Currently Nouveau will unvalidate all buffers if it is forced to wait on
one, and then start revalidating from the beginning. While doing so, it
destroys the operation fence, causing nouveau_fence_emit to crash.
This patch fixes this bug by taking the fence object out of validate_op
and creating it just before emit. The fence pointer is initialized to 0
and unref'ed unconditionally.
In addition to fixing the bug, this prevents its reintroduction and
simplifies the code.
Signed-off-by: Luca Barbieri <luca@luca-barbieri.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The below is mainly an educated guess at what's going on, docs would
sure be handy... NVIDIA? :P
It appears it's possible for a ctxprog to run even while a GPU exception
is pending. The GF8 and up ctxprogs appear to have a small snippet of
code which detects this, and stalls the ctxprog until it's been handled,
which essentially looks like:
if (r2 & 0x00008000) {
r0 |= 0x80000000;
while (r0 & 0x80000000) {}
}
I don't know of any way that flag would get cleared unless the driver
intervenes (and indeed, in the cases I've seen the hang, nothing steps
in to automagically clear it for us). This patch causes the driver to
clear the flag during the PGRAPH IRQ handler.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
There's no good reason for us to have our own anymore, this is left over
from an early port to these TTM interfaces.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
It's mostly a cleanup, but in nv50_fbcon_accel_init gpu lockup
message was printed, but HWACCEL_DISBALED flag was not set.
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Depending on the visual, the colours handed to us in fillrect() can either be
an actual colour, or an index into the pseudo-palette.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This should avoid a race condition on nv0x, if we're doing it with
actual PGRAPH objects and a there's a fence within the FIFO DMA fetch
area when a context switch kicks in.
In that case we get an ILLEGAL_MTHD interrupt as expected, but the
values in PGRAPH_TRAPPED_ADDR aren't calculated correctly and they're
almost useless (e.g. you can see ILLEGAL_MTHDs for the now inactive
channel, with a wrong offset/data pair).
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
It will be useful for various synchronization purposes, mostly stolen
from "[PATCH] drm/nv50: synchronize user channel after buffer object
move on kernel channel" by Maarten Maathuis.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
- Aligning to block size should ensure that the extra size is enough.
- Using roundup, because not all sizes are powers of two.
Signed-off-by: Maarten Maathuis <madman2003@gmail.com>
struct fb_fillrect->color is not a color, but index into pseudo_palette
array
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This partially reverts e4b41066, as this driver is intended to be
useful with any KMS driver for suitable hardware. The missing build
dependency that commit workarounded was DRM_KMS_HELPER.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
This commit has also the following 3 bugfix commits squashed into it from
the nouveau git tree:
drm/nouveau: Fix up the tiling alignment restrictions for nv1x.
drm/nouveau: Fix up the nv2x tiling alignment restrictions.
drm/nv50: fix align typo for g9x
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Lots of ppl keep thinking this is an oops, it was just a warning for
me to see, just make it a printk now.
Signed-off-by: Dave Airlie <airlied@redhat.com>
If userspace (plymouth in this case) asks for a deeper depth,
refuse it as well due to lack of resizing.
This fixes an issue since < 32MB cards went to 8bpp and plymouth
crashes on startup.
Signed-off-by: Dave Airlie <airlied@redhat.com>
With the current DRM code, an output that has been powered off
from userspace will automatically power back on when resuming
from suspend. This patch fixes this behaviour.
Tested only with the Intel i915 driver on an Intel GM45 Express
chipset.
Signed-off-by: David John <davidjon@xenontk.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
pci_dma_mapping_error should be used to test return value of
pci_map_single or pci_map_page.
Signed-off-by: Denis Kirjanov <kirjanov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The id_table field of the struct pci_driver is constant in <linux/pci.h>
so it is worth to make pci_device_id also constant.
The semantic match that finds this kind of pattern is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@r@
identifier I1, I2, x;
@@
struct I1 {
...
const struct I2 *x;
...
