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>
Commit cbda12d77ea590082edb6d30bd342a67ebc459e0 (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 cbda12d77ea59.
[ 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>
if we enable CPUFREQ we can't build omap2 for two reasons,
one of them is fixed by the patch below.
It's failing because the __must_be_array() check in
ARRAY_SIZE() is failing and printing the following message:
arch/arm/mach-omap2/clock2xxx.c:453: error: negative width in bit-field '<anonymous>'
Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
[paul@pwsan.com: commit message updated; changed rate variable name]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Commit 52650505fbf3a6ab851c801f54e73e76c55ab8da added an __initdata
decoration to the structure containing the clk_enable and clk_disable
functions. Once init data was freed, these pointers went to null, and
the next enable or disable call caused the kernel to crash. This
change removes this decoration.
Signed-off-by: Cory Maccarrone <darkstar6262@gmail.com>
[paul@pwsan.com: patch manually split and commit message edited]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
This change adds in some missing clocks that were needed as a result
of 526505... (OMAP1 clock: convert mach-omap1/clock.h to
mach-omap1/clock_data.c). Prior to this, it was just assumed that
these clocks existed for all devices, and it was used directly instead
of calling it out with a clock_get call or similar. So, not having
the CK_7XX meant these clocks weren't being used anymore for omap 7xx
devices, which broke things badly.
Signed-off-by: Cory Maccarrone <darkstar6262@gmail.com>
[paul@pwsan.com: commit message edited]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
The only symbols that should be exported are symbols that are to be
called from loadable kernel modules, e.g., device drivers. In the
context of plat-omap/clock.c, these should only be the Linux clock
interface symbols as defined by include/linux/clk.h. Core code
doesn't need these symbols to be exported. Also, clean up an old
comment while here.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Add necessary definitions to clock framework to allow changing
dpll4_m5_ck rate. This is used by the camera code.
Signed-off-by: Jouni Högander <jouni.hogander@nokia.com>
Signed-off-by: Tuukka Toivonen <tuukka.o.toivonen@nokia.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
The correct parent of the McBSP 2, 3, and 4 functional clocks is
PER_96M_FCLK, not CORE_96M_FCLK. Fix this in the OMAP clock tree.
Reported by Nicole Chalhoub <n-chalhoub@ti.com>.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Nicole Chalhoub <n-chalhoub@ti.com>
UART1 & 2 were missing clockdomains resulting in broken omap_hwmod
init for these devices.
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Commit 10db25fea4c11661070b97832b8cc3d2af495092 causes the following
kernel messages during N800 boot (and presumably all other 2420
boards):
[ 0.000000] BUG: mapping for 0x58000000 at 0xe0000000 overlaps vmalloc space
[ 0.000000] BUG: mapping for 0x59000000 at 0xe1000000 overlaps vmalloc space
[ 0.000000] BUG: mapping for 0x5a000000 at 0xe2000000 overlaps vmalloc space
Fix by remapping the IVA memory areas somewhere outside vmalloc space.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
Out of the three major OMAP2 chip types, OMAP2420, OMAP2430, and OMAP3430,
we only map the IVA on OMAP2420. The memory mapping is not shared between
OMAP2420 and OMAP2430, so it is inappropriate to label those macros as
'24XX'; this patch changes them to '2420'.
Signed-off-by: Paul Walmsley <paul@pwsan.com>
In OMAP2/3 some of the clock-domains which did not have control
facility were being falsely written to and read using the CM_CLKSTCTRL
register though it did not exist for them. One check is added to remove
this flaw.
Signed-off-by: Abhijit Pagare <abhijitpagare@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/linux-2.6-kgdb:
kgdb: Fix kernel-doc format error in kgdb.h
blackfin,kgdb: Do not put PC in gdb_regs into retx.
blackfin,kgdb,probe_kernel: Cleanup probe_kernel_read/write
maccess,probe_kernel: Allow arch specific override probe_kernel_(read|write)
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-tip:
x86, irq: Check move_in_progress before freeing the vector mapping
x86: copy_from_user() should not return -EFAULT
Revert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"
x86/pci: Intel ioh bus num reg accessing fix
x86: Fix size for ex trampoline with 32bit
* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
nfs: fix oops in nfs_rename()
sunrpc: fix build-time warning
sunrpc: on successful gss error pipe write, don't return error
SUNRPC: Fix the return value in gss_import_sec_context()
SUNRPC: Fix up an error return value in gss_import_sec_context_kerberos()
Randy Dunlap Reported printk() format-related warnings reported
on i386 builds in his environment. Dave Chinner provided this
patch to eliminate them.
Signed-off by: Dave Chinner <david@fromorbit.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Without this the kernel doesn't boot, it craches in
omap_mux_package_fixup(), since the package_subset becomes NULL.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Otherwise bringing up new boards can be harder:
Unable to handle kernel NULL pointer dereference at virtual address 00000000
pgd = c0004000
[00000000] *pgd=00000000
Internal error: Oops: 5 [#1]
last sysfs file:
Modules linked in:
CPU: 0 Not tainted (2.6.33-rc2-00015-g0bc9c93-dirty #37)
PC is at omap_mux_init+0xa4/0x3d8
LR is at omap_mux_init+0x3c/0x3d8
...
