linux.git, branch v3.14.44

Linux 3.14.44

2015-06-06T15:19:54+00:00

fs/binfmt_elf.c:load_elf_binary(): return -EINVAL on zero-length mappings

2015-06-06T15:19:40+00:00

commit 2b1d3ae940acd11be44c6eced5873d47c2e00ffa upstream.

load_elf_binary() returns `retval', not `error'.

Fixes: a87938b2e246b81b4fb ("fs/binfmt_elf.c: fix bug in loading of PIE binaries")
Reported-by: James Hogan 
Cc: Michael Davidson 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

vfs: read file_handle only once in handle_to_path

2015-06-06T15:19:40+00:00

commit 161f873b89136eb1e69477c847d5a5033239d9ba upstream.

We used to read file_handle twice.  Once to get the amount of extra
bytes, and once to fetch the entire structure.

This may be problematic since we do size verifications only after the
first read, so if the number of extra bytes changes in userspace between
the first and second calls, we'll have an incoherent view of
file_handle.

Instead, read the constant size once, and copy that over to the final
structure without having to re-read it again.

Signed-off-by: Sasha Levin 
Cc: Al Viro 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

drm/radeon: partially revert "fix VM_CONTEXT*_PAGE_TABLE_END_ADDR handling"

2015-06-06T15:19:40+00:00

commit 7c0411d2fabc2e2702c9871ffb603e251158b317 upstream.

We have that bug for years and some users report side effects when fixing it on older hardware.

So revert it for VM_CONTEXT0_PAGE_TABLE_END_ADDR, but keep it for VM 1-15.

Signed-off-by: Christian König 
Signed-off-by: Alex Deucher 
Signed-off-by: Greg Kroah-Hartman

drm/radeon: fix VM_CONTEXT*_PAGE_TABLE_END_ADDR handling

2015-06-06T15:19:40+00:00

commit 607d48063512707a414e346972e2210dc71ab491 upstream.

The mapping range is inclusive between starting and ending addresses.

Signed-off-by: Christian König 
Signed-off-by: Alex Deucher 
Signed-off-by: Greg Kroah-Hartman

drm/radeon: add new bonaire pci id

2015-06-06T15:19:39+00:00

commit fcf3b54282e4c5a95a1f45f67558bc105acdbc6a upstream.

Signed-off-by: Alex Deucher 
Signed-off-by: Greg Kroah-Hartman

ACPI / init: Fix the ordering of acpi_reserve_resources()

2015-06-06T15:19:39+00:00

commit b9a5e5e18fbf223502c0b2264c15024e393da928 upstream.

Since acpi_reserve_resources() is defined as a device_initcall(),
there's no guarantee that it will be executed in the right order
with respect to the rest of the ACPI initialization code.  On some
systems this leads to breakage if, for example, the address range
that should be reserved for the ACPI fixed registers is given to
the PCI host bridge instead if the race is won by the wrong code
path.

Fix this by turning acpi_reserve_resources() into a void function
and calling it directly from within the ACPI initialization sequence.

Reported-and-tested-by: George McCollister 
Link: http://marc.info/?t=143092384600002&r=1&w=2
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

sd: Disable support for 256 byte/sector disks

2015-06-06T15:19:39+00:00

commit 74856fbf441929918c49ff262ace9835048e4e6a upstream.

256 bytes per sector support has been broken since 2.6.X,
and no-one stepped up to fix this.
So disable support for it.

Signed-off-by: Mark Hounschell 
Signed-off-by: Hannes Reinecke 
Signed-off-by: James Bottomley 
Signed-off-by: Greg Kroah-Hartman

storvsc: Set the SRB flags correctly when no data transfer is needed

2015-06-06T15:19:39+00:00

commit dc45708ca9988656d706940df5fd102672c5de92 upstream.

Set the SRB flags correctly when there is no data transfer.  Without this
change some IHV drivers will fail valid commands such as TEST_UNIT_READY.

Signed-off-by: K. Y. Srinivasan 
Reviewed-by: Long Li 
Signed-off-by: James Bottomley 
Signed-off-by: Greg Kroah-Hartman

Input: elantech - fix semi-mt protocol for v3 HW

2015-06-06T15:19:39+00:00

commit 3c0213d17a09601e0c6c0ae0e27caf70d988290f upstream.

When the v3 hardware sees more than one finger, it uses the semi-mt
protocol to report the touches. However, it currently works when
num_fingers is 0, 1 or 2, but when it is 3 and above, it sends only 1
finger as if num_fingers was 1.

This confuses userspace which knows how to deal with extra fingers
when all the slots are used, but not when some are missing.

Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=90101

Signed-off-by: Benjamin Tissoires 
Signed-off-by: Dmitry Torokhov 
Signed-off-by: Greg Kroah-Hartman