1120) adds an unusual_devs entry for the Nokia 5300.
Maybe once Nokia releases the Symbian code we'll be able to fix all
the problems it has with the USB mass-storage protocol.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Cedric Godin <cedric@belbone.be>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: cdc-acm.c: Fix compile warnings
Takashi Iwai [Thu, 31 Jul 2008 17:06:13 +0000]
USB: cdc-acm.c: Fix compile warnings

The irq flags should be unsigned long.

CC [M] drivers/usb/class/cdc-acm.o
drivers/usb/class/cdc-acm.c: In function 'acm_waker':
drivers/usb/class/cdc-acm.c:527: warning: comparison of distinct pointer types lacks a cast
drivers/usb/class/cdc-acm.c:529: warning: comparison of distinct pointer types lacks a cast

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: BandRich BandLuxe C150/C250 HSPA Data Card Driver
Leon Leong [Fri, 1 Aug 2008 03:40:12 +0000]
USB: BandRich BandLuxe C150/C250 HSPA Data Card Driver

This patch adds the Product ID for the BandLuxe C150/C250 3.5G data
card series from BandRich Inc.
After detection, the data card works fine.

It was patched against kernel 2.6.27-rc1 with -mm patch

Signed-off-by: Leon Leong <upleong@bandrich.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: ftdi_sio: add support for PHI Fisco data cable (FT232BM based, VID/PID 0403...
Lex Ross [Wed, 6 Aug 2008 12:25:08 +0000]
USB: ftdi_sio: add support for PHI Fisco data cable (FT232BM based, VID/PID 0403:e40b)

Support for PHI Fisco USB to Serial data cable (FTDI FT232BM based).
PHI Fisco cable is supplied for connecting Philips Xenium 9@9++ mobile phones.
PIDs were missing.

Tested successfully with PHI Fisco Data Cable (VID/PID 0403:e40b)

Signed-off-by: Lex V. Ross <lross@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: isp1760: don't be noisy about short packets.
Sebastian Siewior [Thu, 17 Jul 2008 18:09:29 +0000]
usb: isp1760: don't be noisy about short packets.

According to Alan Stern, short packets are quite normal under
certain circumstances. This printk was triggered by usb to
serial converters on every packet and some usb sticks triggered
a few of those while plugging the stick.
This printks are now hidden unless USB debug mode is activated.

Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: ISP1760: improve pre-fetch timing
Enrico Scholz [Thu, 17 Jul 2008 18:09:30 +0000]
usb: ISP1760: improve pre-fetch timing

ISP1760 requires a delay of 90ns between programming the address and
reading the data.  Current driver solves this by a mdelay(1) which is
very heavy weighted and slow.  This patch applies the workaround from
the ISP1760 FAQ by using two different banks for PTD and payload data
and using a common wait for them.  This wait is done by an additional
ISP1760 access (whose timing constraints guarantee the 90ns delay).
This improves speed when reading from an USB stick from:

  $ time dd if=/dev/sda of=/dev/zero bs=65536 count=1638
  real    1m 15.43s
  user    0m 0.44s
  sys     0m 39.46s

to

  $ time dd if=/dev/sda of=/dev/zero bs=65536 count=1638
  real    0m 18.53s
  user    0m 0.16s
  sys     0m 12.97s

[bigeasy@linutronix.de: fixed comment formating, moved define into
   header file, obey 80 char rule]

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: return error code instead of 0 in the enqueue function.
Sebastian Siewior [Thu, 17 Jul 2008 18:09:28 +0000]
usb: return error code instead of 0 in the enqueue function.

if the enqueue function returns -ESHUTDOWN or -ENOMEM then
we return 0 instead of an error. This leads to a timeout and
then to a dequeue request of an not enqueued urb.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: serial gadget: rx path data loss fixes
David Brownell [Mon, 7 Jul 2008 19:16:08 +0000]
USB: serial gadget: rx path data loss fixes

Update RX path handling in new serial gadget code to cope better with
RX blockage:  queue every RX packet until its contents can safely be
passed up to the ldisc.  Most of the RX path work is now done in the
RX tasklet, instead of just the final "push to ldisc" step.  This
addresses some cases of data loss:

  - A longstanding serial gadget bug: when tty_insert_flip_string()
    didn't copy the entire buffer, the rest of the characters were
    dropped!  Now that packet stays queued until the rest of its data
    is pushed to the ldisc.

  - Another longstanding issue:  in the unlikely case that an RX
    transfer returns data and also reports a fault, that data is
    no longer discarded.

  - In the recently added RX throttling logic:  it needs to stop
    pushing data into the TTY layer, instead of just not submitting
    new USB read requests.  When the TTY is throttled long enough,
    backpressure will eventually make the OUT endpoint NAK.

Also: an #ifdef is removed (no longer necessary); and start switching
to a better convention for debug messages (prefix them with tty name).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: Hook start_hnp into ohci struct
Dmitry Baryshkov [Sun, 6 Jul 2008 10:26:30 +0000]
USB: Hook start_hnp into ohci struct

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: ohci: make distrust_firmware a quirk
Dmitry Baryshkov [Sun, 6 Jul 2008 19:35:01 +0000]
USB: ohci: make distrust_firmware a quirk

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: update to Documentation
Oliver Neukum [Fri, 4 Jul 2008 08:10:53 +0000]
USB: update to Documentation

this mentions a new deadlock due to advanced power management.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: cxacru: Fix printk format flag in error message
Simon Arlott [Sat, 12 Jul 2008 21:19:48 +0000]
USB: cxacru: Fix printk format flag in error message

"#%x" should have been "%#x"

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: usb-storage Motorola Phone Razr v3xx US_FL_FIX_CAPACITY patch
Jost Diederichs [Sun, 13 Jul 2008 05:37:05 +0000]
USB: usb-storage Motorola Phone Razr v3xx US_FL_FIX_CAPACITY patch

add razr v3xx US_FL_FIX_CAPACITY flag to unusual_devs.h in usb-storage

This is another Motorola phone that incorrectly reports the sector count
(off by one).

