Originally from Nick Piggin, just adapted to the newer branch.
You can't check PageLRU without holding zone->lru_lock. The page
release code can get away with it only because the page refcount is 0 at
that point. Also, you can't reliably remove pages from the LRU unless
the refcount is 0. Ever.
Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Jens Axboe <axboe@suse.de>
[PATCH] splice: page stealing needs to wait_on_page_writeback()
Thanks to Andrew for the good explanation of why this is so. akpm writes:
If a page is under writeback and we remove it from pagecache, it's still
going to get written to disk. But the VFS no longer knows about that page,
nor that this page is about to modify disk blocks.
So there might be scenarios in which those
blocks-which-are-about-to-be-written-to get reused for something else.
When writeback completes, it'll scribble on those blocks.
This won't happen in ext2/ext3-style filesystems in normal mode because the
page has buffers and try_to_release_page() will fail.
But ext2 in nobh mode doesn't attach buffers at all - it just sticks the
page in a BIO, finds some new blocks, points the BIO at those blocks and
lets it rip.
While that write IO's in flight, someone could truncate the file. Truncate
won't block on the writeout because the page isn't in pagecache any more.
So truncate will the free the blocks from the file under the page's feet.
Then something else can reallocate those blocks. Then write data to them.
Now, the original write completes, corrupting the filesystem.
[PATCH] splice: improve writeback and clean up page stealing
By cleaning up the writeback logic (killing write_one_page() and the manual
set_page_dirty()), we can get rid of ->stolen inside the pipe_buffer and
just keep it local in pipe_to_file().
This also adds dirty page balancing logic and O_SYNC handling.
Merge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
IB/ipath: kbuild infrastructure
IB/ipath: infiniband verbs support
IB/ipath: misc infiniband code, part 2
IB/ipath: misc infiniband code, part 1
IB/ipath: infiniband RC protocol support
IB/ipath: infiniband UC and UD protocol support
IB/ipath: infiniband header files
IB/ipath: layering interfaces used by higher-level driver code
IB/ipath: support for userspace apps using core driver
IB/ipath: sysfs and ipathfs support for core driver
IB/ipath: misc driver support code
IB/ipath: chip initialisation code, and diag support
IB/ipath: support for PCI Express devices
IB/ipath: support for HyperTransport devices
IB/ipath: core driver header files
IB/ipath: core device driver
* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
[SPARC]: Wire up sys_sync_file_range() into syscall tables.
[SPARC]: Wire up sys_splice() into the syscall tables.
[SPARC64]: Update defconfig.
[SPARC64]: Align address in huge_pte_alloc().
[SPARC64]: Document the instruction checks we do in do_sparc64_fault().
[SPARC64]: Make tsb_sync() mm comparison more precise.
It doesn't make the splice itself necessarily nonblocking (because the
actual file descriptors that are spliced from/to may block unless they
have the O_NONBLOCK flag set), but it makes the splice pipe operations
nonblocking.
Randy Dunlap [Wed, 29 Mar 2006 19:15:23 +0000 (16:15 -0300)]
V4L/DVB (3667b): cpia2: fix function prototype
Fix address space warning (from sparse):
drivers/media/video/cpia2/cpia2_core.c:2355:6: error: symbol 'cpia2_read' redeclared with different type (originally declared at drivers/media/video/cpia2/cpia2.h:458) - incompatible argument 2 (different address spaces)
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Hans Verkuil [Sat, 1 Apr 2006 21:03:23 +0000 (18:03 -0300)]
V4L/DVB (3702): Make msp3400 routing defines more consistent
Renamed various msp3400 routing defines to be more consistent and less
confusing. Esp. the MSP_DSP_OUT defines were confusing since it is really
a DSP input.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Hans Verkuil [Sat, 1 Apr 2006 19:40:21 +0000 (16:40 -0300)]
V4L/DVB (3700): Remove obsolete commands from tvp5150.c
- Remove old DECODER_ commands from tvp5150.c, replacing them with newer
ones if appropriate.
- Small VIDIOC_G_TUNER fixes in msp3400 and tuner.
- Fix VIDIOC_S_TUNER support in em28xx.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Hans Verkuil [Sat, 1 Apr 2006 18:27:52 +0000 (15:27 -0300)]
V4L/DVB (3697): More msp3400 and bttv fixes
- remove obsolete VIDIOC_S_INPUT i2c call in bttv
- translate VIDIOCSFREQ to VIDIOC_S_FREQUENCY in i2c call
- improve muting during carrier scan in msp3400
- don't start scan unless really needed.
- no longer reset chip for msp3400c/d.
