linux-2.6
14 years agoPM / ACPI: HP G7000 Notebook needs a SCI_EN resume quirk
Bartlomiej Zolnierkiewicz [Wed, 29 Jul 2009 19:07:47 +0000 (21:07 +0200)] 
PM / ACPI: HP G7000 Notebook needs a SCI_EN resume quirk

This fixes regression (battery "vanishing" on resume) introduced by
commit d0c71fe7ebc180f1b7bc7da1d39a07fc19eec768 ("ACPI Suspend: Enable
ACPI during resume if SCI_EN is not set") and also the issue with
the "screaming" IRQ 9.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=13745

Reported-and-tested-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Cc: stable@kernel.org
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Acked-by: Len Brown <lenb@kernel.org>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agodrm/radeon: set fb aperture sizes for framebuffer handoff.
Dave Airlie [Wed, 29 Jul 2009 07:07:38 +0000 (17:07 +1000)] 
drm/radeon: set fb aperture sizes for framebuffer handoff.

This will allow efi/vesa to handoff to radeon.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: fix highuser vs dma32 confusion.
Dave Airlie [Wed, 29 Jul 2009 06:56:52 +0000 (16:56 +1000)] 
drm/ttm: fix highuser vs dma32 confusion.

DMA32 and highmem are sort of exclusive.

Noticed by AndrewR on #radeon.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon: Fix size used for benchmarking BO copies.
Michel Dänzer [Sun, 19 Jul 2009 23:44:03 +0000 (01:44 +0200)] 
drm/radeon: Fix size used for benchmarking BO copies.

The incorrect size caused benchmark results to be inflated by a factor of 4.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon: Add radeon.test parameter for running BO GPU copy tests.
Michel Dänzer [Tue, 21 Jul 2009 09:23:57 +0000 (11:23 +0200)] 
drm/radeon: Add radeon.test parameter for running BO GPU copy tests.

If enabled, during initialization BO GTT->VRAM and VRAM->GTT GPU copies are
tested across the whole GTT aperture.

This has helped uncover the benchmark copy size bug and verify the maximum
aperture size supported by the AGP bridge in my PowerBook.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: allow interruptible waits for objects.
Dave Airlie [Fri, 24 Jul 2009 03:47:45 +0000 (13:47 +1000)] 
drm/radeon/kms: allow interruptible waits for objects.

Blocking here isn't something the X server mouse appreciates,
avoid the block and let userspace retry the waits.

libdrm_radeon userspace library is also expecting EBUSY not ERESTART

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: powerpc: Fix Highmem cache flushing.
Thomas Hellstrom [Fri, 24 Jul 2009 08:22:36 +0000 (10:22 +0200)] 
drm/ttm: powerpc: Fix Highmem cache flushing.

Temporarily maps highmem pages while flushing to get a valid virtual
address to flush.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agox86: Export kmap_atomic_prot() needed for TTM.
Thomas Hellstrom [Fri, 24 Jul 2009 07:57:33 +0000 (09:57 +0200)] 
x86: Export kmap_atomic_prot() needed for TTM.

This functionality is needed to kmap_atomic() highmem pages that may
potentially have or are about to set up other mappings with
non-standard caching attributes.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: Fix ttm in-kernel copying of pages with non-standard caching attributes.
Thomas Hellstrom [Fri, 24 Jul 2009 07:57:34 +0000 (09:57 +0200)] 
drm/ttm: Fix ttm in-kernel copying of pages with non-standard caching attributes.

For x86 this affected highmem pages only, since they were always kmapped
cache-coherent, and this is fixed using kmap_atomic_prot().

For other architectures that may not modify the linear kernel map we
resort to vmap() for now, since kmap_atomic_prot() generally uses the
linear kernel map for lowmem pages. This of course comes with a
performance impact and should be optimized when possible.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: Fix an oops and sync object leak.
Thomas Hellstrom [Tue, 21 Jul 2009 15:45:13 +0000 (17:45 +0200)] 
drm/ttm: Fix an oops and sync object leak.

The code was potentially dereferencig a NULL sync object pointer.
At the same time a sync object reference was potentially leaked.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: vram sizing on certain r100 chips needs workaround.
Dave Airlie [Tue, 21 Jul 2009 10:39:30 +0000 (20:39 +1000)] 
drm/radeon/kms: vram sizing on certain r100 chips needs workaround.

If an rn50/r100/m6/m7 GPU has < 64MB RAM, i.e. 8/16/32, the
aperture used to calculate the MC_FB_LOCATION needs to be worked
out from the CONFIG_APER_SIZE register, and not the actual vram size.

TTM VRAM size was also being initialised wrong, use actual vram size
to initialise it.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon: Pay more attention to object placement requested by userspace.
Michel Dänzer [Tue, 28 Jul 2009 10:30:57 +0000 (12:30 +0200)] 
drm/radeon: Pay more attention to object placement requested by userspace.

Previously we were basically always setting the GTT and VRAM flags regardless of
what userspace requested.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon: Fall back to evicting BOs with memcpy if necessary.
Michel Dänzer [Tue, 28 Jul 2009 10:30:56 +0000 (12:30 +0200)] 
drm/radeon: Fall back to evicting BOs with memcpy if necessary.

Otherwise if there's no GTT space we would fail the eviction, leading to
cascaded failure.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon: Don't unreserve twice on failure to validate.
Michel Dänzer [Tue, 28 Jul 2009 10:30:55 +0000 (12:30 +0200)] 
drm/radeon: Don't unreserve twice on failure to validate.

This is done later in radeon_object_list_unvalidate(). Doing it twice triggers
a BUG in TTM, rendering X on KMS unusable until reboot.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix bandwidth computation on avivo hardware
Jerome Glisse [Mon, 13 Jul 2009 19:04:08 +0000 (21:04 +0200)] 
drm/radeon/kms: fix bandwidth computation on avivo hardware

Fix bandwidth computation and crtc priority in memory controller
so that crtc memory request are fullfill in time to avoid display
artifact.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: add initial colortiling support.
Dave Airlie [Tue, 23 Jun 2009 23:48:08 +0000 (09:48 +1000)] 
drm/radeon/kms: add initial colortiling support.

This adds new set/get tiling interfaces where the pitch
and macro/micro tiling enables can be set. Along with
a flag to decide if this object should have a surface when mapped.

The only thing we need to allocate with a mapped surface should be
the frontbuffer. Note rotate scanout shouldn't require one, and
back/depth shouldn't either, though mesa needs some fixes.

It fixes the TTM interfaces along Thomas's suggestions, and I've tested
the surface stealing code with two X servers and not seen any lockdep issues.

I've stopped tiling the fbcon frontbuffer, as I don't see there being
any advantage other than testing, I've left the testing commands in there,
just flip the fb_tiled to true in radeon_fb.c

Open: Can we integrate endian swapping in with this?

Future features:
texture tiling - need to relocate texture registers TXOFFSET* with tiling info.

