- 04 Nov, 2011 7 commits
-
-
git://git.samba.org/sfrench/cifs-2.6Linus Torvalds authored
* git://git.samba.org/sfrench/cifs-2.6: cifs: Assume passwords are encoded according to iocharset (try #2) CIFS: Fix the VFS brlock cache usage in posix locking case [CIFS] Update cifs version to 1.76 CIFS: Remove extra mutex_unlock in cifs_lock_add_if
-
git://git.kernel.org/pub/scm/linux/kernel/git/davem/netLinus Torvalds authored
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (45 commits) be2net: Add detect UE feature for Lancer be2net: Prevent CQ full condition for Lancer be2net: Fix disabling multicast promiscous mode be2net: Fix endian issue in RX filter command af_packet: de-inline some helper functions MAINTAINERS: Add can-gw include to maintained files net: Add back alignment for size for __alloc_skb net: add missing bh_unlock_sock() calls l2tp: fix race in l2tp_recv_dequeue() ixgbevf: Update release version ixgbe: DCB, return max for IEEE traffic classes ixgbe: fix reading of the buffer returned by the firmware ixgbe: Fix compiler warnings ixgbe: fix smatch splat due to missing NULL check ixgbe: fix disabling of Tx laser at probe ixgbe: Fix link issues caused by a reset while interface is down igb: Fix for I347AT4 PHY cable length unit detection e100: make sure vlan support isn't advertised on old adapters e1000e: demote a debugging WARN to a debug log message net: fix typo in drivers/net/ethernet/xilinx/ll_temac_main.c ...
-
Padmanabh Ratnakar authored
Add code to detect UE in case of Lancer. Signed-off-by:
Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Padmanabh Ratnakar authored
Indicate to HW that the CQ is cleaned up before posting new RX buffers. This prevents the HW to go into CQ full error condition. Signed-off-by:
-
Padmanabh Ratnakar authored
If user tries to disable multicast promiscous mode, the adapter remains in this mode as resetting the multicast promiscous mode was missing in RX filter command. Fixed this. Signed-off-by:
-
Padmanabh Ratnakar authored
Use cpu_to_le32() for mcast_num field in RX filter command as this field is of type u32. Signed-off-by:
-
-
- 03 Nov, 2011 15 commits
-
-
Olof Johansson authored
This popped some compiler errors due to mismatched prototypes. Just remove most manual inlines, the compiler should be able to figure out what makes sense to inline and not. net/packet/af_packet.c:252: warning: 'prb_curr_blk_in_use' declared inline after being called net/packet/af_packet.c:252: warning: previous declaration of 'prb_curr_blk_in_use' was here net/packet/af_packet.c:258: warning: 'prb_queue_frozen' declared inline after being called net/packet/af_packet.c:258: warning: previous declaration of 'prb_queue_frozen' was here net/packet/af_packet.c:248: warning: 'packet_previous_frame' declared inline after being called net/packet/af_packet.c:248: warning: previous declaration of 'packet_previous_frame' was here net/packet/af_packet.c:251: warning: 'packet_increment_head' declared inline after being called net/packet/af_packet.c:251: warning: previous declaration of 'packet_increment_head' was here Signed-off-by:
Olof Johansson <olof@lixom.net> Cc: Chetan Loke <loke.chetan@gmail.com> Signed-off-by:
-
Oliver Hartkopp authored
Commit c1aabdf3 (can-gw: add netlink based CAN routing) added a new include file that's neither referenced by any of the CAN maintainers. Signed-off-by:
Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by:
-
Tony Lindgren authored