- remove v4l2 check in tuner-core (radio stops after using the TV)
- add missing VIDIOC_INT_ strings in v4l2-common.c
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Hans Verkuil [Thu, 30 Mar 2006 22:50:34 +0000 (19:50 -0300)]
V4L/DVB (3693): Fix msp3400c and bttv stereo/mono/bilingual detection/handling
- msp3400c did not detect the second carrier, thus being always mono.
- properly mute the msp3400c while detecting the carrier.
- fix checks on the presence of scart2/3 inputs and scart 2 output.
- implement proper audio mode fallbacks for msp3400c/d, identical to the
way msp3400g works.
- MODE_STEREO no longer produces dual languages when set for a bilingual
transmission, instead it falls back to LANG1. Use LANG1_LANG2 to hear
both languages of a bilingual transmission. This is much more intuitive
for the user and is in accordance with the preferred usage in the v4l2
specification.
- bttv tried to implement v4l2 calls with v4l1 calls to the i2c devices,
completely mangling the audmode/rxsubchans handling. v4l2 calls now do
v4l2 calls to the i2c devices.
- fixed broken i2c_vidiocschan in bttv.
- add start/end lines to LOG_STATUS.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Michael Krufky [Sat, 1 Apr 2006 21:00:41 +0000 (18:00 -0300)]
V4L/DVB (3689): Kconfig: fix VP-3054 Secondary I2C Bus build configuration menu dependencies
This patch fixes a dependency problem that affected the indentation order
within the individual frontend selection support menus for cx88-dvb.
- created a boolean dependency link for VIDEO_CX88_VP3054, so that
it's tristate value will be the same as that of VIDEO_CX88_DVB.
- VIDEO_CX88_VP3054 is automatically selected by VIDEO_CX88_DVB_ALL_FRONTENDS,
but is otherwise selected by VIDEO_CX88_DVB_VP3054, offered as an option
under VIDEO_CX88_DVB_MT352
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Unfortunately on the budget-av board, the CAM reset line is tied to the
frontend reset line, so resetting the CAM also zaps the frontend. This
breaks the tda1004x at least, and causes it to fail to tune until the
budget-av module is reloaded. This patch adds an exported function to dvb_frontend
that allows a card to forcibly reinitialise a frontend. The budget-av now
does this on CAM reset, which corrects this problem.
since they do not tie the CAM reset line to the frontend reset line.
Signed-off-by: Andrew de Quincey <adq_dvb@lidskialf.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Trent Piepho [Thu, 30 Mar 2006 18:53:32 +0000 (15:53 -0300)]
V4L/DVB (3672): Fix memory leak in dvr open
The dvr device could be opened multiple times simultaneously in O_RDONLY mode.
Each open after the first would allocate a new dvr buffer (1880 KB) and leak
the old buffer. The first close would de-allocate the dvr buffer and cause
all other open dvrs to stop working. This patch allows only a single O_RDONLY
open of the drv device, as per the API specification. Multiple O_WRONLY opens
are still allowed and don't appear to cause any problems.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org> Signed-off-by: Andrew de Quincey <adq_dvb@lidskialf.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
C.Y.M [Thu, 30 Mar 2006 07:31:48 +0000 (04:31 -0300)]
V4L/DVB (3671): New module parameter 'tv_standard' (dvb-ttpci driver)
This attached patch was originally proposed by Anssi Hannula to the dvb-kernel
user to choose the default broadcast mode when using the ttpci driver.
NTSC users need to only add the following line to modprobe.d:
options dvb-ttpci tv_standard=1
PAL users will not need to change anything, for this will be the default.
- Issue a warning when more than 80% of the DMA buffer is being used
(probably due to bad IRQ latency). Warnings are rate-limited.
- Introduce a new parameter 'bufsize' (in KByte) which increases the
default DMA buffer of 188 KByte up to 1410 KByte (Activy: 564 KByte).
Michael Krufky [Sun, 2 Apr 2006 06:14:11 +0000 (03:14 -0300)]
V4L/DVB (3653h): Move usb v4l docs into Documentation/video4linux
- Move documentation for usb v4l devices from
Documentation/usb to Documentation/video4linux.
- Removed trailing whitespace.
- Update Kconfig help text links to reflect the new file locations.
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
It should be V4L2_TUNER_MODE_LANG1_LANG2. What the code does is check if
we are NTSC and a SAP channel is available. If so, then the msp3400
should switch to standard 0x21 if the user wants to hear the SAP
channel, which is for audio modes LANG2 (aka SAP) and LANG1_LANG2
(bilingual).
In the msp3400 driver STEREO is abused for bilingual in PAL. Bilingual
never worked with NTSC in the past and I decided that I'd better not
use the bad PAL example.