};
@s@
identifier r.I1, y;
identifier r.x, E;
@@
struct I1 y = {
.x = E,
};
@c@
identifier r.I2;
identifier s.E;
@@
const struct I2 E[] = ... ;
@depends on !c@
identifier r.I2;
identifier s.E;
@@
+ const
struct I2 E[] = ...;
// </smpl>
Signed-off-by: Márton Németh <nm127@freemail.hu>
Cc: Julia Lawall <julia@diku.dk>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
The max junction temperature of Atom N450/D410/D510 CPUs is 100 degrees
Celsius. Since these CPUs are always coupled with Intel NM10 chipset in
one package, the best way to verify whether an Atom CPU is N450/D410/D510
is to check the host bridge device.
Signed-off-by: Yong Wang <yong.y.wang@intel.com>
Acked-by: Huaxu Wan <huaxu.wan@intel.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
The latest version of the Revision Guide for AMD Family 10h Processors
lists two more processor revisions which may be affected by erratum 319.
Change the blacklisting code to correctly detect those processors, by
implementing AMD's recommended algorithm.
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Andreas Herrmann <herrmann.der.user@googlemail.com>
Expose the raw GGRP/GITM interface via debugfs. The hwmon interface is
reverse engineered and the driver tends to break on newer boards...
Using this interface it's possible to poke directly at the ACPI methods
without the need to recompile, reducing the guesswork and the round trips
needed to support a new revision of the interface.
Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
The behaviour is unmodified, this makes easier to override the heuristic (which
is probably needed for some boards).
Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
The ADT7462_PIN28_VOLT value is a 4-bit field, so the corresponding
shift must be 4.
Signed-off-by: Roger Blofeld <blofeldus@yahoo.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
When we search for and find a busy extent during allocation we
force the log out to ensure the extent free transaction is on
disk before the allocation transaction. The current implementation
has a subtle bug in it--it does not handle multiple overlapping
ranges.
That is, if we free lots of little extents into a single
contiguous extent, then allocate the contiguous extent, the busy
search code stops searching at the first extent it finds that
overlaps the allocated range. It then uses the commit LSN of the
transaction to force the log out to.
Unfortunately, the other busy ranges might have more recent
commit LSNs than the first busy extent that is found, and this
results in xfs_alloc_search_busy() returning before all the
extent free transactions are on disk for the range being
allocated. This can lead to potential metadata corruption or
stale data exposure after a crash because log replay won't replay
all the extent free transactions that cover the allocation range.
Modified-by: Alex Elder <aelder@sgi.com>
(Dropped the "found" argument from the xfs_alloc_busysearch trace
event.)
Signed-off-by: Dave Chinner <david@fromorbit.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
Because inodes remain in cache much longer than inode buffers do
under memory pressure, we can get the situation where we have
stale, dirty inodes being reclaimed but the backing storage has
been freed. Hence we should never, ever flush XFS_ISTALE inodes
to disk as there is no guarantee that the backing buffer is in
cache and still marked stale when the flush occurs.
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
We currently have some rather odd code in xfs_setattr for
updating the a/c/mtime timestamps:
- first we do a non-transaction update if all three are updated
together
- second we implicitly update the ctime for various changes
instead of relying on the ATTR_CTIME flag
- third we set the timestamps to the current time instead of the
arguments in the iattr structure in many cases.
This patch makes sure we update it in a consistent way:
- always transactional
- ctime is only updated if ATTR_CTIME is set or we do a size
update, which is a special case
- always to the times passed in from the caller instead of the
current time
The only non-size caller of xfs_setattr that doesn't come from
the VFS is updated to set ATTR_CTIME and pass in a valid ctime
value.
Reported-by: Eric Blake <ebb9@byu.net>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
Using DECLARE_EVENT_CLASS allows us to to use trace event code
instead of duplicating it in the binary. This was not available
before 2.6.33 so it had to be done as a separate step once the
prerequisite was merged.
