| Age | Commit message (Collapse) | Author |
|---|
|
There's a race between shutting down one io scheduler and firing up the
next, in which a new io could enter and cause the io scheduler to be
invoked with bad or NULL data.
To fix this, we need to maintain the queue lock for a bit longer.
Unfortunately we cannot do that, since the elevator init requires to be
run without the lock held. This isn't easily fixable, without also
changing the mempool API. So split the initialization into two parts,
and alloc-init operation and an attach operation. Then we can
preallocate the io scheduler and related structures, and run the attach
inside the lock after we detach the old one.
This patch has survived 30 minutes of 1 second io scheduler switching
with a very busy io load.
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
From: Ralf Baechle <ralf@linux-mips.org>
<linux/mempolicy.h> uses struct mm_struct and relies on a definition or
declaration somehow magically being dragged in which may result in a
build:
[...]
CC mm/mempolicy.o
In file included from mm/mempolicy.c:69:
include/linux/mempolicy.h:150: warning: âstruct mm_structâ declared inside parameter list
include/linux/mempolicy.h:150: warning: its scope is only this definition or declaration, which is probably not what you want
include/linux/mempolicy.h:175: warning: âstruct mm_structâ declared inside parameter list
mm/mempolicy.c:622: error: conflicting types for âdo_migrate_pagesâ
include/linux/mempolicy.h:175: error: previous declaration of âdo_migrate_pagesâ was here
mm/mempolicy.c:1661: error: conflicting types for âmpol_rebind_mmâ
include/linux/mempolicy.h:150: error: previous declaration of âmpol_rebind_mmâ was here
make[1]: *** [mm/mempolicy.o] Error 1
make: *** [mm] Error 2
[ralf@denk linux-ip35]$
Including <linux/sched.h> is a step into direction of include hell so
fixed by adding a forward declaration of struct mm_struct instead.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
From: Andrew Morton <akpm@osdl.org>
drivers/rtc/rtc-m48t86.c: In function `m48t86_rtc_read_time':
drivers/rtc/rtc-m48t86.c:51: error: structure has no member named `ia64_mv'
drivers/rtc/rtc-m48t86.c:55: error: structure has no member named `ia64_mv'
drivers/rtc/rtc-m48t86.c:56: error: structure has no member named `ia64_mv'
drivers/rtc/rtc-m48t86.c:57: error: structure has no member named `ia64_mv'
drivers/rtc/rtc-m48t86.c:58: error: structure has no member named `ia64_mv'
drivers/rtc/rtc-m48t86.c:60: error: structure has no member named `ia64_mv'
readb() and writeb() are macros on ia64.
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
From: Ralf Baechle <ralf@linux-mips.org>
<linux/mmzone.h> uses PAGE_SIZE, PAGE_SHIFT from <asm/page.h> without
including that header itself. For some sparsemem configurations this may
result in build errors like:
CC init/initramfs.o
In file included from include/linux/gfp.h:4,
from include/linux/slab.h:15,
from include/linux/percpu.h:4,
from include/linux/rcupdate.h:41,
from include/linux/dcache.h:10,
from include/linux/fs.h:226,
from init/initramfs.c:2:
include/linux/mmzone.h:498:22: warning: "PAGE_SHIFT" is not defined
In file included from include/linux/gfp.h:4,
from include/linux/slab.h:15,
from include/linux/percpu.h:4,
from include/linux/rcupdate.h:41,
from include/linux/dcache.h:10,
from include/linux/fs.h:226,
from init/initramfs.c:2:
include/linux/mmzone.h:526: error: `PAGE_SIZE' undeclared here (not in a function)
include/linux/mmzone.h: In function `__pfn_to_section':
include/linux/mmzone.h:573: error: `PAGE_SHIFT' undeclared (first use in this function)
include/linux/mmzone.h:573: error: (Each undeclared identifier is reported only once
include/linux/mmzone.h:573: error: for each function it appears in.)
include/linux/mmzone.h: In function `pfn_valid':
include/linux/mmzone.h:578: error: `PAGE_SHIFT' undeclared (first use in this function)
make[1]: *** [init/initramfs.o] Error 1
make: *** [init] Error 2
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Seems-reasonable-to: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
From: Andrew Morton <akpm@osdl.org>
Revert commit ff4da2e262d2509fe1bacff70dd00934be569c66.
It broke APM suspend, probably because APM doesn't switch back to a VT
when suspending.