Hans Verkuil [Wed, 29 Mar 2006 17:56:17 +0000 (14:56 -0300)]
V4L/DVB (3665): Add new NEC uPD64031A and uPD64083 i2c drivers
- Add support for the uPD64031A NEC Electronics Ghost Reduction i2c device
- Add support for the uPD6408x NEC Electronics 3-Dimensional Y/C separation
i2c device.
Hans Verkuil [Wed, 29 Mar 2006 17:31:44 +0000 (14:31 -0300)]
V4L/DVB (3663): Fix msp3400c wait time and better audio mode fallbacks
- The wait time until the first time the audio mode was detected was
1+5=6 seconds instead of just 1 second (wrong statement order).
msp3400c specific bug.
- Implemented audio mode fallback for msp3400c/d just like the msp3400g
does automatically. E.g. fallback to stereo if no second language exists,
etc.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Add support for the Wolfson Microelectronics WM8739
stereo A/D converter from the ivtv driver.
Many thanks to Takahiro Adachi for writing the original driver.
V4L/DVB (3655): Support for a new revision of the WT220U-stick
There seems to be a new version of the USB DVB-T stick from WideView with a new demod-revision inside and thus a new firwmare.
This patch enables support for that.
Thanks to Mikel Martin for early testing.
Signed-off-by: Patrick Boettcher <pb@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Hartmut Hackmann [Mon, 27 Mar 2006 22:39:30 +0000 (19:39 -0300)]
V4L/DVB (3646): Added support for the new Lifeview hybrid cardbus modules
There seem to be many variants of this cards with different
feature sets. This entry supports
analog TV, CVBS and s-video input, FM radio and DVB-T
if they are supported by the hardware.
Mike Isely [Sat, 25 Mar 2006 23:43:14 +0000 (20:43 -0300)]
V4L/DVB (3639): Reduce FWSEND due to certain I2C bus adapter limits
The FWSEND parameter controls the size of the firmware chunks sent
down the I2C bus to the chip. Previously this had been set to 1024
but unfortunately some I2C implementations can't transfer data in such
big gulps. Specifically, the pvrusb2 driver has a hard limit of
around 60 bytes, due to the encapsulation there of I2C traffic into
USB messages. So we have to significantly reduce this parameter. Acked-by: Hans Verkuil <hverkuil@xs4all.nl> Signed-off-by: Mike Isely <isely@pobox.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
After the recent video-buf "generic" adaptation, the PCI wrappers got
completely broken, and all of the DMA sound modules stopped working (and
failed with an oops)
Michael Krufky [Sun, 26 Mar 2006 08:43:36 +0000 (05:43 -0300)]
V4L/DVB (3617): Cxusb: add support for FusionHDTV USB portable remote control
- Added keycodes for the DViCO FusionHDTV portable remote control.
- Enabled the remote control for both versions of FusionHDTV DVB-T USB
and the FusionHDTV 5 USB Gold (ATSC)
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Samuel Thibault [Sun, 2 Apr 2006 05:10:28 +0000 (00:10 -0500)]
Input: add support for Braille devices
- Add KEY_BRL_* input keys and K_BRL_* keycodes;
- Add emulation of how braille keyboards usually combine braille dots
to the console keyboard driver;
- Add handling of unicode U+28xy diacritics.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org> Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Input: synaptics - limit rate to 40pps on Toshiba Protege M300
Toshiba Protege M300 also requires the same workaround as Satellites
and Dynabooks - Synaptics report rate should be lowered to 40pps
(from 80), otherwise KBC starts losing keypresses.
Signed-off-by: Richard Thrippleton <ret28@cam.ac.uk> Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
SNES gamepads and mice share the same type of interface so they both can be
connected to the parallel port using a simple interface. Adding mouse
support to a gamepad driver may sound funny at first, but doing so in this
case makes it possible to connect and SNES gamepads and mice at the same
time, on the same port.
[PATCH] powerpc: iSeries needs slb_initialize to be called
Since the powerpc 64k pages patch went in, systems that have SLBs
(like Power4 iSeries) needed to have slb_initialize called to set up
some variables for the SLB miss handler. This was not being called
on the boot processor on iSeries, so on single cpu iSeries machines,
we would get apparent memory curruption as soon as we entered user mode.
This patch fixes that by calling slb_initialize on the boot cpu if the
processor has an SLB.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
David S. Miller [Sat, 1 Apr 2006 20:48:36 +0000 (12:48 -0800)]
[NET]: Fully fix the memory leaks in sys_accept().
Andi Kleen was right, fput() on sock->file will end up calling
sock_release() if necessary. So here is the rest of his version
of the fix for these leaks.