This also merges Michel's cleanup surfaces regs at init time patch
even though it makes sense on its own, this patch really relies on it.

Some PowerMac firmwares set up a tiling surface at the beginning of VRAM
which messes us up otherwise.
that patch is:
Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoRemove zero-length file drivers/char/vr41xx_giu.c
Jeff Garzik [Wed, 29 Jul 2009 02:36:59 +0000 (22:36 -0400)] 
Remove zero-length file drivers/char/vr41xx_giu.c

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: accept late unlocking of HPA
Tejun Heo [Wed, 15 Jul 2009 08:11:41 +0000 (17:11 +0900)] 
libata: accept late unlocking of HPA

On certain configurations, HPA isn't or can't be unlocked during
probing but it somehow ends up unlocked afterwards.  In the following
thread, the problem can be reliably reproduced after resuming from
STR.  The BIOS turns on HPA during boot but forgets to do it during
resume.

  http://thread.gmane.org/gmane.linux.kernel/858310

This patch updates libata revalidation such that it considers native
n_sectors.  If the device size has increased to match native
n_sectors, it's assumed that HPA has been unlocked involuntarily and
the device is recognized as the same one.  This should be fairly safe
while nicely working around the problem.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Christof Warlich <christof@warlich.name>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: Updates and fixes for pata_at91 driver
Sergey Matyukevich [Thu, 16 Jul 2009 18:38:55 +0000 (22:38 +0400)] 
libata: Updates and fixes for pata_at91 driver

Please consider the following updates and fixes for pata_at91 driver.

* Removed extra headers
Here we need only static memory controller properties, which are
contained in generic header at91sam9_smc.h.
No need to include any specific headers for at91sam9260 SoC.

* No harsh BUG_ON for get_clk in set_smc_timing function
        get_clk is now performed in driver probing function,
        probing fails if master clock is not available

* Fixed uint/ulong mess in calc_mck_cycles function

Signed-off-by: Sergey Matyukevich <geomatsi@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoata_piix: Add new short cable ID
Steve Conklin [Thu, 16 Jul 2009 21:31:10 +0000 (16:31 -0500)] 
ata_piix: Add new short cable ID

OriginalAuthor: Tony Espy <espy@canonical.com>
Signed-off-by: Steve Conklin <sconklin@canonical.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoata_piix: Add new laptop short cable IDs
Steve Conklin [Thu, 16 Jul 2009 21:27:56 +0000 (16:27 -0500)] 
ata_piix: Add new laptop short cable IDs

OriginalAuthor: Michael Frey <michael.frey@canonical.com>
Signed-off-by: Steve Conklin <sconklin@canonical.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoahci: add device IDs for Ibex Peak ahci controllers
David Milburn [Wed, 22 Jul 2009 20:15:56 +0000 (15:15 -0500)] 
ahci: add device IDs for Ibex Peak ahci controllers

Add device IDS for Ibex Peak SATA AHCI Controllers

Signed-off-by: Jaroslav Kysela <jkysela@redhat.com>
Signed-off-by: David Milburn <dmilburn@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: remove superfluous NULL pointer checks
Bartlomiej Zolnierkiewicz [Sun, 26 Jul 2009 14:05:13 +0000 (16:05 +0200)] 
libata: remove superfluous NULL pointer checks

host->ports[] always contain pointers to valid port structures since
a "dummy port" structure is used in case if there is no physical port.

This patch takes care of two entries from Dan's list:

drivers/ata/sata_sil.c +535 sil_interrupt(13) warning: variable derefenced before check 'ap'
drivers/ata/sata_mv.c +2517 mv_unexpected_intr(6) warning: variable derefenced before check 'ap'

and of another needless NULL pointer check in pata_octeon_cf.c.

Reported-by: Dan Carpenter <error27@gmail.com>
Cc: corbet@lwn.net
Cc: eteo@redhat.com
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: add missing NULL pointer check to ata_eh_reset()
Bartlomiej Zolnierkiewicz [Sun, 26 Jul 2009 14:21:01 +0000 (16:21 +0200)] 
libata: add missing NULL pointer check to ata_eh_reset()

drivers/ata/libata-eh.c +2403 ata_eh_reset(80) warning: variable derefenced before check 'slave'

Please note that this is _not_ a real bug at the moment since ata_eh_context
structure is embedded into ata_list structure and the code alwas checks for
'slave' before accessing 'sehc'.

Anyway lets add missing check and always have a valid 'sehc' pointer (which
makes code easier to understand and prevents introducing some possible bugs
in the future).

Reported-by: Dan Carpenter <error27@gmail.com>
Cc: corbet@lwn.net
Cc: eteo@redhat.com
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agopata_pcmcia: add CNF-CDROM-ID
Wolfram Sang [Sun, 26 Jul 2009 16:27:29 +0000 (18:27 +0200)] 
pata_pcmcia: add CNF-CDROM-ID

Fixes this report:
http://article.gmane.org/gmane.linux.kernel.pcmcia.devel/2228/

Reported-by: John McGrath <john@john-mcgrath.com>
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoUSB: xhci: Stall handling bug fixes.
Sarah Sharp [Mon, 27 Jul 2009 19:05:21 +0000 (12:05 -0700)] 
USB: xhci: Stall handling bug fixes.

Correct the xHCI code to handle stalls on USB endpoints.  We need to move
the endpoint ring's dequeue pointer past the stalled transfer, or the HW
will try to restart the transfer the next time the doorbell is rung.

Don't attempt to clear a halt on an endpoint if we haven't seen a stalled
transfer for it.  The USB core will attempt to clear a halt on all
endpoints when it selects a new configuration.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelv...
Linus Torvalds [Tue, 28 Jul 2009 21:31:13 +0000 (14:31 -0700)] 
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: (asus_atk0110) Fix upper limit readings
  hwmon: (smsc47m1) Differentiate between LPC47M233 and LPC47M292

14 years agoUSB: xhci: Support for 64-byte contexts
John Youn [Mon, 27 Jul 2009 19:05:15 +0000 (12:05 -0700)] 
USB: xhci: Support for 64-byte contexts

Adds support for controllers that use 64-byte contexts.  The following context
data structures are affected by this: Device, Input, Input Control, Endpoint,
and Slot.  To accommodate the use of either 32 or 64-byte contexts, a Device or
Input context can only be accessed through functions which look-up and return
pointers to their contained contexts.

Signed-off-by: John Youn <johnyoun@synopsys.com>
Acked-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Always align output device contexts to 64 bytes.
Sarah Sharp [Mon, 27 Jul 2009 19:05:08 +0000 (12:05 -0700)] 
USB: xhci: Always align output device contexts to 64 bytes.

Make sure the xHCI output device context is 64-byte aligned.  Previous
code was using the same structure for both the output device context and
the input control context.  Since the structure had 32 bytes of flags
before the device context, the output device context wouldn't be 64-byte
aligned.  Define a new structure to use for the output device context and
clean up the debugging for these two structures.