This only requires changes to xfs_trace.h and the results are
rather impressive:
hch@brick:~/work/linux-2.6/obj-kvm$ size fs/xfs/xfs.o*
text data bss dec hex filename
607732 41884 3616 653232 9f7b0 fs/xfs/xfs.o
1026732 41884 3808 1072424 105d28 fs/xfs/xfs.o.old
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
Add the STMicroelectronics ST7597 codec and an unknown codec
from the same manufacturer found on the Creative SB 128 card (CT4810).
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
This model needs both 'Headphone Jack Sense' and 'Line Jack Sense' muted
for audible playback, so just add it to the ad1981 jack sense blacklist.
Cc: stable@kernel.org
Tested-by: Pete <x41215201@gmail.com>
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
The file arch/arm/include/asm/cpu.h needs to include 'linux/cpu.h' to
meet its dependency. Otherwise when using "struct cpuinfo_arm" and
including just 'asm/cpu.h' throws below error -
arch/arm/include/asm/cpu.h:16: error: field 'cpu' has incomplete type
To fix this otherway, one can also include both linux/cpu.h and
asm/cpu.h but it shoudn't be that way. So this patch fixes this by
including the linux/cpu.h in asm/cpu.h, so that including alone
asm/cpu.h is enough.
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch fixes a build failure [1] due to missing includes
This should make the arm tree build again with lpd7a404_defconfig
References:
http://kisskb.ellerman.id.au/kisskb/buildresult/1983329/
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
A lot of ARM-defconfigs (those without CONFIG_ISA_DMA_API set) fail to
build [1][2][3] due to the changes of the patch
[PATCH] PCI: Clean up build for CONFIG_PCI_QUIRKS unset
by Rafael J. Wysocki (Sat, 2 Jan 2010 22:57:24 +0100) [4]
as the referenced variable 'isa_dma_bridge_buggy' in asm/dma.h is
enclosed by the CONFIG_ISA_DMA_API conditional all configs without this
setting fail to build.
I'm not sure wether moving the condition is the right way to solve the
issue, but atleast it fixes the issue :)
References:
[1] http://kisskb.ellerman.id.au/kisskb/buildresult/1983354/
[2] http://kisskb.ellerman.id.au/kisskb/buildresult/1983333/
[3] http://kisskb.ellerman.id.au/kisskb/buildresult/1983337/
[4] http://lkml.org/lkml/2010/1/2/102
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Commit cbda12d77e (drm/i915: implement
new pm ops for i915), among other things, removed the .suspend and
.resume pointers from the struct drm_driver object in i915_drv.c,
which broke resume without KMS on my MSI Wind U100.
Fix this by reverting that part of commit cbda12d77e.
[ The DRM layer will not use the class-specific suspend/resume functions
if the driver is marked MODESET-aware, and conversely it will not
register the PCI device if the drievr isn't so marked, so you always
end up with _either_ the drm-class suspend/resume _or_ the PCI layer
PM functionality, never both. - Linus ]
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dynamically allocate the CPUFreq frequency table on OMAP2xxx chips.
This fixes some compilation problems, since the kernel may not know
what chip it is running on until boot-time. This also reduces the size
of the CPUFreq frequency table.
Problem originally reported by Felipe Balbi <felipe.balbi@nokia.com>.
Thanks also for comments on the patch from Felipe and Kevin.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Cc: Kevin Hilman <khilman@deeprootsystems.com>
A subsequent patch adds code on OMAP2xxx to dynamically allocate the
CPUFreq frequency table in clk_init_cpufreq_table(), so for it to
avoid a leak, it will need a corresponding function to free the
memory. This patch adds clk_exit_cpufreq_table() with generic
code to call a chip-specific variant inside the clockfw_lock spinlock via
struct clk_functions.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Revise some of the comments in the OMAP2xxx OPP data for clarity.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Richard Woodruff <r-woodruff2@ti.com>
