- 08 Feb, 2008 40 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/christoph/vmLinus Torvalds authored
* 'slub-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/christoph/vm: SLUB: fix checkpatch warnings Use non atomic unlock SLUB: Support for performance statistics SLUB: Alternate fast paths using cmpxchg_local SLUB: Use unique end pointer for each slab page. SLUB: Deal with annoying gcc warning on kfree()
-
Ingo Molnar authored
fix checkpatch --file mm/slub.c errors and warnings. $ q-code-quality-compare errors lines of code errors/KLOC mm/slub.c [before] 22 4204 5.2 mm/slub.c [after] 0 4210 0 no code changed: text data bss dec hex filename 22195 8634 136 30965 78f5 slub.o.before 22195 8634 136 30965 78f5 slub.o.after md5: 93cdfbec2d6450622163c590e1064358 slub.o.before.asm 93cdfbec2d6450622163c590e1064358 slub.o.after.asm [clameter: rediffed against Pekka's cleanup patch, omitted moves of the name of a function to the start of line] Signed-off-by:Ingo Molnar <mingo@elte.hu> Signed-off-by:
Christoph Lameter <clameter@sgi.com>
-
Nick Piggin authored
Slub can use the non-atomic version to unlock because other flags will not get modified with the lock held. Signed-off-by:
Nick Piggin <npiggin@suse.de> Acked-by:
Andrew Morton <akpm@linux-foundation.org>
-
Christoph Lameter authored
The statistics provided here allow the monitoring of allocator behavior but at the cost of some (minimal) loss of performance. Counters are placed in SLUB's per cpu data structure. The per cpu structure may be extended by the statistics to grow larger than one cacheline which will increase the cache footprint of SLUB. There is a compile option to enable/disable the inclusion of the runtime statistics and its off by default. The slabinfo tool is enhanced to support these statistics via two options: -D Switches the line of information displayed for a slab from size mode to activity mode. -A Sorts the slabs displayed by activity. This allows the display of the slabs most important to the performance of a certain load. -r Report option will report detailed statistics on Example (tbench load): slabinfo -AD ->Shows the most active slabs Name Objects Alloc Free %Fast skbuff_fclone_cache 33 111953835 111953835 99 99 :0000192 2666 5283688 5281047 99 99 :0001024 849 5247230 5246389 83 83 vm_area_struct 1349 119642 118355 91 22 :0004096 15 66753 66751 98 98 :0000064 2067 25297 23383 98 78 dentry 10259 28635 18464 91 45 :0000080 11004 18950 8089 98 98 :0000096 1703 12358 10784 99 98 :0000128 762 10582 9875 94 18 :0000512 184 9807 9647 95 81 :0002048 479 9669 9195 83 65 anon_vma 777 9461 9002 99 71 kmalloc-8 6492 9981 5624 99 97 :0000768 258 7174 6931 58 15 So the skbuff_fclone_cache is of highest importance for the tbench load. Pretty high load on the 192 sized slab. Look for the aliases slabinfo -a | grep 000192 :0000192 <- xfs_btree_cur filp kmalloc-192 uid_cache tw_sock_TCP request_sock_TCPv6 tw_sock_TCPv6 skbuff_head_cache xfs_ili Likely skbuff_head_cache. Looking into the statistics of the skbuff_fclone_cache is possible through slabinfo skbuff_fclone_cache ->-r option implied if cache name is mentioned .... Usual output ... Slab Perf Counter Alloc Free %Al %Fr -------------------------------------------------- Fastpath 111953360 111946981 99 99 Slowpath 1044 7423 0 0 Page Alloc 272 264 0 0 Add partial 25 325 0 0 Remove partial 86 264 0 0 RemoteObj/SlabFrozen 350 4832 0 0 Total 111954404 111954404 Flushes 49 Refill 0 Deactivate Full=325(92%) Empty=0(0%) ToHead=24(6%) ToTail=1(0%) Looks good because the fastpath is overwhelmingly taken. skbuff_head_cache: Slab Perf Counter Alloc Free %Al %Fr -------------------------------------------------- Fastpath 5297262 5259882 99 99 Slowpath 4477 