The copy of the device context in the input control context does *not*
need to be 64-byte aligned.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Scratchpad buffer allocation
John Youn [Mon, 27 Jul 2009 19:05:03 +0000 (12:05 -0700)] 
USB: xhci: Scratchpad buffer allocation

Allocates and initializes the scratchpad buffer array (XHCI 4.20).  This is an
array of 64-bit DMA addresses to scratch pages that the controller may use
during operation.  The number of pages is specified in the "Max Scratchpad
Buffers" field of HCSPARAMS2.  The DMA address of this array is written into
slot 0 of the DCBAA.

Signed-off-by: John Youn <johnyoun@synopsys.com>
Acked-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Fix parsing of SuperSpeed Endpoint Companion descriptor.
Sarah Sharp [Mon, 27 Jul 2009 19:04:52 +0000 (12:04 -0700)] 
USB: Fix parsing of SuperSpeed Endpoint Companion descriptor.

usb_parse_ss_endpoint_companion() was supposed to allocate a structure to
hold the SuperSpeed Endpoint Companion descriptor, and either copy the
values the device returned, or fill in default values if the device
descriptor did not include the companion descriptor.

However, the previous code would miss the last endpoint in a configuration
with no descriptors after it.  Make usb_parse_endpoint() allocate the SS
endpoint companion descriptor and fill it with default values, even if
we've run out of buffer space in this configuration descriptor.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Fail gracefully if there's no SS ep companion descriptor.
Sarah Sharp [Mon, 27 Jul 2009 19:04:38 +0000 (12:04 -0700)] 
USB: xhci: Fail gracefully if there's no SS ep companion descriptor.

This is a work around for a bug in the SuperSpeed Endpoint Companion Descriptor
parsing code.  It fails in some corner cases, which means ep->ss_ep_comp may be
NULL.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Handle babble errors on transfers.
Sarah Sharp [Mon, 27 Jul 2009 19:04:32 +0000 (12:04 -0700)] 
USB: xhci: Handle babble errors on transfers.

Pass back a babble error when this error code is seen in the transfer event TRB.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Setup HW retries correctly.
Sarah Sharp [Mon, 27 Jul 2009 19:04:27 +0000 (12:04 -0700)] 
USB: xhci: Setup HW retries correctly.

The xHCI host controller can be programmed to retry a transfer a certain number
of times per endpoint before it passes back an error condition to the host
controller driver.  The xHC will return an error code when the error count
transitions from 1 to 0.  Programming an error count of 3 means the xHC tries
the transfer 3 times, programming it with a 1 means it tries to transfer once,
and programming it with 0 means the HW tries the transfer infinitely.

We want isochronous transfers to only be tried once, so set the error count to
one.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Check if the host controller died in IRQ handler.
Sarah Sharp [Mon, 27 Jul 2009 19:04:01 +0000 (12:04 -0700)] 
USB: xhci: Check if the host controller died in IRQ handler.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Don't oops if the host doesn't halt.
Sarah Sharp [Mon, 27 Jul 2009 19:03:50 +0000 (12:03 -0700)] 
USB: xhci: Don't oops if the host doesn't halt.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Make debugging more verbose.
Sarah Sharp [Mon, 27 Jul 2009 19:03:46 +0000 (12:03 -0700)] 
USB: xhci: Make debugging more verbose.

Add more debugging to the irq handler, slot context initialization, ring
operations, URB cancellation, and MMIO writes.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Correct Event Handler Busy flag usage.
Sarah Sharp [Mon, 27 Jul 2009 19:03:40 +0000 (12:03 -0700)] 
USB: xhci: Correct Event Handler Busy flag usage.

The Event Handler Busy bit in the event ring dequeue pointer is write 1 to
clear.  Fix the interrupt service routine to clear that bit after the
event handler has run.

xhci_set_hc_event_deq() is designed to update the event ring dequeue pointer
without changing any of the four reserved bits in the lower nibble.  The event
handler busy (EHB) bit is write one to clear, so the new value must always
contain a zero in that bit in order to preserve the EHB value.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Handle short control packets correctly.
Sarah Sharp [Mon, 27 Jul 2009 19:03:36 +0000 (12:03 -0700)] 
USB: xhci: Handle short control packets correctly.

When there is a short packet on a control transfer, the xHCI host controller
hardware will generate two events.  The first event will be for the data stage
TD with a completion code for a short packet.  The second event will be for the
status stage with a successful completion code.  Before this patch, the xHCI
driver would giveback the short control URB when it received the event for the
data stage TD.  Then it would become confused when it saw a status stage event
for the endpoint for an URB it had already finished processing.

Change the xHCI host controller driver to wait for the status stage event when
it receives a short transfer completion code for a data stage TD.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Represent 64-bit addresses with one u64.
Sarah Sharp [Mon, 27 Jul 2009 19:03:31 +0000 (12:03 -0700)] 
USB: xhci: Represent 64-bit addresses with one u64.

There are several xHCI data structures that use two 32-bit fields to
represent a 64-bit address.  Since some architectures don't support 64-bit
PCI writes, the fields need to be written in two 32-bit writes.  The xHCI
specification says that if a platform is incapable of generating 64-bit
writes, software must write the low 32-bits first, then the high 32-bits.
Hardware that supports 64-bit addressing will wait for the high 32-bit
write before reading the revised value, and hardware that only supports
32-bit writes will ignore the high 32-bit write.

Previous xHCI code represented 64-bit addresses with two u32 values.  This
lead to buggy code that would write the 32-bits in the wrong order, or
forget to write the upper 32-bits.  Change the two u32s to one u64 and
create a function call to write all 64-bit addresses in the proper order.
This new function could be modified in the future if all platforms support
64-bit writes.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Use GFP_ATOMIC while holding spinlocks.
Sarah Sharp [Mon, 27 Jul 2009 19:03:23 +0000 (12:03 -0700)] 
USB: xhci: Use GFP_ATOMIC while holding spinlocks.

The xHCI functions to queue an URB onto the hardware rings must be called
with the xhci spinlock held.  Those functions will allocate memory, and
take a gfp_t memory flags argument.  We must pass them the GFP_ATOMIC
flag, since we don't want the memory allocation to attempt to sleep while
waiting for more memory to become available.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Deal with stalled endpoints.
Sarah Sharp [Mon, 27 Jul 2009 19:03:15 +0000 (12:03 -0700)] 
USB: xhci: Deal with stalled endpoints.

When an endpoint on a device under an xHCI host controller stalls, the
host controller driver must let the hardware know that the USB core has
successfully cleared the halt condition.  The HCD submits a Reset Endpoint
Command, which will clear the toggle bit for USB 2.0 devices, and set the
sequence number to zero for USB 3.0 devices.

The xHCI urb_enqueue will accept new URBs while the endpoint is halted,
and will queue them to the hardware rings.  However, the endpoint doorbell
will not be rung until the Reset Endpoint Command completes.