Problem Description: io errors when mounting phone's sd-card via the
phones usb port

Steps to reproduce: mount Motorola Razr v3xx phones sd-card on Linux Desktop
via usb cable. Phones USB port must be in memory card mode.

DEBUG output:
Jul 9 19:32:41 micky kernel: Buffer I/O error on device sdd, logical block 3970048
Jul 9 19:32:41 micky kernel: sd 11:0:0:0: [sdd] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
Jul 9 19:32:41 micky kernel: sd 11:0:0:0: [sdd] Sense Key : Medium Error [current]
Jul 9 19:32:41 micky kernel: sd 11:0:0:0: [sdd] Add. Sense: No additional sense information
Jul 9 19:32:41 micky kernel: end_request: I/O error, dev sdd, sector 3970048

From: Jost Diederichs <jost@qdusa.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB Storage Sierra: TRU-Install feature update
Kevin Lloyd [Thu, 10 Jul 2008 21:14:57 +0000]
USB Storage Sierra: TRU-Install feature update

This patch upgrades the support for the Sierra Wireless TRU-Install
feature (i.e. zeroCD) to allow for future support of Linux enabled
TRU-Install devices.

By default all devices that do not have a Linux enabled TRU-Install
device (i.e. the device does not have a Linux package on the virtual CD
partition) will be switched into "modem mode." Devices that do contain a
Linux package in the TRU-Install virtual CD will be allowed to enumerate
as a CD-Rom so that either (a) a user can install the packaged software
or (b) a user-space application (e.g. udev) can switch it to modem mode.

This patch does allow for manual override by adding a usb-storage module
parameter 'swi_tru_install' which can force the modem into either mode
regardless of what packages it contains.

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB Serial Sierra: TRU-Install feature update
Kevin Lloyd [Thu, 10 Jul 2008 21:14:54 +0000]
USB Serial Sierra: TRU-Install feature update

Moves responsbility of TRU-Install (i.e. ZeroCD) to the usb-storage
driver. See patch 04/04 of this set.

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB Serial Sierra: Dynamic interface detection
Kevin Lloyd [Thu, 10 Jul 2008 21:14:51 +0000]
USB Serial Sierra: Dynamic interface detection

This patch changes the method by which the number of ports per interface is
assigned so that it is more dynamic and calculated on the fly (as opposed to
hard coding it). This will allow for faster and easier addition of products.

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB Serial Sierra: clean-up
Kevin Lloyd [Thu, 10 Jul 2008 21:14:47 +0000]
USB Serial Sierra: clean-up

Very minor changes to clean up sierra code. Adds a prefix to debug messages so
that Sierra messages are easily recognized. Removes extraneous code.

This targets kernel 2.6.26-rc9

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: pl2023: Remove USB id (4348:5523) handled by ch341
Tollef Fog Heen [Tue, 22 Jul 2008 07:28:05 +0000]
USB: pl2023: Remove USB id (4348:5523) handled by ch341

USB ID 4348:5523 is handled by the ch341 driver.  Remove it from the
pl2023 driver.

Reverts 002e8f2c80c6be76bb312940bc278fc10b2b2487.

Signed-off-by: Tollef Fog Heen <tfheen@err.no>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb-storage: unusual_devs entries for iRiver T10 and Datafab CF+SM reader
Alan Stern [Mon, 21 Jul 2008 14:08:28 +0000]
usb-storage: unusual_devs entries for iRiver T10 and Datafab CF+SM reader

This patch (as1115) adds unusual_devs entries with the IGNORE_RESIDE
flag for the iRiver T10 and the Simple Tech/Datafab CF+SM card
reader.  Apparently these devices provide reasonable residue values
for READ and WRITE operations, but not for others like INQUIRY or READ
CAPACITY.

This fixes the iRiver T10 problem reported in Bugzilla #11125.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: Move usb/mon/ up to misc options in Kconfig
Simon Arlott [Mon, 21 Jul 2008 19:44:50 +0000]
USB: Move usb/mon/ up to misc options in Kconfig

This makes "USB Monitor" appear under "Miscellaneous USB options"
section instead of in the middle of device specific drivers in the
"USB Imaging devices" section.

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb-storage: automatically recognize bad residues
Alan Stern [Tue, 29 Jul 2008 15:58:06 +0000]
usb-storage: automatically recognize bad residues

This patch (as1119) will help to reduce the clutter of usb-storage's
unusual_devs file by automatically detecting some devices that need
the IGNORE_RESIDUE flag.  The idea is that devices should never return
a non-zero residue for an INQUIRY or a READ CAPACITY command unless
they failed to transfer all the requested data.  So if one of these
commands transfers a standard amount of data but there is a positive
residue, we know that the residue is bogus and we can set the flag.