39586 0 0 Page Alloc 937 824 0 0 Add partial 0 2515 0 0 Remove partial 1691 824 0 0 RemoteObj/SlabFrozen 2621 9684 0 0 Total 5301739 5299468 Deactivate Full=2620(100%) Empty=0(0%) ToHead=0(0%) ToTail=0(0%) Descriptions of the output: Total: The total number of allocation and frees that occurred for a slab Fastpath: The number of allocations/frees that used the fastpath. Slowpath: Other allocations Page Alloc: Number of calls to the page allocator as a result of slowpath processing Add Partial: Number of slabs added to the partial list through free or alloc (occurs during cpuslab flushes) Remove Partial: Number of slabs removed from the partial list as a result of allocations retrieving a partial slab or by a free freeing the last object of a slab. RemoteObj/Froz: How many times were remotely freed object encountered when a slab was about to be deactivated. Frozen: How many times was free able to skip list processing because the slab was in use as the cpuslab of another processor. Flushes: Number of times the cpuslab was flushed on request (kmem_cache_shrink, may result from races in __slab_alloc) Refill: Number of times we were able to refill the cpuslab from remotely freed objects for the same slab. Deactivate: Statistics how slabs were deactivated. Shows how they were put onto the partial list. In general fastpath is very good. Slowpath without partial list processing is also desirable. Any touching of partial list uses node specific locks which may potentially cause list lock contention. Signed-off-by: -
Christoph Lameter authored
Provide an alternate implementation of the SLUB fast paths for alloc and free using cmpxchg_local. The cmpxchg_local fast path is selected for arches that have CONFIG_FAST_CMPXCHG_LOCAL set. An arch should only set CONFIG_FAST_CMPXCHG_LOCAL if the cmpxchg_local is faster than an interrupt enable/disable sequence. This is known to be true for both x86 platforms so set FAST_CMPXCHG_LOCAL for both arches. Currently another requirement for the fastpath is that the kernel is compiled without preemption. The restriction will go away with the introduction of a new per cpu allocator and new per cpu operations. The advantages of a cmpxchg_local based fast path are: 1. Potentially lower cycle count (30%-60% faster) 2. There is no need to disable and enable interrupts on the fast path. Currently interrupts have to be disabled and enabled on every slab operation. This is likely avoiding a significant percentage of interrupt off / on sequences in the kernel. 3. The disposal of freed slabs can occur with interrupts enabled. The alternate path is realized using #ifdef's. Several attempts to do the same with macros and inline functions resulted in a mess (in particular due to the strange way that local_interrupt_save() handles its argument and due to the need to define macros/functions that sometimes disable interrupts and sometimes do something else). [clameter: Stripped preempt bits and disabled fastpath if preempt is enabled] Signed-off-by:
Pekka Enberg <penberg@cs.helsinki.fi> Cc: <linux-arch@vger.kernel.org> Signed-off-by:
-
Christoph Lameter authored
We use a NULL pointer on freelists to signal that there are no more objects. However the NULL pointers of all slabs match in contrast to the pointers to the real objects which are in different ranges for different slab pages. Change the end pointer to be a pointer to the first object and set bit 0. Every slab will then have a different end pointer. This is necessary to ensure that end markers can be matched to the source slab during cmpxchg_local. Bring back the use of the mapping field by SLUB since we would otherwise have to call a relatively expensive function page_address() in __slab_alloc(). Use of the mapping field allows avoiding a call to page_address() in various other functions as well. There is no need to change the page_mapping() function since bit 0 is set on the mapping as also for anonymous pages. page_mapping(slab_page) will therefore still return NULL although the mapping field is overloaded. Signed-off-by:
-
Christoph Lameter authored
gcc 4.2 spits out an annoying warning if one casts a const void * pointer to a void * pointer. No warning is generated if the conversion is done through an assignment. Signed-off-by: -
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6Linus Torvalds authored
* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (48 commits) [SCSI] aacraid: do not set valid bit in sense information [SCSI] ses: add new Enclosure ULD [SCSI] enclosure: add support for enclosure services [SCSI] sr: fix test unit ready responses [SCSI] u14-34f: fix data direction bug [SCSI] aacraid: pci_set_dma_max_seg_size opened up for late model controllers [SCSI] fix BUG when sum(scatterlist) > bufflen [SCSI] arcmsr: updates (1.20.00.15) [SCSI] advansys: make 3 functions static [SCSI] Small cleanups for scsi_host.h [SCSI] dc395x: fix uninitialized var warning [SCSI] NCR53C9x: remove driver [SCSI] remove m68k NCR53C9x based drivers [SCSI] dec_esp: Remove driver [SCSI] kernel-doc: fix scsi docbook [SCSI] update my email address [SCSI] add protocol definitions [SCSI] sd: handle bad lba in sense information [SCSI] qla2xxx: Update version number to 8.02.00-k8. [SCSI] qla2xxx: Correct issue where incorrect init-fw mailbox command was used on non-NPIV capable ISPs. ...
-
git://git.linux-nfs.org/pub/linux/nfs-2.6Linus Torvalds authored
* git://git.linux-nfs.org/pub/linux/nfs-2.6: SUNRPC xptrdma: simplify build configuration NFS: Fix a potential file corruption issue when writing
-
David Sterba authored
The device is manufactured by IPWireless. In some countries (for example Czech Republic, T-Mobile ISP) this card is shipped for service called UMTS 4G. It's a piece of PCMCIA "4G" UMTS PPP networking hardware that presents itself as a serial character device (i.e. looks like usual modem to userspace, accepts AT commands, etc). Rewieved-by:
Jiri Slaby <jslaby@suse.cz> Signed-off-by:
Ben Martel <benm@symmetric.co.nz> Signed-off-by:
Stephen Blackheath <stephen@symmetric.co.nz> Signed-off-by:
David Sterba <dsterba@suse.cz> Signed-off-by:
Jiri Kosina <jkosina@suse.cz> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
James Lentini authored
Trond and Bruce, This is a patch for 2.6.25. This is the same version that was sent out on December 12 for review (no comments to date). To simplify the RPC/RDMA client and server build configuration, make SUNRPC_XPRT_RDMA a hidden config option that continues to depend on SUNRPC and INFINIBAND. The value of SUNRPC_XPRT_RDMA will be: - N if either SUNRPC or INFINIBAND are N - M if both SUNRPC and INFINIBAND are on (M or Y) and at least one is M - Y if both SUNRPC and INFINIBAND are Y In 2.6.25, all of the RPC/RDMA related files are grouped in net/sunrpc/xprtrdma and the net/sunrpc/xprtrdma/Makefile builds both the client and server RPC/RDMA support using this config option. Signed-off-by:
James Lentini <jlentini@netapp.com> Signed-off-by:
Trond Myklebust <Trond.Myklebust@netapp.com>
-
Salyzyn, Mark authored