Don't queue a reset endpoint command for root hubs.  khubd clears halt
conditions on the roothub during the initialization process, but the roothub
isn't a real device, so the xHCI host controller doesn't need to know about the
cleared halt.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Set TD size in transfer TRB.
Sarah Sharp [Mon, 27 Jul 2009 19:03:07 +0000 (12:03 -0700)] 
USB: xhci: Set TD size in transfer TRB.

The 0.95 xHCI specification requires software to set the "TD size" field
in each transaction request block (TRB).  This field gives the host
controller an indication of how much data is remaining in the TD
(including the buffer in the current TRB).  Set this field in bulk TRBs
and data stage TRBs for control transfers.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: fix less- and greater than confusion
Roel Kluin [Sun, 12 Jul 2009 22:23:47 +0000 (00:23 +0200)] 
USB: xhci: fix less- and greater than confusion

Without this change the loops won't start

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Cc: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: usbtest: no need for USB_DEVICEFS
Alan Stern [Tue, 28 Jul 2009 15:56:17 +0000 (11:56 -0400)] 
USB: usbtest: no need for USB_DEVICEFS

THis patch (as1270) allows the usbtest module to be built even when
USB_DEVICEFS isn't configured.  Tests can be performed without
USB_DEVICEFS, using the /dev/bus/usb/*/* device files.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: fix CONFIGDATA register read issue
Ajay Kumar Gupta [Mon, 27 Jul 2009 21:32:15 +0000 (14:32 -0700)] 
USB: musb: fix CONFIGDATA register read issue

INDEX register has to be set to '0' before reading
CONFIGDATA register which is only present in TI musb
platforms.

Currently the default register access mode is set to
FLAT_MODE thus INDEX register is not getting set
properly with musb_ep_select() which is just a nop
operation in FLAT_MODE.This invalid register read is
causing module reinset failure.

Fixing the issue by moving INDEX register write part to
musb_read_configdata() function itself.

Signed-off-by: Vikram Pandita <vikram.pandita@ti.com>
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: Refer to musb_otg_timer_func under correct #ifdef
Amit Kucheria [Mon, 27 Jul 2009 09:03:19 +0000 (12:03 +0300)] 
USB: musb: Refer to musb_otg_timer_func under correct #ifdef

musb_otg_timer_func() is defined under #ifdef CONFIG_USB_MUSB_OTG.
Make sure any reference to it is also under the same #ifdef.

Without this fix, the driver failes to compile when USB_OTG is defined
but USB_MUSB_OTG isn't.

Signed-off-by: Amit Kucheria <amit.kucheria@canonical.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb_gadget_ep0: fix typo in service_zero_data_request()
Sergei Shtylyov [Fri, 17 Jul 2009 14:30:03 +0000 (17:30 +0300)] 
USB: musb_gadget_ep0: fix typo in service_zero_data_request()

This function uses wrong bit mask to prevent clearing RXCSR status
bits when halting an endpoint -- which results in clearing SentStall
and RxPktRdy bits (that the code actually tries to avoid); must be
a result of cut-and-paste...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option.c to support Alcatel X060S/X200 broadband modems
Javier Martin [Thu, 23 Jul 2009 02:03:43 +0000 (04:03 +0200)] 
USB: option.c to support Alcatel X060S/X200 broadband modems

Added support for the Alcatel X060S/X200 broadband modems to the option
driver. The device starts in cd-rom emulation mode (1bbb:f000) and
requires the use of the usb_modeswitch tool to switch it to modem mode
(1bbb:0000).

Signed-off-by: Javier Martin <jmartinj@iname.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: serial: option: Add ZTE AC8710 usb modem device.
Peng Huang [Wed, 22 Jul 2009 09:54:48 +0000 (11:54 +0200)] 
USB: serial: option: Add ZTE AC8710 usb modem device.

Signed-off-by: Peng Huang <shawn.p.huang@gmail.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: aten uc2324 is really a moschip 7840
Russell Lang [Fri, 17 Jul 2009 09:29:20 +0000 (19:29 +1000)] 
USB: aten uc2324 is really a moschip 7840

I've opened up the case, and the chips in the ATEN UC2324 are:

Moschip
MCS7840CV-AA
69507-6B1
0650
(USB to 4-port serial)

(logo with AF kerned together)  0748
24BC02
SINGLP
(unknown 8-pin chip)

(logo looks like 3 or Z in circle)
ZT3243LEEA    0752
B7A16420.T
(4 chips, so this will be RS232 line driver)
(Probably equivalent of Sipex SP3243)

So the ATEN 2324 (aten2011.c driver), is definitely the Moschip 7840,
and should use the mos7840.c driver.  I expect you will remove the
aten2011.c driver from the staging area.

From the aten2011.c source code, the device ID for the UC2322 (2 port
serial) is 0x7820, just like the Moschip evaluation board.  This value
should be added to the device id table of mos7840.c.

Here's a patch that adds these devices to the driver.

From: Russell Lang <gsview@ghostgum.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option: Add USB ID for Novatel MC727/U727/USB727 refresh
Tim Gardner [Fri, 17 Jul 2009 13:55:36 +0000 (07:55 -0600)] 
USB: option: Add USB ID for Novatel MC727/U727/USB727 refresh

BugLink: https://bugs.launchpad.net/bugs/365291
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option: add ZTE device ids and remove ONDA ids
Herton Ronaldo Krzesinski [Wed, 15 Jul 2009 20:10:26 +0000 (17:10 -0300)] 
USB: option: add ZTE device ids and remove ONDA ids

Current listed Onda ids are ZTE devices. Replace them with ZTE id define
and add more ZTE device ids. Also remove 19d2:2000, this is the id when
device is first plugged in and is a CD-only device, before the switch
using eject.

These changes are based on a previous patch by Ming Zhao
<zhao.ming9@zte.com.cn>

Signed-off-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br>
Cc: Ming Zhao <zhao.ming9@zte.com.cn>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: ehci-orion: Call ehci_reset before ehci_halt
Simon Kagstrom [Mon, 13 Jul 2009 06:25:09 +0000 (08:25 +0200)] 
USB: ehci-orion: Call ehci_reset before ehci_halt

I noticed that USB initialization didn't setup correctly on my kirkwood
based board (OpenRD base) if I hadn't initialized USB in U-boot first.
The error message looks like this:

  ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
  orion-ehci orion-ehci.0: Marvell Orion EHCI
  orion-ehci orion-ehci.0: new USB bus registered, assigned bus number 1
  orion-ehci orion-ehci.0: can't setup
  orion-ehci orion-ehci.0: USB bus 1 deregistered
  orion-ehci orion-ehci.0: init orion-ehci.0 fail, -110
  orion-ehci: probe of orion-ehci.0 failed with error -110

which is caused by ehci_halt() timing out in the handshake() call. I
noticed that U-boot does a reset before calling handshake(), so this
patch does the same thing for Linux. USB now works for me.

Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
Acked-by: Nicolas Pitre <nico@marvell.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: CP210x Add new device IDs
Craig Shelley [Sun, 12 Jul 2009 20:52:33 +0000 (21:52 +0100)] 
USB: CP210x Add new device IDs

Signed-off-by: Craig Shelley <craig@microtron.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: ftdi_sio: Add support for GN Otometrics Aurical USB Audiometer
Ville Sundberg [Tue, 14 Jul 2009 21:27:28 +0000 (00:27 +0300)] 
USB: ftdi_sio: Add support for GN Otometrics Aurical USB Audiometer

The patch adds support for the GN Otometrics Aurical USB Audiometer
(FT232BM-based).
A new VID and a new PID is added.

Signed-off-by: Ville Sundberg <vsundber@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: OMAP: OHCI: hc_driver's stop method should call ohci_stop
Anand Gadiyar [Tue, 14 Jul 2009 11:11:00 +0000 (16:41 +0530)] 
USB: OMAP: OHCI: hc_driver's stop method should call ohci_stop

OMAP: OHCI: hc_driver's stop method should call ohci_stop

Without this, the ohci-omap driver will not cleanup the debugfs
nodes when the driver is unloaded. So the next insmod will fail,
if CONFIG_DEBUG_FS and CONFIG_USB_DEBUG are both selected.

Reported-by: vikram pandita <vikram.pandita@ti.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: storage: raise timeout in usb_stor_Bulk_max_lun
Giacomo Lozito [Mon, 13 Jul 2009 21:23:33 +0000 (23:23 +0200)] 
USB: storage: raise timeout in usb_stor_Bulk_max_lun

Requests to get max LUN, for certain USB storage devices, require a
longer timeout before a correct reply is returned. This happens for a
Realtek USB Card Reader (0bda:0152), which has a max LUN of 3 but is set
to 0, thus losing functionality, because of the timeout occurring too
quickly.

Raising the timeout value fixes the issue and might help other devices
to return a correct max LUN value as well.

Signed-off-by: Giacomo Lozito <james@develia.org>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: let the option driver compile without CONFIG_PM
Oliver Neukum [Mon, 13 Jul 2009 21:24:10 +0000 (23:24 +0200)] 
USB: let the option driver compile without CONFIG_PM

This is needed for compilation without CONFIG_PM.

Signed-off-by: Oliver Neukum <oliver@neukum.org>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option: Remove unused variable
Carlos R. Mafra [Mon, 13 Jul 2009 19:45:03 +0000 (21:45 +0200)] 
USB: option: Remove unused variable

After commit f092c240494f2d807401d93f95f683909b90af96 ("USB: option:
remove unnecessary and erroneous code") the variable 'serial' becomes
unused, as gcc-4.3.2 points out:

drivers/usb/serial/option.c: In function 'option_instat_callback':
drivers/usb/serial/option.c:834: warning: unused variable 'serial'
drivers/usb/serial/option.c: In function 'option_open':
drivers/usb/serial/option.c:930: warning: unused variable 'serial'

So I removed it.

Signed-off-by: Carlos R. Mafra <crmafra@aei.mpg.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoMerge branch 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvar...
Linus Torvalds [Tue, 28 Jul 2009 21:31:02 +0000 (14:31 -0700)] 
Merge branch 'i2c-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  i2c/tsl2550: Fix lux value in dark environment

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable
Linus Torvalds [Tue, 28 Jul 2009 21:27:06 +0000 (14:27 -0700)] 
Merge git://git./linux/kernel/git/mason/btrfs-unstable

* git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable: (22 commits)
  Btrfs: Fix async caching interaction with unmount
  Btrfs: change how we unpin extents
  Btrfs: Correct redundant test in add_inode_ref
  Btrfs: find smallest available device extent during chunk allocation
  Btrfs: clear all space_info->full after removing a block group
  Btrfs: make flushoncommit mount option correctly wait on ordered_extents
  Btrfs: Avoid delayed reference update looping
  Btrfs: Fix ordering of key field checks in btrfs_previous_item
  Btrfs: find_free_dev_extent doesn't handle holes at the start of the device
  Btrfs: Remove code duplication in comp_keys
  Btrfs: async block group caching
  Btrfs: use hybrid extents+bitmap rb tree for free space
  Btrfs: Fix crash on read failures at mount
  Btrfs: remove of redundant btrfs_header_level
  Btrfs: adjust NULL test
  Btrfs: Remove broken sanity check from btrfs_rmap_block()
  Btrfs: convert nested spin_lock_irqsave to spin_lock
  Btrfs: make sure all dirty blocks are written at commit time
  Btrfs: fix locking issue in btrfs_find_next_key
  Btrfs: fix double increment of path->slots[0] in btrfs_next_leaf
  ...

14 years agomISDN: Fix handling of receive buffer size in L1oIP
Andreas Eversberg [Mon, 27 Jul 2009 17:24:04 +0000 (19:24 +0200)] 
mISDN: Fix handling of receive buffer size in L1oIP

The size of receive buffer pointer was used to get size of
receive buffer instead of recvbuf_size itself, so only 4/8
bytes could be transfered.

This is a regression to 2.6.30 introduced by commit
8c90e11e3543d7de612194a042a148caeaab5f1d ("mISDN: Use
kernel_{send,recv}msg instead of open coding")

Signed-off-by: Andreas Eversberg <andreas@eversberg.eu>
Signed-off-by: Karsten Keil <keil@b1-systems.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoeCryptfs: parse_tag_3_packet check tag 3 packet encrypted key size
Ramon de Carvalho Valle [Tue, 28 Jul 2009 18:58:22 +0000 (13:58 -0500)] 
eCryptfs: parse_tag_3_packet check tag 3 packet encrypted key size

The parse_tag_3_packet function does not check if the tag 3 packet contains a
encrypted key size larger than ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES.

Signed-off-by: Ramon de Carvalho Valle <ramon@risesecurity.org>
[tyhicks@linux.vnet.ibm.com: Added printk newline and changed goto to out_free]
Signed-off-by: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
Cc: stable@kernel.org (2.6.27 and 30)
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoeCryptfs: Check Tag 11 literal data buffer size
Tyler Hicks [Tue, 28 Jul 2009 18:57:01 +0000 (13:57 -0500)] 
eCryptfs: Check Tag 11 literal data buffer size

Tag 11 packets are stored in the metadata section of an eCryptfs file to
store the key signature(s) used to encrypt the file encryption key.
After extracting the packet length field to determine the key signature
length, a check is not performed to see if the length would exceed the
key signature buffer size that was passed into parse_tag_11_packet().

Thanks to Ramon de Carvalho Valle for finding this bug using fsfuzzer.

Signed-off-by: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
Cc: stable@kernel.org (2.6.27 and 30)
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agostaging: udlfb: Add vmalloc.h include
Amit Kucheria [Mon, 27 Jul 2009 09:01:03 +0000 (12:01 +0300)] 
staging: udlfb: Add vmalloc.h include

Required for vmalloc_32 and vfree declarations on non-x86 platforms.