Tracked down by Matt Mackall <mpm@selenic.com>
Rafael sayeth:
"It only fixed the theoretical issue that a quick-handed user could
switch to X after processes have been frozen and before the devices
are suspended.
With the current userland suspend tools it shouldn't be necessary."
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
* master.kernel.org:/pub/scm/linux/kernel/git/davej/agpgart:
[AGPGART] VIA PT880 Ultra support.
[AGPGART] Fix Nforce3 suspend on amd64.
[AGPGART] Enable SIS AGP driver on x86-64 for EM64T systems
|
|
Remove the numbered SW_* entries from the input system and assign names
to the existing users.
Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
|
|
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
|
|
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
These flags are needed by userspace - move them outside __KERNEL__
(Pointed out by dwmw2)
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
This patch enables agpgart on a Via "PT880 Ultra" based motherboard
(Asus P4V800D-X). The PCI ID of the PT880 Ultra is 0x0308 instead of
0x0258 of the PT880.
The patched via-agp passes testgart.
Signed-off-by: Magnus Kessler <Magnus.Kessler@gmx.net>
Signed-off-by: Dave Jones <davej@redhat.com>
|
|
* master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb: (33 commits)
V4L/DVB (3965): Fix CONFIG_VIDEO_VIVI=y build bug
V4L/DVB (3964): Bt8xx/bttv-cards.c: fix off-by-one errors
V4L/DVB (3914): Vivi build fix
V4L/DVB (3912): Sparc32 vivi fix
V4L/DVB (3832): Get_dvb_firmware: download nxt2002 firmware from new driver location
V4L/DVB (3829): Fix frequency values in the ranges structures of the LG TDVS H06xF tuners
V4L/DVB (3826): Saa7134: Missing 'break' in Terratec Cinergy 400 TV initialization
V4L/DVB (3825): Remove broken 'fast firmware load' from cx25840.
V4L/DVB (3819): Cxusb-bluebird: bug-fix: power down corrupts frontend
V4L/DVB (3813): Add support for TCL M2523_5N_E tuner.
V4L/DVB (3804): Tweak bandselect setup fox cx24123
V4L/DVB (3803): Various correctness fixes to tuning.
V4L/DVB (3797): Always wait for diseqc queue to become ready before transmitting a diseqc message
V4L/DVB (3796): Add several debug messages to cx24123 code
V4L/DVB (3795): Fix for CX24123 & low symbol rates
V4L/DVB (3792): Kbuild: DVB_BT8XX must select DVB_ZL10353
V4L/DVB (3790): Use after free in drivers/media/video/em28xx/em28xx-video.c
V4L/DVB (3788): Fix compilation with V4L1_COMPAT
V4L/DVB (3782): Removed uneeded stuff from pwc Makefile
V4L/DVB (3775): Add VIVI Kconfig stuff
...
|
|
Andy added code to buddy allocator which does not require the zone's
endpoints to be aligned to MAX_ORDER. An issue is that the buddy allocator
requires the node_mem_map's endpoints to be MAX_ORDER aligned. Otherwise
__page_find_buddy could compute a buddy not in node_mem_map for partial
MAX_ORDER regions at zone's endpoints. page_is_buddy will detect that
these pages at endpoints are not PG_buddy (they were zeroed out by bootmem
allocator and not part of zone). Of course the negative here is we could
waste a little memory but the positive is eliminating all the old checks
for zone boundary conditions.
SPARSEMEM won't encounter this issue because of MAX_ORDER size constraint
when SPARSEMEM is configured. ia64 VIRTUAL_MEM_MAP doesn't need the logic
either because the holes and endpoints are handled differently. This
leaves checking alloc_remap and other arches which privately allocate for
node_mem_map.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
- remove the following global function that is both unused and
unimplemented:
- register_firmware()
- make the following needlessly global function static:
- firmware_class_uevent()
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
This driver supports the SPI controller on the MPC83xx SoC devices from
Freescale. Note, this driver supports only the simple shift register SPI
controller and not the descriptor based CPM or QUICCEngine SPI controller.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
command and cleanups
Patch from Pavel Pisa
There has been problems that for some paths that clock are not stopped
during new command programming and initiation. Result is issuing
of incorrect command to the card. Some other problems are cleaned too.
Noisy report of known ERRATUM #4 has been suppressed.
Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
We need to be able to have a "SPI bus 0" matching chip numbering; but
that number was wrongly used to flag dynamic allocation of a bus number.
This patch resolves that issue; now negative numbers trigger dynamic alloc.
It also updates the how-to-write-a-controller-driver overview to mention
this stuff.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Add spi_device hook for LSB-first word encoding, and update all the
(in-tree) controller drivers to reject such devices. Eventually,
some controller drivers will be updated to support lsb-first encodings
on the wire; no current drivers need this.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
and export it
Renamed bitbang_transfer_setup to follow convention of other exported symbols
from spi-bitbang. Exported spi_bitbang_setup_transfer to allow users of
spi-bitbang to use the function in their own setup_transfer.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This removes superfluous whitespace in the <linux/spi/spi.h> header.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Some protocols (like one for some bitmap displays) require different clock
speed or word size settings for each transfer in an SPI message. This adds
those parameters to struct spi_transfer. They are to be used when they are
nonzero; otherwise the defaults from spi_device are to be used.
The patch also adds a setup_transfer callback to spi_bitbang, uses it for
messages that use those overrides, and implements it so that the pure
bitbanging code can help resolve any questions about how it should work.
Signed-off-by: Imre Deak <imre.deak@nokia.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
can_share_swap_page() is used to check if the page has the last reference.
This avoids allocating a new page for COW if it's the last page.
However, if CONFIG_SWAP is not set, can_share_swap_page() is defined as 0,
thus always causes a copy for the last COW page. The below simple patch
fixes it.
Signed-off-by: Hua Zhong <hzhong@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
slab_is_available() indicates slab based allocators are available for use.
SPARSEMEM code needs to know this as it can be called at various times
during the boot process.
Signed-off-by: Mike Kravetz <kravetz@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
Even since a previous patch:
Fix race between CONFIG_DEBUG_SLABALLOC and modules
Sun, 27 Jun 2004 17:55:19 +0000 (17:55 +0000)
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/old-2.6-bkcvs.git;a=commit;h=92b3db26d31cf21b70e3c1eadc56c179506d8fbe
The function symbol_put_addr() will deadlock the kernel.
symbol_put_addr() would acquire modlist_lock, then while holding the lock call
two functions kernel_text_address() and module_text_address() which also try
to acquire the same lock. This deadlocks the kernel of course.
This patch changes symbol_put_addr() to not acquire the modlist_lock, it
doesn't need it since it never looks at the module list directly. Also, it
now uses core_kernel_text() instead of kernel_text_address(). The latter has
an additional check for addr inside a module, but we don't need to do that
since we call module_text_address() (the same function kernel_text_address
uses) ourselves.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Cc: Zwane Mwaikambo <zwane@fsmlabs.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
With "Paul E. McKenney" <paulmck@us.ibm.com>
Introduce rcu_needs_cpu() interface. This can be used to tell if there
will be a new rcu batch on a cpu soon by looking at the curlist pointer.
This can be used to avoid to enter a tickless idle state where the cpu
would miss that a new batch is ready when rcu_start_batch would be called
on a different cpu.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
V4L1 API is depreciated and should be removed soon from kernel. This patch
adds two new options, one to disable V4L1 drivers, and another to disable
V4L1 compat module. This way, it would be easy to check what still depends
on V4L1 stuff, allowing also to test if app works fine with V4L2 only support.
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
|
|
* master.kernel.org:/home/rmk/linux-2.6-serial:
[SERIAL] 8250: add locking to console write function
[SERIAL] Remove unconditional enable of TX irq for console
[SERIAL] 8250: set divisor register correctly for AMD Alchemy SoC uart
[SERIAL] AMD Alchemy UART: claim memory range
[SERIAL] Clean up serial locking when obtaining a reference to a port
|
|
* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
[NET_SCHED]: HFSC: fix thinko in hfsc_adjust_levels()
[IPV6]: skb leakage in inet6_csk_xmit
[BRIDGE]: Do sysfs registration inside rtnl.
[NET]: Do sysfs registration as part of register_netdevice.
[TG3]: Fix possible NULL deref in tg3_run_loopback().
[NET] linkwatch: Handle jiffies wrap-around
[IRDA]: Switching to a workqueue for the SIR work
[IRDA]: smsc-ircc: Minimal hotplug support.
[IRDA]: Removing unused EXPORT_SYMBOLs
[IRDA]: New maintainer.
[NET]: Make netdev_chain a raw notifier.