This fixes the problems reported in Bugzilla #11125.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: fix interface unregistration logic
Alan Stern [Wed, 30 Jul 2008 15:31:50 +0000]
USB: fix interface unregistration logic

This patch (as1122) fixes a bug: When an interface is unregistered,
its children (sysfs files and endpoint devices) are unregistered after
it instead of before.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Kay Sievers <kay.sievers@vrfy.org>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb-serial: don't release unregistered minors
Alan Stern [Tue, 29 Jul 2008 16:01:04 +0000]
usb-serial: don't release unregistered minors

This patch (as1121) fixes a bug in the USB serial core.  When a device
is unregistered, the core will give back its minors -- even if the
device hasn't been assigned any!

The patch reserves the highest minor value (255) to mean that no minor
was assigned.  It also removes some dead code and does a small style
fixup.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: gadget: protect gadget_chips.h from been included twice
Felipe Balbi [Thu, 17 Jul 2008 14:26:49 +0000]
usb: gadget: protect gadget_chips.h from been included twice

Without it, we might have trouble when trying to write
some composite gadget drivers.

Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: usb-storage: quirk around v1.11 firmware on Nikon D4
Dave Jones [Mon, 14 Jul 2008 17:28:34 +0000]
USB: usb-storage: quirk around v1.11 firmware on Nikon D4

usb-storage: quirk around v1.11 firmware on Nikon D40

https://bugzilla.redhat.com/show_bug.cgi?id=454028
Just as in earlier firmware versions, we need to perform this
quirk for the latest version too.

Speculatively do the entry for the D80 too, as they seem to
have the same firmware problems historically.

Signed-off-by: Dave Jones <davej@redhat.com>
Cc: Johannes Berg <johannes@sipsolutions.net>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: OHCI: fix system hang caused by earlier patch
Alan Stern [Mon, 21 Jul 2008 13:56:26 +0000]
USB: OHCI: fix system hang caused by earlier patch

This patch (as1114) fixes a problem that was revealed by an earlier
patch (as1069b).  Some broken controllers seem never to turn off their
RHCS interrupt status bit, even when told to do so.  As a result they
generate an interrupt storm and hang the system.

The patch avoids enabling RHSC interrupt requests when the RHCS status
bit is already set.  This should have no adverse affects on normal
controllers, since they won't set the status bit until a root-hub
status change actually occurs, in which case we wouldn't enable RHSC
interrupt requests anyway -- we would wait until the status change had
been processed and cleared.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested by: Andrey Borzenkov <arvidjaar@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb/core/driver: fix warning
Alexander Beregalov [Tue, 29 Jul 2008 01:54:31 +0000]
usb/core/driver: fix warning

usb/core/driver: fix warning:
drivers/usb/core/driver.c:834: warning: 'do_unbind_rebind' defined but not used

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: add missing kerneldoc line for "needs_binding"
Alan Stern [Mon, 28 Jul 2008 14:39:28 +0000]
USB: add missing kerneldoc line for "needs_binding"

This patch (as1117) adds a kerneldoc line for the "needs_binding"
field in struct usb_interface.  It was accidentally omitted when the
field was added.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: ftdi_sio: add support for Luminance Stellaris Evaluation/Development Kits
Frederik Kriewitz [Wed, 30 Jul 2008 14:53:41 +0000]
USB: ftdi_sio: add support for Luminance Stellaris Evaluation/Development Kits

The Patch adds support for Luminance Stellaris Evaluation/Development
Kits (FTDI 2232C based).
The PIDs were missing.

Successfully tested with a Stellaris LM3S8962 Evaluation kit.

Signed-off-by: Frederik Kriewitz <frederik@kriewitz.eu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: ftdi_sio: Add USB Product Id for ELV HS485
André Schenk [Mon, 28 Jul 2008 13:48:46 +0000]
USB: ftdi_sio: Add USB Product Id for ELV HS485

USB product id registration for the ELV HS485 USB adapter (www.elv.de) to
their home automation bus system. Applies to 2.6.26.

Signed-off-by: Andre Schenk <andre@melior.s.bawue.de>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoUSB: fix USB boot crash, ecm_do_notify(), list_add corruption. prev->next should...
David Brownell [Sat, 26 Jul 2008 15:06:24 +0000]
USB: fix USB boot crash, ecm_do_notify(), list_add corruption. prev->next should be next (ffff88003b8f82f8)

This fixes a BUG() turned up by Ingo via randconfig testing, where
CONFIG_LIST_DEBUG turned up list corruption.  The corruption was
caused by the dummy_hcd (single-machine test harness for gadget and
HCD code) trashing the request queue when driven by the new CDC
composite gadget an I/O pattern that was previously uncommon.
Fix suggested by Alan Stern.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Tested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: auerswald: remove driver (obsolete)
Wolfgang Mües [Tue, 29 Jul 2008 09:54:43 +0000]
usb: auerswald: remove driver (obsolete)

This patch removes the auerswald USB driver from the linux kernel
2.6.26.

This driver was included into the kernel mainly to connect to the ISDN
framework. This was done in linux 2.4.x. For 2.6.x, due to the fragile
and moving ISDN support, this connection was never realized, and the
only use of this driver was for device configuration. In the age of DSL,
the demand of ISDN support is getting very low.

Meanwhile, with the advent of libusb, an userspace driver was done for
the device configuration which works fine for linux and mac. (Thanks to
the libusb developers!). The userspace driver is downloadable from the
auerswald web site.

So this driver is obsolete now and has to be removed. Many thanks to all
developers which helped me to bring this driver up and working.