Signed-off-by: Amit Kucheria <amit.kucheria@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agostaging: remove aten2011 driver
Greg Kroah-Hartman [Mon, 20 Jul 2009 06:37:32 +0000 (23:37 -0700)] 
staging: remove aten2011 driver

This driver is not needed, as the existing mos7840 driver works
properly for this device.

Thanks to Russell Lang for doing the work to figure this out.

Cc: Russell Lang <gsview@ghostgum.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoStaging: android: lowmemorykiller.c: fix it for "oom: move oom_adj value from task_st...
David Rientjes [Tue, 16 Jun 2009 23:42:53 +0000 (16:42 -0700)] 
Staging: android: lowmemorykiller.c: fix it for "oom: move oom_adj value from task_struct to mm_struct"

I'm about to merge "oom: move oom_adj value from task_struct to
mm_struct", and this fixup is needed to repair linux-next's
drivers/staging/android/lowmemorykiller.c.

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoStaging: serqt_usb2: fix memory leak in error case
Oliver Neukum [Mon, 13 Jul 2009 08:46:57 +0000 (10:46 +0200)] 
Staging: serqt_usb2: fix memory leak in error case

a standard memory leak, as later allocations may fail even if prior
allocations did not. Then the prior allocations must be undone.

Signed-off-by: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoStaging: serqt_usb2: add missing calls to tty_kref_put()
Bill Pemberton [Tue, 14 Jul 2009 19:33:52 +0000 (15:33 -0400)] 
Staging: serqt_usb2: add missing calls to tty_kref_put()

tty_port_tty_get() was called without a corresponding tty_kref_put()
in qt_read_bulk_callback() and qt_close().

Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodriver core: documentation: make it clear that sysfs is optional
Lucian Adrian Grijincu [Mon, 27 Jul 2009 16:06:42 +0000 (09:06 -0700)] 
driver core: documentation: make it clear that sysfs is optional

The original text suggested that sysfs is mandatory and always
compiled in the kernel.

Signed-off-by: Lucian Adrian Grijincu <lgrijincu@ixiacom.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodriver core: sysdev: do not send KOBJ_ADD uevent if kobject_init_and_add fails
Xiaotian Feng [Fri, 24 Jul 2009 09:31:41 +0000 (17:31 +0800)] 
driver core: sysdev: do not send KOBJ_ADD uevent if kobject_init_and_add fails

If kobject_init_and_add fails, sysdev_register should not send KOBJ_ADD
uevent to userspace.

Signed-off-by: Xiaotian Feng <dfeng@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoDynamic debug: fix typo: -/->
Roel Kluin [Wed, 15 Jul 2009 18:29:07 +0000 (20:29 +0200)] 
Dynamic debug: fix typo: -/->

The member was intended, not the local variable.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Greg Banks <gnb@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodriver core: firmware_class:fix memory leak of page pointers array
Ming Lei [Sat, 11 Jul 2009 03:11:10 +0000 (11:11 +0800)] 
driver core: firmware_class:fix memory leak of page pointers array

The page pointers array is allocated in fw_realloc_buffer() called by
firmware_data_write(), and should be freed in release function of firmware
device.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Reported-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosysfs: fix hardlink count on device_move
Peter Oberparleiter [Tue, 14 Jul 2009 15:56:15 +0000 (17:56 +0200)] 
sysfs: fix hardlink count on device_move

Update directory hardlink count when moving kobjects to a new parent.
Fixes the following problem which occurs when several devices are
moved to the same parent and then unregistered:

> ls -laF /sys/devices/css0/defunct/
> total 0
> drwxr-xr-x 4294967295 root root    0 2009-07-14 17:02 ./
> drwxr-xr-x        114 root root    0 2009-07-14 17:02 ../
> drwxr-xr-x          2 root root    0 2009-07-14 17:01 power/
> -rw-r--r--          1 root root 4096 2009-07-14 17:01 uevent

Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoi2c/tsl2550: Fix lux value in dark environment
Michele Jr De Candia [Tue, 28 Jul 2009 14:33:03 +0000 (16:33 +0200)] 
i2c/tsl2550: Fix lux value in dark environment

I've tested TSL2550 driver and I've found a bug: when light is off,
returned value from tsl2550_calculate_lux function is -1 when it should
be 0 (sensor correctly read that light was off).

I think the bug is that a zero c0 value (approximated value of ch0) is
misinterpreted as an error.

Signed-off-by: Michele Jr De Candia <michele.decandia@valueteam.com>
Acked-by: Rodolfo Giometti <giometti@linux.it>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agohwmon: (asus_atk0110) Fix upper limit readings
Luca Tettamanti [Tue, 28 Jul 2009 14:31:39 +0000 (16:31 +0200)] 
hwmon: (asus_atk0110) Fix upper limit readings

On newer Asus boards the "upper" limit of a sensor is encoded as
delta from the "lower" limit. Fix the driver to correctly handle
this case.

Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Tested-by: Alex Macfarlane Smith <nospam@archifishal.co.uk>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agohwmon: (smsc47m1) Differentiate between LPC47M233 and LPC47M292
Jean Delvare [Tue, 28 Jul 2009 14:31:39 +0000 (16:31 +0200)] 
hwmon: (smsc47m1) Differentiate between LPC47M233 and LPC47M292

The SMSC LPC47M233 and LPC47M292 chips have the same device ID but
are not compatible.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Juerg Haefliger <juergh@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
14 years agoBtrfs: Fix async caching interaction with unmount
Yan Zheng [Tue, 28 Jul 2009 12:41:57 +0000 (08:41 -0400)] 
Btrfs: Fix async caching interaction with unmount

- don't stop the caching thread until btrfs_commit_super return.

- if caching is interrupted by umount, set last to (u64)-1.
  otherwise the un-scanned range of block group will be considered
  as free extent.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
14 years agomm: Remove duplicate definitions in MIPS and SH
Benjamin Herrenschmidt [Tue, 28 Jul 2009 00:16:48 +0000 (10:16 +1000)] 
mm: Remove duplicate definitions in MIPS and SH

Those definitions are already provided by asm-generic

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agousb_serial: Fix remaining ref count/lock bugs
Alan Cox [Mon, 27 Jul 2009 23:34:58 +0000 (00:34 +0100)] 
usb_serial: Fix remaining ref count/lock bugs

This fixes
- locking bug that was hidden by ecc2e05e739c30870c8e4f252b63a0c4041f2724
- Regression #13821
- Spurious warning when closing and blocking for data write out

With these changes my PL2303 always ends up as ttyUSB0 when it should and
the module refcounts stay correct.

I'll do a more wholesale split & tidy of _open in the next release or two
as we get a standard tty_port_open and port->ops->init port->ops->shutdown
call backs.