Commit 87fb4b7b (net: more accurate skb truesize) changed the alignment of size. This can cause problems at least on some machines with NFS root: Unhandled fault: alignment exception (0x801) at 0xc183a43a Internal error: : 801 [#1] PREEMPT Modules linked in: CPU: 0 Not tainted (3.1.0-08784-g5eeee4a #733) pc : [<c02fbba0>] lr : [<c02fbb9c>] psr: 60000013 sp : c180fef8 ip : 00000000 fp : c181f580 r10: 00000000 r9 : c044b28c r8 : 00000001 r7 : c183a3a0 r6 : c1835be0 r5 : c183a412 r4 : 000001f2 r3 : 00000000 r2 : 00000000 r1 : ffffffe6 r0 : c183a43a Flags: nZCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment kernel Control: 0005317f Table: 10004000 DAC: 00000017 Process swapper (pid: 1, stack limit = 0xc180e270) Stack: (0xc180fef8 to 0xc1810000) fee0: 00000024 00000000 ff00: 00000000 c183b9c0 c183b8e0 c044b28c c0507ccc c019dfc4 c180ff2c c0503cf8 ff20: c180ff4c c180ff4c 00000000 c1835420 c182c740 c18349c0 c05233c0 00000000 ff40: 00000000 c00e6bb8 c180e000 00000000 c04dd82c c0507e7c c050cc18 c183b9c0 ff60: c05233c0 00000000 00000000 c01f34f4 c0430d70 c019d364 c04dd898 c04dd898 ff80: c04dd82c c0507e7c c180e000 00000000 c04c584c c01f4918 c04dd898 c04dd82c ffa0: c04ddd28 c180e000 00000000 c0008758 c181fa60 3231d82c 00000037 00000000 ffc0: 00000000 c04dd898 c04dd82c c04ddd28 00000013 00000000 00000000 00000000 ffe0: 00000000 c04b2224 00000000 c04b21a0 c001056c c001056c 00000000 00000000 Function entered at [<c02fbba0>] from [<c019dfc4>] Function entered at [<c019dfc4>] from [<c01f34f4>] Function entered at [<c01f34f4>] from [<c01f4918>] Function entered at [<c01f4918>] from [<c0008758>] Function entered at [<c0008758>] from [<c04b2224>] Function entered at [<c04b2224>] from [<c001056c>] Code: e1a00005 e3a01028 ebfa7cb0 e35a0000 (e5858028) Here PC is at __alloc_skb and &shinfo->dataref is unaligned because skb->end can be unaligned without this patch. As explained by Eric Dumazet <eric.dumazet@gmail.com>, this happens only with SLOB, and not with SLAB or SLUB: * Eric Dumazet <eric.dumazet@gmail.com> [111102 15:56]: > > Your patch is absolutely needed, I completely forgot about SLOB :( > > since, kmalloc(386) on SLOB gives exactly ksize=386 bytes, not nearest > power of two. > > [ 60.305763] malloc(size=385)->ffff880112c11e38 ksize=386 -> nsize=2 > [ 60.305921] malloc(size=385)->ffff88007c92ce28 ksize=386 -> nsize=2 > [ 60.306898] malloc(size=656)->ffff88007c44ad28 ksize=656 -> nsize=272 > [ 60.325385] malloc(size=656)->ffff88007c575868 ksize=656 -> nsize=272 > [ 60.325531] malloc(size=656)->ffff88011c777230 ksize=656 -> nsize=272 > [ 60.325701] malloc(size=656)->ffff880114011008 ksize=656 -> nsize=272 > [ 60.346716] malloc(size=385)->ffff880114142008 ksize=386 -> nsize=2 > [ 60.346900] malloc(size=385)->ffff88011c777690 ksize=386 -> nsize=2 Signed-off-by:
Tony Lindgren <tony@atomide.com> Acked-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
Eric Dumazet authored
Simon Kirby reported lockdep warnings and following messages : [104661.897577] huh, entered softirq 3 NET_RX ffffffff81613740 preempt_count 00000101, exited with 00000102? [104661.923653] huh, entered softirq 3 NET_RX ffffffff81613740 preempt_count 00000101, exited with 00000102? Problem comes from commit 0e734419 (ipv4: Use inet_csk_route_child_sock() in DCCP and TCP.) If inet_csk_route_child_sock() returns NULL, we should release socket lock before freeing it. Another lock imbalance exists if __inet_inherit_port() returns an error since commit 093d2823 ( tproxy: fix hash locking issue when using port redirection in __inet_inherit_port()) a backport is also needed for >= 2.6.37 kernels. Reported-by:
Simon Kirby <sim@hostway.ca> Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Tested-by:
-
Eric Dumazet authored
Misha Labjuk reported panics occurring in l2tp_recv_dequeue() If we release reorder_q.lock, we must not keep a dangling pointer (tmp), since another thread could manipulate reorder_q. Instead we must restart the scan at beginning of list. Reported-by:
Misha Labjuk <spiked.yar@gmail.com> Tested-by:
-
git://git.linux-mips.org/pub/scm/ralf/upstream-linusLinus Torvalds authored
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (37 commits) MIPS: O32: Provide definition of registers ta0 .. ta3. MIPS: perf: Add Octeon support for hardware perf. MIPS: perf: Add support for 64-bit perf counters. MIPS: perf: Reorganize contents of perf support files. MIPS: perf: Cleanup formatting in arch/mips/kernel/perf_event.c MIPS: Add accessor macros for 64-bit performance counter registers. MIPS: Add probes for more Octeon II CPUs. MIPS: Add more CPU identifiers for Octeon II CPUs. MIPS: XLR, XLS: Add comment for smp setup MIPS: JZ4740: GPIO: Check correct IRQ in demux handler MIPS: JZ4740: GPIO: Simplify IRQ demuxer MIPS: JZ4740: Use generic irq chip MIPS: Alchemy: remove all CONFIG_SOC_AU1??? defines MIPS: Alchemy: kill au1xxx.h header MIPS: Alchemy: clean DMA code of CONFIG_SOC_AU1??? defines MIPS, IDE: Alchem, au1xxx-ide: Remove pb1200/db1200 header dep MIPS: Alchemy: Redo PCI as platform driver MIPS: Alchemy: more base address cleanup MIPS: Alchemy: rewrite USB platform setup. MIPS: Alchemy: abstract USB block control register access ... Fix up trivial conflicts in: arch/mips/alchemy/devboards/db1x00/platform.c drivers/ide/Kconfig drivers/mmc/host/au1xmmc.c drivers/video/Kconfig sound/mips/Kconfig
-
Josh Boyer authored
The sb_edac driver is marginally useful on a 32-bit kernel, and currently has 64-bit divide compile errors when building that config. For now, make this build on only for 64-bit kernels. Signed-off-by:
Josh Boyer <jwboyer@redhat.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@redhat.com> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
git://github.com/kernelslacker/cpufreqLinus Torvalds authored
* 'next' of git://github.com/kernelslacker/cpufreq: [CPUFREQ] db8500: support all frequencies [CPUFREQ] db8500: remove unneeded for loop iteration over freq_table [CPUFREQ] ARM Exynos4210 PM/Suspend compatibility with different bootloaders [CPUFREQ] ARM: ux500: send cpufreq notification for all cpus [CPUFREQ] e_powersaver: Allow user to lower maximum voltage [CPUFREQ] e_powersaver: Check BIOS limit for CPU frequency [CPUFREQ] e_powersaver: Additional checks [CPUFREQ] exynos4210: Show list of available frequencies
-
git://git.infradead.org/users/sameo/mfd-2.6Linus Torvalds authored
* 'for-next' of git://git.infradead.org/users/sameo/mfd-2.6: (80 commits) mfd: Fix missing abx500 header file updates mfd: Add missing <linux/io.h> include to intel_msic x86, mrst: add platform support for MSIC MFD driver mfd: Expose TurnOnStatus in ab8500 sysfs mfd: Remove support for early drop ab8500 chip mfd: Add support for ab8500 v3.3 mfd: Add ab8500 interrupt disable hook mfd: Convert db8500-prcmu panic() into pr_crit() mfd: Refactor db8500-prcmu request_clock() function mfd: Rename db8500-prcmu init function mfd: Fix db5500-prcmu defines mfd: db8500-prcmu voltage domain consumers additions mfd: db8500-prcmu reset code retrieval mfd: db8500-prcmu tweak for modem wakeup mfd: Add db8500-pcmu watchdog accessor functions for watchdog mfd: hwacc power state db8500-prcmu accessor mfd: Add db8500-prcmu accessors for PLL and SGA clock mfd: Move to the new db500 PRCMU API mfd: Create a common interface for dbx500 PRCMU drivers mfd: Initialize DB8500 PRCMU regs ... Fix up trivial conflicts in arch/arm/mach-imx/mach-mx31moboard.c arch/arm/mach-omap2/board-omap3beagle.c arch/arm/mach-u300/include/mach/irqs.h drivers/mfd/wm831x-spi.c