Luben Tuikov [mailto:ltuikov@yahoo.com] sez: > Just as in your case and Tony's case, which I presume > uses the same RAID firmware vendor, it would've > probably been better if the RAID firmware vendor > fixed the firmware to not set the VALID bit if the > INFORMATION field is not valid. Point taken regarding the aacraid driver. Dropped the VALID bit, and then did some cleanup/simplification of the set_sense procedure and the associated parameters. Mike did some preliminary tests when the VALID bit was dropped before the 'Re: [PATCH] [SCSI] sd: make error handling more robust' patches came on the scene. The change in the SCSI subsystem does make this enclosed aacraid patch unnecessary, so this aacraid patch is merely post battle ground cleanup. If the simplification is an issue, repugnant, too much for a back-port to the stable trees or clouds the point, this patch could be happily distilled down to: diff -ru a/drivers/scsi/aacraid/aachba.c b/drivers/scsi/aacraid/aachba.c --- a/drivers/scsi/aacraid/aachba.c 2008-02-06 16:26:45.834938955 -0500 +++ b/drivers/scsi/aacraid/aachba.c 2008-02-06 16:32:01.109035329 -0500 @@ -865,7 +865,7 @@ u32 residue) { - sense_buf[0] = 0xF0; /* Sense data valid, err code 70h (current error) */ + sense_buf[0] = 0x70; /* Sense data invalid, err code 70h (current error) */ sense_buf[1] = 0; /* Segment number, always zero */ if (incorrect_length) { Signed-off-by:Mark Salyzyn <aacraid@adaptec.com> Signed-off-by:
James Bottomley <James.Bottomley@HansenPartnership.com>
-
Trond Myklebust authored
If the inode is flagged as having an invalid mapping, then we can't rely on the PageUptodate() flag. Ensure that we don't use the "anti-fragmentation" write optimisation in nfs_updatepage(), since that will cause NFS to write out areas of the page that are no longer guaranteed to be up to date. A potential corruption could occur in the following scenario: client 1 client 2 =============== =============== fd=open("f",O_CREAT|O_WRONLY,0644); write(fd,"fubar\n",6); // cache last page close(fd); fd=open("f",O_WRONLY|O_APPEND); write(fd,"foo\n",4); close(fd); fd=open("f",O_WRONLY|O_APPEND); write(fd,"bar\n",4); close(fd); ----- The bug may lead to the file "f" reading 'fubar\n\0\0\0\nbar\n' because client 2 does not update the cached page after re-opening the file for write. Instead it keeps it marked as PageUptodate() until someone calls invaldate_inode_pages2() (typically by calling read()). Signed-off-by: -
James Bottomley authored
This adds support to SCSI for enclosure services devices. It also makes use of the enclosure services added in an earlier patch to display the enclosure topology in sysfs. At the moment, the enclosures are SAS specific, but if anyone actually has a non-SAS enclosure that follows the SES-2 standard, we can add that as well. On my Vitesse based system, the enclosures show up like this: sparkweed:~# ls -l /sys/class/enclosure/0\:0\:1\:0/ total 0 -r--r--r-- 1 root root 4096 2008-02-03 15:44 components lrwxrwxrwx 1 root root 0 2008-02-03 15:44 device -> ../../../devices/pci0000:01/0000:01:02.0/host0/port-0:0/expander-0:0/port-0:0:12/end_device-0:0:12/target0:0:1/0:0:1:0 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 000 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 001 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 002 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 003 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 004 drwxr-xr-x 2 root root 0 2008-02-03 15:44 SLOT 005 lrwxrwxrwx 1 root root 0 2008-02-03 15:44 subsystem -> ../../enclosure --w------- 1 root root 4096 2008-02-03 15:44 uevent And the individual occupied slots like this: sparkweed:~# ls -l /sys/class/enclosure/0\:0\:1\:0/SLOT\ 001/ total 0 -rw-r--r-- 1 root root 4096 2008-02-03 15:45 active lrwxrwxrwx 1 root root 0 2008-02-03 15:45 device -> ../../../../devices/pci0000:01/0000:01:02.0/host0/port-0:0/expander-0:0/port-0:0:11/end_device-0:0:11/target0:0:0/0:0:0:0 -rw-r--r-- 1 root root 4096 2008-02-03 15:45 fault -rw-r--r-- 1 root root 4096 2008-02-03 15:45 locate -rw-r--r-- 1 root root 4096 2008-02-03 15:45 status lrwxrwxrwx 1 root