Signed-off-by: Wolfgang Muees <wolfgang@iksw-muees.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosparc64: Handle stack trace attempts before irqstacks are setup.
David S. Miller [Thu, 14 Aug 2008 00:17:52 +0000]
sparc64: Handle stack trace attempts before irqstacks are setup.

Things like lockdep can try to do stack backtraces before
the irqstack blocks have been setup.  So don't try to match
their ranges so early on.

Also, remove unused variable in save_stack_trace().

Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agoCRED: Introduce credential access wrappers
David Howells [Wed, 13 Aug 2008 15:20:04 +0000]
CRED: Introduce credential access wrappers

The patches that are intended to introduce copy-on-write credentials for 2.6.28
require abstraction of access to some fields of the task structure,
particularly for the case of one task accessing another's credentials where RCU
will have to be observed.

Introduced here are trivial no-op versions of the desired accessors for current
and other tasks so that other subsystems can start to be converted over more
easily.

Wrappers are introduced into a new header (linux/cred.h) for UID/GID,
EUID/EGID, SUID/SGID, FSUID/FSGID, cap_effective and current's subscribed
user_struct.  These wrappers are macros because the ordering between header
files mitigates against making them inline functions.

linux/cred.h is #included from linux/sched.h.

Further, XFS is modified such that it no longer defines and uses parameterised
versions of current_fs[ug]id(), thus getting rid of the namespace collision
otherwise incurred.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>

10 years agonetns: Fix crash by making igmp per namespace
Daniel Lezcano [Wed, 13 Aug 2008 23:15:57 +0000]
netns: Fix crash by making igmp per namespace

This patch makes the multicast socket to be per namespace.

When a network namespace is created, other than the init_net and a
multicast packet is received, the kernel goes to a hang or a kernel panic.

How to reproduce ?

 * create a child network namespace
 * create a pair virtual device veth
    * ip link add type veth
 * move one side to the pair network device to the child namespace
    * ip link set netns <childpid> dev veth1
 * ping -I veth0 224.0.0.1

The bug appears because the function ip_mc_init_dev does not initialize
the different multicast fields as it exits because it is not the init_net.

BUG: soft lockup - CPU#0 stuck for 61s! [avahi-daemon:2695]
Modules linked in:
irq event stamp: 50350
hardirqs last  enabled at (50349): [<c03ee949>] _spin_unlock_irqrestore+0x34/0x39
hardirqs last disabled at (50350): [<c03ec639>] schedule+0x9f/0x5ff
softirqs last  enabled at (45712): [<c0374d4b>] ip_setsockopt+0x8e7/0x909
softirqs last disabled at (45710): [<c03ee682>] _spin_lock_bh+0x8/0x27

Pid: 2695, comm: avahi-daemon Not tainted (2.6.27-rc2-00029-g0872073 #3)
EIP: 0060:[<c03ee47c>] EFLAGS: 00000297 CPU: 0
EIP is at __read_lock_failed+0x8/0x10
EAX: c4f38810 EBX: c4f38810 ECX: 00000000 EDX: c04cc22e
ESI: fb0000e0 EDI: 00000011 EBP: 0f02000a ESP: c4e3faa0
 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
CR0: 8005003b CR2: 44618a40 CR3: 04e37000 CR4: 000006d0
DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
DR6: ffff0ff0 DR7: 00000400
 [<c02311f8>] ? _raw_read_lock+0x23/0x25
 [<c0390666>] ? ip_check_mc+0x1c/0x83
 [<c036d478>] ? ip_route_input+0x229/0xe92
 [<c022e2e4>] ? trace_hardirqs_on_thunk+0xc/0x10
 [<c0104c9c>] ? do_IRQ+0x69/0x7d
 [<c0102e64>] ? restore_nocheck_notrace+0x0/0xe
 [<c036fdba>] ? ip_rcv+0x227/0x505
 [<c0358764>] ? netif_receive_skb+0xfe/0x2b3
 [<c03588d2>] ? netif_receive_skb+0x26c/0x2b3
 [<c035af31>] ? process_backlog+0x73/0xbd
 [<c035a8cd>] ? net_rx_action+0xc1/0x1ae
 [<c01218a8>] ? __do_softirq+0x7b/0xef
 [<c0121953>] ? do_softirq+0x37/0x4d
 [<c035b50d>] ? dev_queue_xmit+0x3d4/0x40b
 [<c0122037>] ? local_bh_enable+0x96/0xab
 [<c035b50d>] ? dev_queue_xmit+0x3d4/0x40b
 [<c012181e>] ? _local_bh_enable+0x79/0x88
 [<c035fcb8>] ? neigh_resolve_output+0x20f/0x239
 [<c0373118>] ? ip_finish_output+0x1df/0x209
 [<c0373364>] ? ip_dev_loopback_xmit+0x62/0x66
 [<c0371db5>] ? ip_local_out+0x15/0x17
 [<c0372013>] ? ip_push_pending_frames+0x25c/0x2bb
 [<c03891b8>] ? udp_push_pending_frames+0x2bb/0x30e
 [<c038a189>] ? udp_sendmsg+0x413/0x51d
 [<c038a1a9>] ? udp_sendmsg+0x433/0x51d
 [<c038f927>] ? inet_sendmsg+0x35/0x3f
 [<c034f092>] ? sock_sendmsg+0xb8/0xd1
 [<c012d554>] ? autoremove_wake_function+0x0/0x2b
 [<c022e6de>] ? copy_from_user+0x32/0x5e
 [<c022e6de>] ? copy_from_user+0x32/0x5e
 [<c034f238>] ? sys_sendmsg+0x18d/0x1f0
 [<c0175e90>] ? pipe_write+0x3cb/0x3d7
 [<c0170347>] ? do_sync_write+0xbe/0x105
 [<c012d554>] ? autoremove_wake_function+0x0/0x2b
 [<c03503b2>] ? sys_socketcall+0x176/0x1b0
 [<c01085ea>] ? syscall_trace_enter+0x6c/0x7b
 [<c0102e1a>] ? syscall_call+0x7/0xb

Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Version update
Eilon Greenstein [Wed, 13 Aug 2008 22:59:45 +0000]
bnx2x: Version update

Version update

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Checkpatch compliance
Eilon Greenstein [Wed, 13 Aug 2008 22:59:25 +0000]
bnx2x: Checkpatch compliance

Checkpatch compliance
The latest version of checkpatch found the following style errors in the
code

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Spelling mistakes
Eilon Greenstein [Wed, 13 Aug 2008 22:59:08 +0000]
bnx2x: Spelling mistakes

Spelling mistakes
Spelling has to L's in it...

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Minor code improvements
Eilon Greenstein [Wed, 13 Aug 2008 22:58:49 +0000]
bnx2x: Minor code improvements

Minor code improvements
Small changes to make the code a little bit more efficient and mostly
more readable:
- Using unified macros for EMAC_RD/WR which looks like normal REG_RD/WR
- Removing the NIG_WR since it did nothing and was only confusing
- On bnx2x_panic_dump, print only the used parts of the rings
- define parameters only on the branch they are needed and not at the
  beginning of the function
- using NETIF_MSG_INTR and not private BNX2X_MSG_SP for debug prints

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Driver info
Eilon Greenstein [Wed, 13 Aug 2008 22:58:30 +0000]
bnx2x: Driver info

Driver info
The internal FW which is downloaded by the driver should not be
displayed - it is only causing confusion and it is redundant since it
can be concluded from the driver version. Display only FW which is
burned on the board nvram

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: 1G LED does not turn off
Eilon Greenstein [Wed, 13 Aug 2008 22:58:12 +0000]
bnx2x: 1G LED does not turn off

1G LED does not turn off
The 1G LED was not switched to off when the link was lost

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: 8073 PHY changes
Yaniv Rosner [Wed, 13 Aug 2008 22:57:28 +0000]
bnx2x: 8073 PHY changes

8073 PHY changes
The initial support we had for this PHY needs some serious changing. The
major change is that this PHY should be initialized only when the first
function is loaded and not for each function. The official SPI-ROM of
this PHY was released and it requires some changes in the initialization
code as well

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Change GPIO for any port
Eilon Greenstein [Wed, 13 Aug 2008 22:56:59 +0000]
bnx2x: Change GPIO for any port

Change GPIO for any port
The set GPIO function should receive the port index to allow changing
the GPIO of another port. This is needed for the common init phase (one
the first driver is loaded for the chip)

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Pause settings
Yaniv Rosner [Wed, 13 Aug 2008 22:56:17 +0000]
bnx2x: Pause settings

Pause settings
- 1G pause was not working due to missing write to the emac block
  (TX_MODE_FLOW_EN)
- The flow control should use the negotiated result (after autoneg) so
  we should save both the requested autoneg and the result
- The HW credits with flow control at 1G speed were not optimized and
  caused low throughput
- It is recommended to turn off flow control if the MTU is bigger than
  5000B due to internal buffers size

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Link order with external PHY
Yaniv Rosner [Wed, 13 Aug 2008 22:55:28 +0000]
bnx2x: Link order with external PHY

Link order with external PHY
When external PHY exists (second chip with the PHY to translate to
another physical medium) the link with the eternal PHY and the network
should be established before setting the link between the 5771x and the
PHY. This is the right order and it is important when using autoneg -
the link to the network should use the autoneg and the link between the
two chips should be forced to the network result.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: No LRO without Rx checksum
Vladislav Zolotarov [Wed, 13 Aug 2008 22:53:38 +0000]
bnx2x: No LRO without Rx checksum

No LRO without Rx checksum
Disabling LRO when Rx checksum is disabled

Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Wrong structure size
Yitchak Gertner [Wed, 13 Aug 2008 22:53:12 +0000]
bnx2x: Wrong structure size

Wrong structure size
The wrong structure was used in the sizeof to clear (luckily both
structures have the same size in this version...)

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: WoL capability
Eilon Greenstein [Wed, 13 Aug 2008 22:52:46 +0000]
bnx2x: WoL capability

WoL capability
All designs reported WoL capability regardless of HW limitations - check
if this device is actually capable of WoL

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Clearing MAC addresses filters
Yitchak Gertner [Wed, 13 Aug 2008 22:52:28 +0000]
bnx2x: Clearing MAC addresses filters

Clearing MAC addresses filters
When the driver unloads, it should clear the MAC addresses filters in
the HW - this prevents packets from entering the chip when the driver is
re-loaded before initializing the right filters

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Delay in while loops
Yitchak Gertner [Wed, 13 Aug 2008 22:52:08 +0000]
bnx2x: Delay in while loops

Delay in while loops
The delay in the loop should be after the change. This has very little
effect (can save one delay) but it is the right thing to do

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: PBA Table Page Alignment Workaround
Eilon Greenstein [Wed, 13 Aug 2008 22:51:48 +0000]
bnx2x: PBA Table Page Alignment Workaround

PBA Table Page Alignment Workaround
The PBA table starts on the middle of the page and that's causing very
low performance with virtualization. The solution is not to update via
the BAR directly but via chip access to the same memory