-
git://github.com/pmundt/linux-shLinus Torvalds authored
* 'sh-latest' of git://github.com/pmundt/linux-sh: sh: Add default uImage rule for sh7757lcr sh: modify the asm/sh_eth.h to linux/sh_eth.h in sh7757lcr sh: userimask.c needs linux/stat.h sh: pfc: Add GPIO IRQ support sh: modify the asm/sh_eth.h to linux/sh_eth.h in some boards sh: pfc: Remove unused gpio_in_use member sh: add parameters for EHCI and RIIC in clock-sh7757.c sh: kexec: Add PHYSICAL_START SH: irq: Remove IRQF_DISABLED sh: pfc: get_config_reg() shift clean up sh: intc: Add IRQ trigger bit field check sh: drop unused Kconfig symbol sh: Fix implicit declaration of function numa_node_id sh: kexec: Register crashk_res sh: ecovec: add renesas_usbhs DMAEngine support
-
git://git.kernel.org/pub/scm/linux/kernel/git/ohad/hwspinlockLinus Torvalds authored
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ohad/hwspinlock: hwspinlock: add MAINTAINERS entries hwspinlock/omap: omap_hwspinlock_remove should be __devexit hwspinlock/u8500: add hwspinlock driver hwspinlock/core: register a bank of hwspinlocks in a single API call hwspinlock/core: remove stubs for register/unregister hwspinlock/core: use a mutex to protect the radix tree hwspinlock/core/omap: fix id issues on multiple hwspinlock devices hwspinlock/omap: simplify allocation scheme hwspinlock/core: simplify 'owner' handling hwspinlock/core: simplify Kconfig Fix up trivial conflicts (addition of omap_hwspinlock_pdata, removal of omap_spinlock_latency) in arch/arm/mach-omap2/hwspinlock.c Also, do an "evil merge" to fix a compile error in omap_hsmmc.c which for some reason was reported in the same email thread as the "please pull hwspinlock changes".
-
git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hidLinus Torvalds authored
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid: Revert "HID: multitouch: decide if hid-multitouch needs to handle mt devices" HID: drivers/hid/hid-roccat.c: eliminate a null pointer dereference HID: hid-apple: add device ID of another wireless aluminium HID: Add device IDs for Macbook Pro 8 keyboards
-
Linus Torvalds authored
This reverts commit 144060fe. It causes a resume regression for Andi on his Acer Aspire 1830T post 3.1. The screen just stays black after wakeup. Also, it really looks like the wrong way to suspend and resume perf events: I think they should be done as part of the CPU suspend and resume, rather than as a notifier that does smp_call_function(). Reported-by:
Andi Kleen <andi@firstfloor.org> Acked-by:
Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/steve/linux-dmLinus Torvalds authored
* git://git.kernel.org/pub/scm/linux/kernel/git/steve/linux-dm: dm: raid fix device status indicator when array initializing dm log userspace: add log device dependency dm log userspace: fix comment hyphens dm: add thin provisioning target dm: add persistent data library dm: add bufio dm: export dm get md dm table: add immutable feature dm table: add always writeable feature dm table: add singleton feature dm kcopyd: add dm_kcopyd_zero to zero an area dm: remove superfluous smp_mb dm: use local printk ratelimit dm table: propagate non rotational flag
-
git://git.selinuxproject.org/~jmorris/linux-securityLinus Torvalds authored
* 'for-linus' of git://git.selinuxproject.org/~jmorris/linux-security: TOMOYO: Fix interactive judgment functionality.