root 0 2008-02-03 15:45 subsystem -> ../../../enclosure_component -r--r--r-- 1 root root 4096 2008-02-03 15:45 type --w------- 1 root root 4096 2008-02-03 15:45 uevent You can flash the various blinky lights by echoing to the fault and locate files. >From the device's point of view, you can see it has an enclosure like this: sparkweed:~# ls /sys/class/scsi_disk/0\:0\:0\:0/device/ block:sda generic queue_depth state bsg:0:0:0:0 iocounterbits queue_type subsystem bus iodone_cnt rescan timeout delete ioerr_cnt rev type device_blocked iorequest_cnt scsi_device:0:0:0:0 uevent driver modalias scsi_disk:0:0:0:0 vendor enclosure_component:SLOT 001 model scsi_generic:sg0 evt_media_change power scsi_level Note the enclosure_component:SLOT 001 which shows where in the enclosure this device fits. The astute will notice that I'm using SCSI VPD Inquiries to identify the devices. This, unfortunately, won't work for SATA devices unless we do some really nasty hacking about on the SAT because the only think that knows the SAS addresses for SATA devices is libsas, not libata where the SAT resides. Signed-off-by: -
James Bottomley authored
The enclosure misc device is really just a library providing sysfs support for physical enclosure devices and their components. Signed-off-by: -
James Bottomley authored
Commit 210ba1d1 updated sr.c to use the scsi_test_unit_ready() function. Unfortunately, this has the wrong characteristic of eating NOT_READY returns which sr.c relies on for tray status. Fix by rolling an internal sr_test_unit_ready() that doesn't do this. Tested-by:
Daniel Drake <dsd@gentoo.org> Signed-off-by:
-
Roel Kluin authored
Direction of data transfer 'DMA_FROM_DEVICE' was tested twice. DTD_OUT means transfer from host to device. This should occur when the direction of data transfer (sc_data_direction) is 'DMA_TO_DEVICE'. Signed-off-by:
Roel Kluin <12o3l@tiscali.nl> Signed-off-by:
-
Salyzyn, Mark authored
This patch ensures that the modern adapters get a maximum sg segment size on par with the maximum transfer size. Added some localized janitor fixes to the discussion patch I used with Fujita. FUJITA Tomonori [mailto:fujita.tomonori@lab.ntt.co.jp] sez: > I think that setting the proper maximum segment size for the late > model cards (as you did above) makes sense. Signed-off-by:
-
Tony Battersby authored
When sending a SCSI command to a tape drive via the SCSI Generic (sg) driver, if the command has a data transfer length more than scatter_elem_sz (32 KB default) and not a multiple of 512, then I either hit BUG_ON(!valid_dma_direction(direction)) in dma_unmap_sg() or else the command never completes (depending on the LLDD). When constructing scatterlists, the sg driver rounds up the scatterlist element sizes to be a multiple of 512. This can result in sum(scatterlist lengths) > bufflen. In this case, scsi_req_map_sg() incorrectly sets bio->bi_size to sum(scatterlist lengths) rather than to bufflen. When the command completes, req_bio_endio() detects that bio->bi_size != 0, and so it doesn't call bio_endio(). This causes the command to be resubmitted, resulting in BUG_ON or the command never completing. This patch makes scsi_req_map_sg() set bio->bi_size to bufflen rather than to sum(scatterlist lengths), which fixes the problem. Signed-off-by:
Tony Battersby <tonyb@cybernetics.com> Acked-by:
Mike Christie <michaelc@cs.wisc.edu> Signed-off-by:
-
Nick Cheng authored