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Self-test false positive
Yitchak Gertner [Wed, 13 Aug 2008 22:51:28 +0000]
bnx2x: Self-test false positive

Self-test false positive
- The memory test should use a mask according to the chip type
- In the register test, check the port only once and not inside the for
  loop (not causing a failure - just ugly)

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Memory allocation
Eilon Greenstein [Wed, 13 Aug 2008 22:51:07 +0000]
bnx2x: Memory allocation

Memory allocation
- The CQE ring was allocated to the max size even for a chip that does
  not support it. Fixed to allocate according to the chip type to save
  memory
- The rx_page_ring was not freed on driver unload

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: HW attention lock
Eilon Greenstein [Wed, 13 Aug 2008 22:50:45 +0000]
bnx2x: HW attention lock

HW attention lock
Making sure that only one function will handle the HW attention. This
makes the device parameter aeu_mask redundant so it is removed

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: HW lock mechanism
Yitchak Gertner [Wed, 13 Aug 2008 22:50:23 +0000]
bnx2x: HW lock mechanism

HW lock mechanism
Enhancing the HW lock to work per function and not only per port - this
is needed for the next patch that protects races over HW attention
detection between the different functions. At this chance, changing the
functions names to be more inline with the current naming convention

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Load/Unload under traffic
Vladislav Zolotarov [Wed, 13 Aug 2008 22:50:00 +0000]
bnx2x: Load/Unload under traffic

Load/Unload under traffic
Few issues were found when loading and unloading under traffic:
- When receiving Tx interrupt call netif_wake_queue if the queue is
  stopped but the state is open
- Check that interrupts are enabled before doing anything else on the
  msix_fp_int function
- In nic_load, enable the interrupts only when needed and ready for it
- Function stop_leading returns status since it can fail
- Add 1ms delay when unloading the driver to validate that there are no
  open transactions that already started by the FW
- Splitting the "has work" function into Tx and Rx so the same function
  will be used on unload and interrupts
- Do not request for WoL if only resetting the device (save the time
  that it takes the FW to set the link after reset)
- Fixing the device reset after iSCSI boot and before driver load - all
  internal buffers must be cleared before the driver is loaded

Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: FW Internal Memory structure
Eilon Greenstein [Wed, 13 Aug 2008 22:49:35 +0000]
bnx2x: FW Internal Memory structure

FW Internal Memory structure
The FW uses data structures on the chip internal memory to aggregate the
connections when TPA is enabled. The driver was clearing the wrong offsets
and therefore one function could cause another function to loose packets.
Changing the initialization of the chip internal memory to clear only the
relevant memory for each function which is being loaded

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Statistics
Yitchak Gertner [Wed, 13 Aug 2008 22:49:05 +0000]
bnx2x: Statistics

Statistics
- Making sure that each drop is accounted for in the driver statistics
- Clearing the FW statistics when driver is loaded to prevent
  inconsistency with HW statistics
- Once error is detected (bnx2x_panic_dump), stop the statistics
  before other actions (currently it is stopped last and can corrupt
  the data) - Adding HW checksum error counter to the statistics
- Removing unused variable stats_ticks
- Using macros instead of magic numbers to indicate which statistics are
  shared per port and which are per function

Signed-off-by: Yitchak Gertner <gertner@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: Not dropping packets with L3/L4 checksum error
Eilon Greenstein [Wed, 13 Aug 2008 22:48:29 +0000]
bnx2x: Not dropping packets with L3/L4 checksum error

Not dropping packets with L3/L4 checksum error
Those packets should be passed to the OS. The problem is clear in
forwarding mode.

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agobnx2x: FW (bootcode) interface fixes
Eilon Greenstein [Wed, 13 Aug 2008 22:47:33 +0000]
bnx2x: FW (bootcode) interface fixes

FW (bootcode) interface fixes
- Making sure that the device will not cause kernel panic of the
  bootcode is corrupted or missing
- Removing module debug parameter "nomcp" since no one should work
  without the bootcode (this is a left over from the chip bring up days)
- Instead of waiting fix amount of time for bootcode response, sample it
  every 10ms (usually the answer is ready after less than 10ms)

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Wed, 13 Aug 2008 22:24:35 +0000]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()
  crypto: hash - Add missing top-level functions
  crypto: hash - Fix digest size check for digest type
  crypto: tcrypt - Fix AEAD chunk testing
  crypto: talitos - Add handling for SEC 3.x treatment of link table

10 years agopkt_sched: Protect gen estimators under est_lock.
Jarek Poplawski [Wed, 13 Aug 2008 22:20:24 +0000]
pkt_sched: Protect gen estimators under est_lock.

gen_kill_estimator() required rtnl_lock() protection, but since it is
moved to an RCU callback __qdisc_destroy() let's use est_lock instead.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agopkt_sched: Fix queue quiescence testing in dev_deactivate().
David S. Miller [Wed, 13 Aug 2008 22:18:38 +0000]
pkt_sched: Fix queue quiescence testing in dev_deactivate().

Based upon discussions with Jarek P. and Herbert Xu.

First, we're testing the wrong qdisc.  We just reset the device
queue qdiscs to &noop_qdisc and checking it's state is completely
pointless here.

We want to wait until the previous qdisc that was sitting at
the ->qdisc pointer is not busy any more.  And that would be
->qdisc_sleeping.

Because of how we propagate the samples qdisc pointer down into
qdisc_run and friends via per-cpu ->output_queue and netif_schedule,
we have to wait also for the __QDISC_STATE_SCHED bit to clear as
well.

Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agoMerge git://oss.sgi.com:8090/xfs/linux-2.6
Linus Torvalds [Wed, 13 Aug 2008 22:17:49 +0000]
Merge git://oss.sgi.com:8090/xfs/linux-2.6

* git://oss.sgi.com:8090/xfs/linux-2.6: (45 commits)
  [XFS] Fix use after free in xfs_log_done().
  [XFS] Make xfs_bmap_*_count_leaves void.
  [XFS] Use KM_NOFS for debug trace buffers
  [XFS] use KM_MAYFAIL in xfs_mountfs
  [XFS] refactor xfs_mount_free
  [XFS] don't call xfs_freesb from xfs_unmountfs
  [XFS] xfs_unmountfs should return void
  [XFS] cleanup xfs_mountfs
  [XFS] move root inode IRELE into xfs_unmountfs
  [XFS] stop using file_update_time
  [XFS] optimize xfs_ichgtime
  [XFS] update timestamp in xfs_ialloc manually
  [XFS] remove the sema_t from XFS.
  [XFS] replace dquot flush semaphore with a completion
  [XFS] replace inode flush semaphore with a completion
  [XFS] extend completions to provide XFS object flush requirements
  [XFS] replace the XFS buf iodone semaphore with a completion
  [XFS] clean up stale references to semaphores
  [XFS] use get_unaligned_* helpers
  [XFS] Fix compile failure in xfs_buf_trace()
  ...

10 years agopkt_sched: Fix oops in htb_delete.
Jarek Poplawski [Wed, 13 Aug 2008 22:16:43 +0000]
pkt_sched: Fix oops in htb_delete.

Recent changes introduced a bug in htb_delete(): cl->parent->children
counter update misses checking cl->parent for NULL, which is used for
root classes, so deleting them causes an oops.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm
Linus Torvalds [Wed, 13 Aug 2008 22:16:10 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: rename structs
  dlm: add missing kfrees

10 years agopktgen: prevent pktgen from using bad tx queue
Andrew Gallatin [Wed, 13 Aug 2008 22:16:00 +0000]
pktgen: prevent pktgen from using bad tx queue

With the new multi-queue transmit code, it is possible to accidentally
make pktgen pick a non-existing tx queue simply by using a stale
script to drive pktgen.  Access to this non-existing tx queue will
then trigger a bad memory access and kill the machine.

For example, setting "queue_map_max 2" will cause my machine to die
when accessing a garbage spinlock in the non-existing tx queue:

BUG: spinlock bad magic on CPU#0, kpktgend_0/564
  lock: ffff88001ddf6718, .magic: ffffffff, .owner: /-1, .owner_cpu: 0
Pid: 564, comm: kpktgend_0 Not tainted 2.6.27-rc3 #35

Call Trace:
  [<ffffffff803a1228>] spin_bug+0xa4/0xac
  [<ffffffff803a1253>] _raw_spin_lock+0x23/0x123
  [<ffffffff8055b06f>] _spin_lock_bh+0x17/0x1b
  [<ffffffff804cb57d>] pktgen_thread_worker+0xa97/0x1002
  [<ffffffff8022874d>] ? finish_task_switch+0x38/0x97
  [<ffffffff80242077>] ? autoremove_wake_function+0x0/0x36
  [<ffffffff80242077>] ? autoremove_wake_function+0x0/0x36
  [<ffffffff804caae6>] ? pktgen_thread_worker+0x0/0x1002
  [<ffffffff80241a40>] kthread+0x44/0x6d
  [<ffffffff8020c399>] child_rip+0xa/0x11
  [<ffffffff802419fc>] ? kthread+0x0/0x6d
  [<ffffffff8020c38f>] ? child_rip+0x0/0x11

The attached patch adds some sanity checking to prevent
these sorts of configuration errors.

Signed-off-by: Andrew Gallatin <gallatin@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years ago[h8300] move include/asm-h8300 to arch/h8300/include/asm
Linus Torvalds [Wed, 13 Aug 2008 21:26:22 +0000]
[h8300] move include/asm-h8300 to arch/h8300/include/asm

Done as a script (well, a single "git mv" actually) on request from
Yoshinori Sato as a way to avoid a huge diff.

Requested-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agodccp: change L/R must have at least one byte in the dccpsf_val field
Arnaldo Carvalho de Melo [Wed, 13 Aug 2008 20:48:39 +0000]
dccp: change L/R must have at least one byte in the dccpsf_val field

Thanks to Eugene Teo for reporting this problem.

Signed-off-by: Eugene Teo <eugenete@kernel.sg>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agoxfrm: remove unnecessary variable in xfrm_output_resume() 2nd try
Jean-Christophe DUBOIS [Wed, 13 Aug 2008 20:35:37 +0000]
xfrm: remove unnecessary variable in xfrm_output_resume() 2nd try

Small fix removing an unnecessary intermediate variable.

Signed-off-by: Jean-Christophe DUBOIS <jcd@tribudubois.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

10 years agodlm: rename structs
David Teigland [Thu, 31 Jul 2008 14:31:53 +0000]
dlm: rename structs

Add a dlm_ prefix to the struct names in config.c.  This resolves a
conflict with struct node in particular, when include/linux/node.h
happens to be included.

Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Teigland <teigland@redhat.com>

10 years agodlm: add missing kfrees
David Teigland [Tue, 29 Jul 2008 20:21:19 +0000]
dlm: add missing kfrees

A couple of unlikely error conditions were missing a kfree on the error
exit path.