[IPV4]: ip_options_fragment() has no effect on fragmentation
[NET]: Add missing operstates documentation.
|
|
master.kernel.org:/pub/scm/linux/kernel/git/shemminger/netdev-2.6
* 'upstream' of master.kernel.org:/pub/scm/linux/kernel/git/shemminger/netdev-2.6:
sis900: phy for FoxCon motherboard
dl2k: use DMA_48BIT_MASK constant
phy: mdiobus_register(): initialize all phy_map entries
sky2: ifdown kills irq mask
|
|
Typo will be harder with this one.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
|
|
The last step of netdevice registration was being done by a delayed
call, but because it was delayed, it was impossible to return any error
code if the class_device registration failed.
Side effects:
* one state in registration process is unnecessary.
* register_netdevice can sleep inside class_device registration/hotplug
* code in netdev_run_todo only does unregistration so it is simpler.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This is a backout of earlier patch.
The whole rescheduling hack was a bad idea. It doesn't really solve
the problem and it makes the code more complicated for no good reason.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
|
|
Atomically create attributes when class device is added. This avoids
the race between registering class_device (which generates hotplug
event), and the creation of attribute groups.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Extend the support of attribute groups in class_device's to allow
groups to be created as part of the registration process. This allows
network device's to avoid race between registration and creating
groups.
Note that unlike attributes that are a property of the class object,
the groups are a property of the class_device object. This is done
because there are different types of network devices (wireless for
example).
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
[PATCH] powerpc: Use the ibm,pa-features property if available
powerpc: Fix incorrect might_sleep in __get_user/__put_user on kernel addresses
[PATCH] ppc32 CPM_UART: fixes and improvements
[PATCH] ppc32 CPM_UART: Fixed break send on SCC
[PATCH] powerpc/kprobes: fix singlestep out-of-line
[PATCH] powerpc/pseries: avoid crash in PCI code if mem system not up
|
|
* master.kernel.org:/home/rmk/linux-2.6-mmc:
[MMC] Move set_ios debugging into mmc.c
[MMC] Correct mmc_request_done comments
[MMC] PXA: reduce the number of lines PXAMCI debug uses
[MMC] PXA and i.MX: don't avoid sending stop command on error
[MMC] extend data timeout for writes
[ARM] 3485/1: i.MX: MX1 SD/MMC fix of unintentional double start possibility
|
|
* 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block:
[PATCH] compat_sys_vmsplice: one-off in UIO_MAXIOV check
[PATCH] splice: redo page lookup if add_to_page_cache() returns -EEXIST
[PATCH] splice: rename remaining info variables to pipe
[PATCH] splice: LRU fixups
[PATCH] splice: fix unlocking of page on error ->prepare_write()
|
|
Signed-off-by: Jing Min Zhao <zhaojingmin@users.sourceforge.net>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Nick says that the current construct isn't safe. This goes back to the
original, but sets PIPE_BUF_FLAG_LRU on user pages as well as they all
seem to be on the LRU in the first place.
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
A number of small issues are fixed, and added the header file, missed from the
original series. With this, driver should be pretty stable as tested among
both platform-device-driven and "old way" boards. Also added missing GPL
statement , and updated year field on existing ones to reflect
code update.
Signed-off-by: Vitaly Bordug <vbordug@ru.mvista.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
|
|
The CSD contains a "read2write factor" which determines the multiplier to
be applied to the read timeout to obtain the write timeout. We were
ignoring this parameter, resulting in the possibility for writes being
timed out too early.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
Apply the same rules as the anon pipe pages, only allow stealing
if no one else is using the page.
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
Currently we rely on the PIPE_BUF_FLAG_LRU flag being set correctly
to know whether we need to fiddle with page LRU state after stealing it,
however for some origins we just don't know if the page is on the LRU
list or not.
So remove PIPE_BUF_FLAG_LRU and do this check/add manually in pipe_to_file()
instead.