-
- 02 Nov, 2011 18 commits
-
-
Greg Rose authored
Signed-off-by:
Greg Rose <gregory.v.rose@intel.com> Tested-by:
Sibai Li <sibai.li@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
John Fastabend authored
Returning the max traffic classes on get requests simplifies user space configurations because applications will know explicitly how many traffic classes can be used. Typical switch implementations use 2 or 3 traffic classes so this not seen often today. And user space can learn the number of traffic classes by return codes but this allows user space to configure ixgbe correctly at the start. Signed-off-by:
John Fastabend <john.r.fastabend@intel.com> Tested-by:
Ross Brattain <ross.b.brattain@intel.com> Signed-off-by:
-
Emil Tantilov authored
This patch fixes some issues found in the buffer read portion of ixgbe_host_interface_command() - use `bi` as the buffer index counter instead of `i` - add conversion to native cpu byte ordering on register read - fix conversion from bytes to dword - use dword_len instead of buf_len when reading the register Signed-off-by:
Emil Tantilov <emil.s.tantilov@intel.com> Tested-by:
Phil Schmitt <phillip.j.schmitt@intel.com> Signed-off-by:
-
Greg Rose authored
Wrap SR-IOV specific functions in CONFIG_PCI_IOV to avoid compiler warnings. Signed-off-by:
-
John Fastabend authored
ixgbe_ieee_ets and ixgbe_ieee_pfc are intialized at the same time. Do a check for both before configuring IEEE802.1Qaz. Also max_frame was causing a sparse warning resolved here as well. Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Emil Tantilov authored
register_netdev() calls ndo_set_features() which may result in HW reset which in turn will bring the laser back up. This patch moves ixgbe_laser_tx_disable() below register_netdev() in ixgbe_probe() to make sure laser is shut off on load. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-edacLinus Torvalds authored
* 'linux_next' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-edac: (21 commits) MAINTAINERS: add an entry for Edac Sandy Bridge driver edac: tag sb_edac as EXPERIMENTAL, as it requires more testing EDAC: Fix incorrect edac mode reporting in sb_edac edac: sb_edac: Add it to the building system edac: Add an experimental new driver to support Sandy Bridge CPU's i7300_edac: Fix error cleanup logic i7core_edac: Initialize memory name with cpu, channel, bank i7core_edac: Fix compilation on 32 bits arch i7core_edac: scrubbing fixups EDAC: Correct Kconfig dependencies i7core_edac: return -ENODEV if no MC is found i7core_edac: use edac's own way to print errors MAINTAINERS: remove dropped edac_mce.* from the file i7core_edac: Drop the edac_mce facility x86, MCE: Use notifier chain only for MCE decoding EDAC i7core: Use mce socketid for better compatibility i7core_edac: Don't enable memory scrubbing for Xeon 35xx i7core_edac: Add scrubbing support edac: Move edac main structs to include/linux/edac.h i7core_edac: Fix oops when trying to inject errors ...
-
Emil Tantilov authored
Interface fails to obtain link on 82599 SFP in the following scenario: 1. Set advertised speed to GB: ethtool -s eth0 advertise 0x20 2. Bring interface down ip link set eth0 down 3. Issue any command that leads to a reset: ethtool -t eth0 4. Bring link back up: ip link set eth0 up Following patch makes sure that the driver flaps the Tx laser every time ixgbe_start_hw() is called, and not only when the speed is set. Signed-off-by:
-
Kantecki, Tomasz authored
The PHY cable length unit detection was not using the correct the correct PHY data variable for I347AT4. Signed-off-by:
Tomasz Kantecki <tomasz.kantecki@intel.com> Signed-off-by:
-
Jesse Brandeburg authored
e100 parts don't support vlan offload but they generally do allow use of vlans in higher software layers via the 8021q module. That said, there are a couple of really old revisions of e100 hardware that don't even allow the longer frame sizes required for vlan use with standard MTU. Use the VLAN_CHALLENGED flag to prevent vlan binding to these devices. Reported-by:
Michael Tokarev <mjt@tls.msk.ru> Signed-off-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> CC: Michael Tokarev <mjt@tls.msk.ru> CC: David Lamparter <equinox@diac24.net> Tested-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
-
Bruce Allan authored
This debugging message was recently added but it does not need to be as alarming as a WARN. Signed-off-by:
Bruce Allan <bruce.w.allan@intel.com> Tested-by:
-
git://linux-nfs.org/~bfields/linuxLinus Torvalds authored
* 'for-3.2' of git://linux-nfs.org/~bfields/linux: nfsd4: typo logical vs bitwise negate in nfsd4_decode_share_access
-
git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linuxLinus Torvalds authored
* 'misc-3.2' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux: MAINTAINERS: Update entry for IA64 [IA64] gpio: GENERIC_GPIO default must be n [IA64[ add CONFIG_NET_VENDOR_INTEL=y to default config files where needed [IA64] agp/hp-agp: Allow binding user memory to the AGP GART [IA64] sn2: add missing put_cpu()
-
Linus Torvalds authored
Says Andrew: "60 patches. That's good enough for -rc1 I guess. I have quite a lot of detritus to be rechecked, work through maintainers, etc. - most of the remains of MM - rtc - various misc - cgroups - memcg - cpusets - procfs - ipc - rapidio - sysctl - pps - w1 - drivers/misc - aio" * akpm: (60 commits) memcg: replace ss->id_lock with a rwlock aio: allocate kiocbs in batches drivers/misc/vmw_balloon.c: fix typo in code comment drivers/misc/vmw_balloon.c: determine page allocation flag can_sleep outside loop w1: disable irqs in critical section drivers/w1/w1_int.c: multiple masters used same init_name drivers/power/ds2780_battery.c: fix deadlock upon insertion and removal drivers/power/ds2780_battery.c: add a nolock function to w1 interface drivers/power/ds2780_battery.c: create central point for calling w1 interface w1: ds2760 and ds2780, use ida for id and ida_simple_get() to get it pps gpio client: add missing dependency pps: new client driver using GPIO pps: default echo function include/linux/dma-mapping.h: add dma_zalloc_coherent() sysctl: make CONFIG_SYSCTL_SYSCALL default to n sysctl: add support for poll() RapidIO: documentation update drivers/net/rionet.c: fix ethernet address macros for LE platforms RapidIO: fix potential null deref in rio_setup_device() RapidIO: add mport driver for Tsi721 bridge ...
-
Andrew Bresticker authored
While back-porting Johannes Weiner's patch "mm: memcg-aware global reclaim" for an internal effort, we noticed a significant performance regression during page-reclaim heavy workloads due to high contention of the ss->id_lock. This lock protects idr map, and serializes calls to idr_get_next() in css_get_next() (which is used during the memcg hierarchy walk). Since idr_get_next() is just doing a look up, we need only serialize it with respect to idr_remove()/idr_get_new(). By making the ss->id_lock a rwlock, contention is greatly reduced and performance improves. Tested: cat a 256m file from a ramdisk in a 128m container 50 times on each core (one file + container per core) in parallel on a NUMA machine. Result is the time for the test to complete in 1 of the containers. Both kernels included Johannes' memcg-aware global reclaim patches. Before rwlock patch: 1710.778s After rwlock patch: 152.227s Signed-off-by:
Andrew Bresticker <abrestic@google.com> Cc: Paul Menage <menage@gmail.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Acked-by:
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Jeff Moyer authored
In testing aio on a fast storage device, I found that the context lock takes up a fair amount of cpu time in the I/O submission path. The reason is that we take it for every I/O submitted (see __aio_get_req). Since we know how many I/Os are passed to io_submit, we can preallocate the kiocbs in batches, reducing the number of times we take and release the lock. In my testing, I was able to reduce the amount of time spent in _raw_spin_lock_irq by .56% (average of 3 runs). The command I used to test this was: aio-stress -O -o 2 -o 3 -r 8 -d 128 -b 32 -i 32 -s 16384 <dev> I also tested the patch with various numbers of events passed to io_submit, and I ran the xfstests aio group of tests to ensure I didn't break anything. Signed-off-by:
Jeff Moyer <jmoyer@redhat.com> Cc: Daniel Ehrenberg <dehrenberg@google.com> Signed-off-by:
-
Rakib Mullick authored
Fix typo in code comment. Signed-off-by:
Rakib Mullick <rakib.mullick@gmail.com> Acked-by:
Dmitry Torokhov <dtor@vmware.com> Signed-off-by:
-
Rakib Mullick authored
In vmballoon_reserve_page(), flags has been passed from the callee function (vmballoon_inflate here). So, we can determine can_sleep outside the loop. Signed-off-by:
-