Copy sent to Alan Stern and Carlos Mafra just to confirm it fixes all the
reports but it passes local testing with the same hardware as Alan Stern.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for-linus' of git://git.infradead.org/users/eparis/notify
Linus Torvalds [Mon, 27 Jul 2009 22:54:10 +0000 (15:54 -0700)] 
Merge branch 'for-linus' of git://git.infradead.org/users/eparis/notify

* 'for-linus' of git://git.infradead.org/users/eparis/notify:
  inotify: use GFP_NOFS under potential memory pressure
  fsnotify: fix inotify tail drop check with path entries
  inotify: check filename before dropping repeat events
  fsnotify: use def_bool in kconfig instead of letting the user choose
  inotify: fix error paths in inotify_update_watch
  inotify: do not leak inode marks in inotify_add_watch
  inotify: drop user watch count when a watch is removed

14 years agopty: quickfix for the pty ENXIO timing problems
Alan Cox [Mon, 27 Jul 2009 21:17:51 +0000 (22:17 +0100)] 
pty: quickfix for the pty ENXIO timing problems

This also makes close stall in the normal case which is apparently
needed to fix emacs

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Mon, 27 Jul 2009 20:42:47 +0000 (13:42 -0700)] 
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (45 commits)
  cnic: Fix ISCSI_KEVENT_IF_DOWN message handling.
  net: irda: init spinlock after memcpy
  ixgbe: fix for 82599 errata marking UDP checksum errors
  r8169: WakeOnLan fix for the 8168
  netxen: reset ring consumer during cleanup
  net/bridge: use kobject_put to release kobject in br_add_if error path
  smc91x.h: add config for Nomadik evaluation kit
  NET: ROSE: Don't use static buffer.
  eepro: Read buffer overflow
  tokenring: Read buffer overflow
  at1700: Read buffer overflow
  fealnx: Write outside array bounds
  ixgbe: remove unnecessary call to device_init_wakeup
  ixgbe: Don't priority tag control frames in DCB mode
  ixgbe: Enable FCoE offload when DCB is enabled for 82599
  net: Rework mdio-ofgpio driver to use of_mdio infrastructure
  register at91_ether using platform_driver_probe
  skge: Enable WoL by default if supported
  net: KS8851 needs to depend on MII
  be2net: Bug fix in the non-lro path. Size of received packet was not updated in statistics properly.
  ...

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp
Linus Torvalds [Mon, 27 Jul 2009 19:23:47 +0000 (12:23 -0700)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/bp/bp

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp:
  amd64_edac: read the right F2 maskoffset reg

14 years agotty: Fix a USB serial crash/scribble
Alan Cox [Mon, 27 Jul 2009 09:58:08 +0000 (10:58 +0100)] 
tty: Fix a USB serial crash/scribble

The port lock is used to protect the port state. However the port structure
is freed on a hangup, then the lock taken on a close. The right fix is to
drop the port on tty->shutdown() but we can't yet do that due to sleep v
non-sleeping rules. Instead do the next best thing and fix it up when we are
not in -rc season.

Reported-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Tested-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'fixes-for-linus' of git://git.monstr.eu/linux-2.6-microblaze
Linus Torvalds [Mon, 27 Jul 2009 19:18:27 +0000 (12:18 -0700)] 
Merge branch 'fixes-for-linus' of git://git.monstr.eu/linux-2.6-microblaze

* 'fixes-for-linus' of git://git.monstr.eu/linux-2.6-microblaze:
  microblaze: Makefile cleanup
  microblaze: Typo fix for cpu param inconsistency
  microblaze: Add support for R_MICROBLAZE_64_NONE
  microblaze: Get module loading working
  microblaze: remove sys_ipc
  microblaze: Support unaligned address for put/get_user macros
  microblaze: Detect new Microblaze 7.20 versions
  microblaze: Fix do_page_fault for no context
  microblaze: Add _PAGE_FILE macros to pgtable.h
  microblaze: Fix put_user macro for 64bits arguments
  microblaze: Clear print messages for DTB passing via r7
  microblaze: Not to clear r7 after copying DTB to kernel
  microblaze: Add messages about FDT blob
  microblaze: Final support for statically linked DTB
  microblaze: remove duplicated #include
  microblaze: Define tlb_flush macro

14 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 27 Jul 2009 19:18:09 +0000 (12:18 -0700)] 
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: geode: Mark mfgpt irq IRQF_TIMER to prevent resume failure
  x86, amd: Don't probe for extended APIC ID if APICs are disabled
  x86, mce: Rename incorrect macro name "CONFIG_X86_THRESHOLD"
  x86-64: Fix bad_srat() to clear all state
  x86, mce: Fix set_trigger() accessor
  x86: Fix movq immediate operand constraints in uaccess.h
  x86: Fix movq immediate operand constraints in uaccess_64.h
  x86: Add reboot fixup for SBC-fitPC2
  x86: Include all of .data.* sections in _edata on 64-bit
  x86: Add quirk for Intel DG45ID board to avoid low memory corruption

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Mon, 27 Jul 2009 19:17:29 +0000 (12:17 -0700)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: hda - Fix mute control with some ALC262 models
  ALSA: snd_usb_caiaq: add support for Audio2DJ
  ALSA: pcm - Fix hwptr buffer-size overlap bug
  ALSA: pcm - Fix warnings in debug loggings
  ALSA: pcm - Add logging of hwptr updates and interrupt updates
  ASoC: tlv320aic3x: Enable PLL when not bypassed
  ALSA: hda - Restore GPIO1 properly at resume with AD1984A
  ALSA: ctxfi - Fix uninitialized error checks
  ALSA: hda - Use snprintf() to be safer
  ALSA: usb-audio - Volume control quirk for QuickCam E 3500
  ALSA: pcm - Fix regressions with VMware

14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Mon, 27 Jul 2009 19:16:57 +0000 (12:16 -0700)] 
Merge branch 'for_linus' of git://git./linux/kernel/git/mchehab/linux-2.6

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (34 commits)
  V4L/DVB (12303): cx23885: check pointers before dereferencing in dprintk macro
  V4L/DVB (12302): cx23885-417: fix broken IOCTL handling
  V4L/DVB (12300): bttv: fix regression: tvaudio must be loaded before tuner
  V4L/DVB (12291): b2c2: fix frontends compiled into kernel
  V4L/DVB (12286): sn9c20x: reorder includes to be like other drivers
  V4L/DVB (12284): gspca - jpeg subdrivers: Check the result of kmalloc(jpeg header).
  V4L/DVB (12283): gspca - sn9c20x: New subdriver for sn9c201 and sn9c202 bridges.
  V4L/DVB (12282): gspca - main: Support for vidioc_g_chip_ident and vidioc_g/s_register.
  V4L/DVB (12269): af9013: auto-detect parameters in case of garbage given by app
  V4L/DVB (12267): gspca - sonixj: Bad sensor init of non ov76xx sensors.
  V4L/DVB (12265): em28xx: fix tuning problem in HVR-900 (R1)
  V4L/DVB (12263): em28xx: set demod profile for Pinnacle Hybrid Pro 320e
  V4L/DVB (12262): em28xx: Make sure the tuner is initialized if generic empia USB id was used
  V4L/DVB (12261): em28xx: set GPIO properly for Pinnacle Hybrid Pro analog support
  V4L/DVB (12260): em28xx: make support work for the Pinnacle Hybrid Pro (eb1a:2881)
  V4L/DVB (12258): em28xx: fix typo in mt352 init sequence for Terratec Cinergy T XS USB
  V4L/DVB (12257): em28xx: make tuning work for Terratec Cinergy T XS USB (mt352 variant)
  V4L/DVB (12245): em28xx: add support for mt9m001 webcams
  V4L/DVB (12244): em28xx: adjust vinmode/vinctl based on the stream input format
  V4L/DVB (12243): em28xx: allow specifying sensor xtal frequency
  ...