Signed-off-by: Tony Lindgren <tony@atomide.com>
`!' has a higher precedence than `&' so parentheses are required.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Keys: 'right arrow', 'up arrow' and 'select' were mapped
wrongly. This patch corrects them.
This patch also adds one missing key present in the board,
currently I added it as 'unknown' key, as I am not able to
find proper description for this key.
One key entry (r: 7, c: 5) is present in the keymap, which
is really not present in the board, removing it.
Signed-off-by: Vimal Singh <vimalsingh@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Commit f62349ee9788b1d94c55eb6c291d74a1f69bdd9e makes it possible to
have some other than first uart port as ttyS0, which breaks the workaround
serial_in_override() function which will try to address the first uart
port (for ttyS0) and not the one that was initialized.
Signed-off-by: Alexander Shishkin <virtuoso@slind.org>
CC: Mika Westerberg <ext-mika.1.westerberg@nokia.com>
Acked-by: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Commit 9905a43b made struct backlight_ops const. Omap was
setting check_fb dynamically, which caused the following
compile error:
drivers/video/backlight/omap1_bl.c: In function 'omapbl_probe':
drivers/video/backlight/omap1_bl.c:142: error: assignment of read-only variable 'omapbl_ops'
Turns out pdata->check_fb is not being used, so just remove
it to fix the compile.
Cc: Emese Revfy <re.emese@gmail.com>
Cc: Richard Purdie <rpurdie@linux.intel.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Commit 35c9049b27040d09461bc90928ad770be7ddf661 added
drivers/spi/omap_spi_100k.c.
This patch add the related clocks and pin muxing
entries to make the driver work on omap7xx platforms.
Signed-off-by: Cory Maccarrone <darkstar6262@gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Some chips, namely any OMAP1 chips using METHOD_MPUIO,
OMAP15xx and OMAP7xx, cannot be setup to respond to on-chip GPIO
interrupts in both rising and falling edge directions -- they can
only respond to one direction or the other, depending on how the
ICR is configured.
Additionally, current code forces rising edge detection if both
flags are specified:
if (trigger & IRQ_TYPE_EDGE_RISING)
l |= 1 << gpio;
else if (trigger & IRQ_TYPE_EDGE_FALLING)
l &= ~(1 << gpio);
else
goto bad;
This change implements a toggle function that will modify the ICR
to flip the direction of interrupt for IRQs that are requested with
both rising and falling flags. The toggle function is not called
for chips and GPIOs it does not apply to through the use of a flip_mask
that's added on a per-bank basis. The mask is only set for those
GPIOs where a toggle is necessary. Edge detection starts out the
same as above with FALLING mode first.
The toggle happens on EACH interrupt; without it, we have the
following sequence of actions on GPIO transition:
ICR GPIO Result
0x1 0 -> 1 (rising) Interrupt
0x1 1 -> 0 (falling) No interrupt
(set ICR to 0x0 manually)
0x0 0 -> 1 (rising) No interrupt
0x0 1 -> 0 (falling) Interrupt
That is, with the ICR set to 1 for a gpio, only rising edge interrupts
are caught, and with it set to 0, only falling edge interrupts are
caught. If we add in the toggle, we get this:
ICR GPIO Result
0x1 0 -> 1 (rising) Interrupt (ICR set to 0x0)
0x0 1 -> 0 (falling) Interrupt (ICR set to 0x1)
0x1 0 -> 1 ...
so, both rising and falling are caught, per the request for both
(IRQ_TYPE_EDGE_RISING | IRQ_TYPE_EDGE_FALLING).
Signed-off-by: Cory Maccarrone <darkstar6262@gmail.com>
Acked-by: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
In its current form, the omap_mcbsp_request() function can return after
irq_request() failure without any cleanups, effectively locking out the port
forever with clocks left running. Fix it.
Signed-off-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Acked-by: Jarkko Nikula <jhnikula@gmail.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@nokia.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
[IA64] __per_cpu_idtrs[] is a memory hog
[IA64] sanity in #include files. Move fnptr to types.h
[IA64] use helpers for rlimits
[IA64] cpumask_of_node() should handle -1 as a node
The ref counting for the bh returned by gfs2_ea_find() was
wrong. This patch ensures that we always drop the ref count
to that bh correctly.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The rename code was taking a resource group lock in cases where
it wasn't actually needed, this caused problems if the rename
was resulting in an inode being unlinked. The patch ensures that
we only take the rgrp lock early if it is really needed.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The VFS reads the inode size during generic_file_aio_write() but
with no locking around it. In order to get the expected result
from O_APPEND opens, this patch updated the inode size before
calling generic_file_aio_write()
There is of course still a race here, in that there is nothing to
prevent another node coming in and extending the file in the
mean time. On the other hand, when used with file locking this
will ensure that the expected results are obtained.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
If for any reason we haven't installed handler we shouldn't try to
enable IRQ/MSI on the hw so we don't get unhandled IRQ/MSI which
makes the kernel sad.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
In some case vblank might not be initialized and we shouldn't
try to use associated function. This patch make sure this is
the case. It also export drm_vblank_cleanup so driver can cleanup
vblank if for any reason IRQ/MSI is not working.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