Signed-off-by: David S. Miller <davem@davemloft.net>
Will Schmidt [Fri, 31 Mar 2006 15:07:48 +0000 (09:07 -0600)]
[PATCH] powerpc/pseries: misc lparcfg fixes
This fixes several problems with the lparcfg code. In case
someone gets a sense of deja-vu, part of this was submitted last Sep, I
thought the changes went in, but either got backed out, or just got
lost.
First, change the local_buffer declaration to be unsigned char *. We
had a bad-math problem in a 2.4 tree which was built with a
"-fsigned-char" parm. I dont believe we ever build with that parm
now-a-days, but to be safe, I'd prefer the declaration be explicit.
Second, fix a bad math calculation for splpar_strlen.
Third, on the rtas_call for get-system-parameter, pass in
RTAS_DATA_BUF_SIZE for the rtas_data_buf size, instead of letting random
data determine the size. Until recently, we've had a sufficiently
large 'random data' value get passed in, so the function just happens to
have worked OK. Now it's getting passed a '0', which causes the
rtas_call to return success, but no data shows up in the buffer.
(oops!). This was found by the LTC test org.
This is in a branch of code that only gets run on SPLPAR systems.
Tested on power5 Lpar.
Signed-off-by: Will Schmidt <willschm@us.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
[PATCH] powerpc: Extends HCALL interface for InfiniBand usage
This extends the HCALL interface for InfiniBand usage. I've
made the patch against the linux-2.6 git tree and Segher's patch:
[PATCH] Change H_StudlyCaps to H_SHOUTING_CAPS
We moved this into the common powerpc code based on comments we
got after posting the first eHCA InfiniBand device driver patch.
Signed-off-by: Heiko j Schick <schickhj@de.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
Linas Vepstas [Wed, 29 Mar 2006 21:29:18 +0000 (15:29 -0600)]
[PATCH] powerpc/pseries: mutex lock to serialize EEH event processing
This forces the processing of EEH PCI events to be serialized,
using a very simple mutex lock. This serialization is required to
avoid races involving additional PCI device failures that may occur
during the recovery phase of a previous failure.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch extends current iptables compatibility layer in order to get
32bit iptables to work on 64bit kernel. Current layer is insufficient due
to alignment checks both in kernel and user space tools.
Patch is for current net-2.6.17 with addition of move of ipt_entry_{match|
target} definitions to xt_entry_{match|target}.
Signed-off-by: Dmitry Mishin <dim@openvz.org> Acked-off-by: Kirill Korotaev <dev@openvz.org> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes expectation notifier unregistration on module unload to
use ip_conntrack_expect_unregister_notifier(). This bug causes a soft
lockup at the first expectation created after a rmmod ; insmod of this
module.
Should go into -stable as well.
Signed-off-by: Martin Josefsson <gandalf@wlug.westbo.se> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[NETFILTER]: fix ifdef for connmark support in nf_conntrack_netlink
When ctnetlink was ported from ip_conntrack to nf_conntrack two #ifdef's
for connmark support were left unchanged and this code was never
compiled.
Problem noticed by Daniel De Graaf.
Signed-off-by: Martin Josefsson <gandalf@wlug.westbo.se> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[NETFILTER]: x_tables: unify IPv4/IPv6 multiport match
This unifies ipt_multiport and ip6t_multiport to xt_multiport.
As a result, this addes support for inversion and port range match
to IPv6 packets.
Signed-off-by: Yasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
This unifies ipt_esp and ip6t_esp to xt_esp. Please note that now
a user program needs to specify IPPROTO_ESP as protocol to use esp match
with IPv6. This means that ip6tables requires '-p esp' like iptables.
Signed-off-by: Yasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Sat, 1 Apr 2006 08:52:46 +0000 (00:52 -0800)]
[IPSEC]: Kill unused decap state argument
This patch removes the decap_state argument from the xfrm input hook.
Previously this function allowed the input hook to share state with
the post_input hook. The latter has since been removed.
The only purpose for it now is to check the encap type. However, it
is easier and better to move the encap type check to the generic
xfrm_rcv function. This allows us to get rid of the decap state
argument altogether.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
MMIOs timeout more quickly that PCI config cycles and some
of these SRAM accesses can take a very long time, triggering
the MMIO limits on some sparc64 PCI controllers and thus
resulting in bus timeouts and bus errors.
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 31 Mar 2006 08:36:25 +0000 (00:36 -0800)]
[SPARC64]: Align address in huge_pte_alloc().
We are about to fill in all HPAGE_SIZE's worth
of PAGE_SIZE ptes, so we have to give the first
pte in that set else we scribble over random memory
when we fill in the ptes.
Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-2.6 / log