14 years agoMerge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Mon, 27 Jul 2009 19:16:38 +0000 (12:16 -0700)] 
Merge branch 'for-linus' of git://git390.marist.edu/linux-2.6

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6:
  [S390] zcrypt: fix scheduling of hrtimer ap_poll_timer
  [S390] vdso: clock_gettime of CLOCK_THREAD_CPUTIME_ID with noexec=on
  [S390] vdso: fix per cpu area allocation
  [S390] hibernation: fix register corruption on machine checks
  [S390] hibernation: fix lowcore handling

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm
Linus Torvalds [Mon, 27 Jul 2009 19:16:21 +0000 (12:16 -0700)] 
Merge git://git./linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm:
  dm table: pass correct dev area size to device_area_is_valid
  dm: remove queue next_ordered workaround for barriers
  dm raid1: wake kmirrord when requeueing delayed bios after remote recovery

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/shaggy...
Linus Torvalds [Mon, 27 Jul 2009 19:15:56 +0000 (12:15 -0700)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/shaggy/jfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/shaggy/jfs-2.6:
  jfs: Fix early release of acl in jfs_get_acl

14 years agoupdate the comment in kthread_stop()
Oleg Nesterov [Fri, 19 Jun 2009 00:51:13 +0000 (02:51 +0200)] 
update the comment in kthread_stop()

Commit 63706172f332fd3f6e7458ebfb35fa6de9c21dc5 ("kthreads: rework
kthread_stop()") removed the limitation that the thread function mysr
not call do_exit() itself, but forgot to update the comment.

Since that commit it is OK to use kthread_stop() even if kthread can
exit itself.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agomodule: use MODULE_SYMBOL_PREFIX with module_layout
Mike Frysinger [Thu, 23 Jul 2009 14:12:08 +0000 (23:42 +0930)] 
module: use MODULE_SYMBOL_PREFIX with module_layout

The check_modstruct_version() needs to look up the symbol "module_layout"
in the kernel, but it does so literally and not by a C identifier.  The
trouble is that it does not include a symbol prefix for those ports that
need it (like the Blackfin and H8300 port).  So make sure we tack on the
MODULE_SYMBOL_PREFIX define to the front of it.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux...
Linus Torvalds [Mon, 27 Jul 2009 19:12:10 +0000 (12:12 -0700)] 
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs-2.6

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6:
  jbd: fix race between write_metadata_buffer and get_write_access
  ext3: Get rid of extenddisksize parameter of ext3_get_blocks_handle()
  jbd: Fix a race between checkpointing code and journal_get_write_access()
  ext3: Fix truncation of symlinks after failed write
  jbd: Fail to load a journal if it is too short

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Mon, 27 Jul 2009 19:11:43 +0000 (12:11 -0700)] 
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  [CIFS] fix sparse warning
  cifs: fix sb->s_maxbytes so that it casts properly to a signed value
  cifs: disable serverino if server doesn't support it

14 years agomm: Pass virtual address to [__]p{te,ud,md}_free_tlb()
Benjamin Herrenschmidt [Wed, 22 Jul 2009 05:44:28 +0000 (15:44 +1000)] 
mm: Pass virtual address to [__]p{te,ud,md}_free_tlb()

mm: Pass virtual address to [__]p{te,ud,md}_free_tlb()

Upcoming paches to support the new 64-bit "BookE" powerpc architecture
will need to have the virtual address corresponding to PTE page when
freeing it, due to the way the HW table walker works.

Basically, the TLB can be loaded with "large" pages that cover the whole
virtual space (well, sort-of, half of it actually) represented by a PTE
page, and which contain an "indirect" bit indicating that this TLB entry
RPN points to an array of PTEs from which the TLB can then create direct
entries. Thus, in order to invalidate those when PTE pages are deleted,
we need the virtual address to pass to tlbilx or tlbivax instructions.

The old trick of sticking it somewhere in the PTE page struct page sucks
too much, the address is almost readily available in all call sites and
almost everybody implemets these as macros, so we may as well add the
argument everywhere. I added it to the pmd and pud variants for consistency.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Howells <dhowells@redhat.com> [MN10300 & FRV]
Acked-by: Nick Piggin <npiggin@suse.de>
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> [s390]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agocnic: Fix ISCSI_KEVENT_IF_DOWN message handling.
Michael Chan [Mon, 27 Jul 2009 18:25:58 +0000 (11:25 -0700)] 
cnic: Fix ISCSI_KEVENT_IF_DOWN message handling.

When a net device goes down or when the bnx2i driver is unloaded,
the code was not generating the ISCSI_KEVENT_IF_DOWN message
properly and this could cause the userspace driver to crash.

This is fixed by sending the message properly in the shutdown path.
cnic_uio_stop() is also added to send the message when bnx2i is
unregistering.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoBtrfs: change how we unpin extents
Josef Bacik [Mon, 27 Jul 2009 17:57:01 +0000 (13:57 -0400)] 
Btrfs: change how we unpin extents

We are racy with async block caching and unpinning extents.  This patch makes
things much less complicated by only unpinning the extent if the block group is
cached.  We check the block_group->cached var under the block_group->lock spin
lock.  If it is set to BTRFS_CACHE_FINISHED then we update the pinned counters,
and unpin the extent and add the free space back.  If it is not set to this, we
start the caching of the block group so the next time we unpin extents we can
unpin the extent.  This keeps us from racing with the async caching threads,
lets us kill the fs wide async thread counter, and keeps us from having to set
DELALLOC bits for every extent we hit if there are caching kthreads going.

One thing that needed to be changed was btrfs_free_super_mirror_extents.  Now
instead of just looking for LOCKED extents, we also look for DIRTY extents,
since we could have left some extents pinned in the previous transaction that
will never get freed now that we are unmounting, which would cause us to leak
memory.  So btrfs_free_super_mirror_extents has been changed to
btrfs_free_pinned_extents, and it will clear the extents locked for the super
mirror, and any remaining pinned extents that may be present.  Thank you,

Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>