Reported-by: Juha Leppanen <juha_motorsportcom@luukku.com>
Signed-off-by: David Teigland <teigland@redhat.com>

10 years agocrypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()
Suresh Siddha [Wed, 13 Aug 2008 12:02:26 +0000]
crypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()

Wolfgang Walter reported this oops on his via C3 using padlock for
AES-encryption:

##################################################################

BUG: unable to handle kernel NULL pointer dereference at 000001f0
IP: [<c01028c5>] __switch_to+0x30/0x117
*pde = 00000000
Oops: 0002 [#1] PREEMPT
Modules linked in:

Pid: 2071, comm: sleep Not tainted (2.6.26 #11)
EIP: 0060:[<c01028c5>] EFLAGS: 00010002 CPU: 0
EIP is at __switch_to+0x30/0x117
EAX: 00000000 EBX: c0493300 ECX: dc48dd00 EDX: c0493300
ESI: dc48dd00 EDI: c0493530 EBP: c04cff8c ESP: c04cff7c
 DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
Process sleep (pid: 2071, ti=c04ce000 task=dc48dd00 task.ti=d2fe6000)
Stack: dc48df30 c0493300 00000000 00000000 d2fe7f44 c03b5b43 c04cffc8 00000046
       c0131856 0000005a dc472d3c c0493300 c0493470 d983ae00 00002696 00000000
       c0239f54 00000000 c04c4000 c04cffd8 c01025fe c04f3740 00049800 c04cffe0
Call Trace:
 [<c03b5b43>] ? schedule+0x285/0x2ff
 [<c0131856>] ? pm_qos_requirement+0x3c/0x53
 [<c0239f54>] ? acpi_processor_idle+0x0/0x434
 [<c01025fe>] ? cpu_idle+0x73/0x7f
 [<c03a4dcd>] ? rest_init+0x61/0x63
 =======================

Wolfgang also found out that adding kernel_fpu_begin() and kernel_fpu_end()
around the padlock instructions fix the oops.

Suresh wrote:

These padlock instructions though don't use/touch SSE registers, but it behaves
similar to other SSE instructions. For example, it might cause DNA faults
when cr0.ts is set. While this is a spurious DNA trap, it might cause
oops with the recent fpu code changes.

This is the code sequence  that is probably causing this problem:

a) new app is getting exec'd and it is somewhere in between
   start_thread() and flush_old_exec() in the load_xyz_binary()

b) At pont "a", task's fpu state (like TS_USEDFPU, used_math() etc) is
   cleared.

c) Now we get an interrupt/softirq which starts using these encrypt/decrypt
   routines in the network stack. This generates a math fault (as
   cr0.ts is '1') which sets TS_USEDFPU and restores the math that is
   in the task's xstate.

d) Return to exec code path, which does start_thread() which does
   free_thread_xstate() and sets xstate pointer to NULL while
   the TS_USEDFPU is still set.

e) At the next context switch from the new exec'd task to another task,
   we have a scenarios where TS_USEDFPU is set but xstate pointer is null.
   This can cause an oops during unlazy_fpu() in __switch_to()

Now:

1) This should happen with or with out pre-emption. Viro also encountered
   similar problem with out CONFIG_PREEMPT.

2) kernel_fpu_begin() and kernel_fpu_end() will fix this problem, because
   kernel_fpu_begin() will manually do a clts() and won't run in to the
   situation of setting TS_USEDFPU in step "c" above.

3) This was working before the fpu changes, because its a spurious
   math fault  which doesn't corrupt any fpu/sse registers and the task's
   math state was always in an allocated state.

With out the recent lazy fpu allocation changes, while we don't see oops,
there is a possible race still present in older kernels(for example,
while kernel is using kernel_fpu_begin() in some optimized clear/copy
page and an interrupt/softirq happens which uses these padlock
instructions generating DNA fault).

This is the failing scenario that existed even before the lazy fpu allocation
changes:

0. CPU's TS flag is set

1. kernel using FPU in some optimized copy  routine and while doing
kernel_fpu_begin() takes an interrupt just before doing clts()

2. Takes an interrupt and ipsec uses padlock instruction. And we
take a DNA fault as TS flag is still set.

3. We handle the DNA fault and set TS_USEDFPU and clear cr0.ts

4. We complete the padlock routine

5. Go back to step-1, which resumes clts() in kernel_fpu_begin(), finishes
the optimized copy routine and does kernel_fpu_end(). At this point,
we have cr0.ts again set to '1' but the task's TS_USEFPU is stilll
set and not cleared.

6. Now kernel resumes its user operation. And at the next context
switch, kernel sees it has do a FP save as TS_USEDFPU is still set
and then will do a unlazy_fpu() in __switch_to(). unlazy_fpu()
will take a DNA fault, as cr0.ts is '1' and now, because we are
in __switch_to(), math_state_restore() will get confused and will
restore the next task's FP state and will save it in prev tasks's FP state.
Remember, in __switch_to() we are already on the stack of the next task
but take a DNA fault for the prev task.

This causes the fpu leakage.

Fix the padlock instruction usage by calling them inside the
context of new routines irq_ts_save/restore(), which clear/restore cr0.ts
manually in the interrupt context. This will not generate spurious DNA
in the  context of the interrupt which will fix the oops encountered and
the possible FPU leakage issue.

Reported-and-bisected-by: Wolfgang Walter <wolfgang.walter@stwm.de>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>