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/viro/audit-current
* 'audit.b10' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/audit-current:
[PATCH] Audit Filter Performance
[PATCH] Rework of IPC auditing
[PATCH] More user space subject labels
[PATCH] Reworked patch for labels on user space messages
[PATCH] change lspp ipc auditing
[PATCH] audit inode patch
[PATCH] support for context based audit filtering, part 2
[PATCH] support for context based audit filtering
[PATCH] no need to wank with task_lock() and pinning task down in audit_syscall_exit()
[PATCH] drop task argument of audit_syscall_{entry,exit}
[PATCH] drop gfp_mask in audit_log_exit()
[PATCH] move call of audit_free() into do_exit()
[PATCH] sockaddr patch
[PATCH] deal with deadlocks in audit_free()
|
|
When iptables userspace adds an ipt_standard_target, it calculates the size
of the entire entry as:
sizeof(struct ipt_entry) + XT_ALIGN(sizeof(struct ipt_standard_target))
ipt_standard_target looks like this:
struct xt_standard_target
{
struct xt_entry_target target;
int verdict;
};
xt_entry_target contains a pointer, so when compiled for 64 bit the
structure gets an extra 4 byte of padding at the end. On 32 bit
architectures where iptables aligns to 8 byte it will also have 4
byte padding at the end because it is only 36 bytes large.
The compat_ipt_standard_fn in the kernel adjusts the offsets by
sizeof(struct ipt_standard_target) - sizeof(struct compat_ipt_standard_target),
which will always result in 4, even if the structure from userspace
was already padded to a multiple of 8. On x86 this works out by
accident because userspace only aligns to 4, on all other
architectures this is broken and causes incorrect adjustments to
the size and following offsets.
Thanks to Linus for lots of debugging help and testing.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
* 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block:
[PATCH] vmsplice: allow user to pass in gift pages
[PATCH] pipe: enable atomic copying of pipe data to/from user space
[PATCH] splice: call handle_ra_miss() on failure to lookup page
[PATCH] Add ->splice_read/splice_write to def_blk_fops
[PATCH] pipe: introduce ->pin() buffer operation
[PATCH] splice: fix bugs in pipe_to_file()
[PATCH] splice: fix bugs with stealing regular pipe pages
|
|
If SPLICE_F_GIFT is set, the user is basically giving this pages away to
the kernel. That means we can steal them for eg page cache uses instead
of copying it.
The data must be properly page aligned and also a multiple of the page size
in length.
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
The pipe ->map() method uses kmap() to virtually map the pages, which
is both slow and has known scalability issues on SMP. This patch enables
atomic copying of pipe pages, by pre-faulting data and using kmap_atomic()
instead.
lmbench bw_pipe and lat_pipe measurements agree this is a Good Thing. Here
are results from that on a UP machine with highmem (1.5GiB of RAM), running
first a UP kernel, SMP kernel, and SMP kernel patched.
Vanilla-UP:
Pipe bandwidth: 1622.28 MB/sec
Pipe bandwidth: 1610.59 MB/sec
Pipe bandwidth: 1608.30 MB/sec
Pipe latency: 7.3275 microseconds
Pipe latency: 7.2995 microseconds
Pipe latency: 7.3097 microseconds
Vanilla-SMP:
Pipe bandwidth: 1382.19 MB/sec
Pipe bandwidth: 1317.27 MB/sec
Pipe bandwidth: 1355.61 MB/sec
Pipe latency: 9.6402 microseconds
Pipe latency: 9.6696 microseconds
Pipe latency: 9.6153 microseconds
Patched-SMP:
Pipe bandwidth: 1578.70 MB/sec
Pipe bandwidth: 1579.95 MB/sec
Pipe bandwidth: 1578.63 MB/sec
Pipe latency: 9.1654 microseconds
Pipe latency: 9.2266 microseconds
Pipe latency: 9.1527 microseconds
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
The ->map() function is really expensive on highmem machines right now,
since it has to use the slower kmap() instead of kmap_atomic(). Splice
rarely needs to access the virtual address of a page, so it's a waste
of time doing it.
Introduce ->pin() to take over the responsibility of making sure the
page data is valid. ->map() is then reduced to just kmap(). That way we
can also share a most of the pipe buffer ops between pipe.c and splice.c
Signed-off-by: Jens Axboe <axboe@suse.de>
|
|
Found by Oleg Nesterov <oleg@tv-sign.ru>, fixed by me.
- Only allow full pages to go to the page cache.
- Check page != buf->page instead of using PIPE_BUF_FLAG_STOLEN.
- Remember to clear 'stolen' if add_to_page_cache() fails.
And as a cleanup on that:
- Make the bottom fall-through logic a little less convoluted. Also make
the steal path hold an extra reference to the page, so we don't have
to differentiate between stolen and non-stolen at the end.
Signed-off-by: Jens Axboe <axboe@suse.de>
|