- add arcmsr_enable_eoi_mode()and readl(reg->iop2drv_doorbell_reg) in arcmsr_handle_hbb_isr() on adapter Type B in case of the doorbell interrupt clearance is cached - add conditional declaration for arcmsr_pci_error_detected() and arcmsr_pci_slot_reset - check if the sg list member number exceeds arcmsr default limit in arcmsr_build_ccb() - change the returned value type of arcmsr_build_ccb()from "void" to "int" returns FAILED in arcmsr_queue_command() - modify arcmsr_drain_donequeue() to ignore unknown command and let kernel process command timeout. This could handle IO request violating maximum segments, i.e. Linux XFS over DM-CRYPT. Thanks to Milan Broz's comments <mbroz@redhat.com> - fix the release of dma memory for type B in arcmsr_free_ccb_pool() - fix the arcmsr_polling_hbb_ccbdone() Signed-off-by:
Nick Cheng <nick.cheng@areca.com.tw> Cc: Milan Broz <mbroz@redhat.com> Cc: <thenzl@redhat.com> Signed-off-by:
-
Adrian Bunk authored
Signed-off-by:
Adrian Bunk <bunk@kernel.org> Signed-off-by:
-
Pavel Machek authored
Small cleanups in scsi_host.h. Few #defines make me wonder if their description is still up to date..? Signed-off-by:
Pavel Machek <pavel@suse.cz> Signed-off-by:
-
Andrew Morton authored
drivers/scsi/dc395x.c: In function 'dc395x_init_one': drivers/scsi/dc395x.c:4270: warning: 'ptr' may be used uninitialized in this function Signed-off-by:
-
James Bottomley authored
Acked-by:
Boaz Harrosh <bharrosh@panasas.com> Signed-off-by:
-
James Bottomley authored
These drivers depend on the deprecated NCR53C9X core and need to be converted to the esp_scsi core. Acked-by:
-
James Bottomley authored
This driver depends on the deprecated NCR53C9X core and needs to be converted to the esp_scsi core. Acked-by:
-
Randy Dunlap authored
Add missing function parameter descriptions. Make function short description fit on one line as required. Signed-off-by:
Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by:
-
James Bottomley authored
This updates steeleye -> hansenpartnership in the documentation since some email has been going astray because of this. Signed-off-by: -
James Bottomley authored
A lot of SCSI command replies have a protocol ID field. Add definitions for the interpretation of that from SPC-3. Signed-off-by: -
James Bottomley authored
Some devices report medium error locations incorrectly. Add guards to make sure the reported bad lba is actually in the request that caused it. Additionally remove the large case statment for sector sizes and replace it with the proper u64 divisions. Tested-by:
Mike Snitzer <snitzer@gmail.com> Cc: Stable Tree <stable@kernel.org> Cc: Tony Battersby <tonyb@cybernetics.com> Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by:
-
Andrew Vasquez authored
[SCSI] qla2xxx: Correct issue where incorrect init-fw mailbox command was used on non-NPIV capable ISPs. BIT_2 of the firmware attributes is only valid on FW-interface-2 type HBAs. Code in commit c48339de would cause the incorrect initialize-firmware mailbox command to be issued for non-NPIV capable ISPs. Correct this by reverting to previously used (and correct) pre-condition 'if' check. Signed-off-by:
-
Seokmann Ju authored
For following fc_host specific attributes, vports rely on the pport. So, this patch changed way to access the data for those attributes so that they can access pport's. - get_host_speed (speed) - get_host_port_state (port_state) - get_host_port_type (port_type) - get_fc_host_stats Also, added PORT_SPEED_8GB case in the speed attribute for 8Gb HBAs. Signed-Off-by:
Seokmann Ju <seokmann.ju@qlogic.com> Signed-off-by:
-
Seokmann Ju authored
While running IO simultaneously through physical port and virtual port, if user changes Data Rate (from scli utility), IO through virtual port fails. It failed because the vport had not received the ISP_ABORT_NEEDED notification. Signed-Off-by:
-
Andrew Vasquez authored
There's no functional change involved with this update, instead it simply migrates the "set cleared interrupt state" codes to a more approprate method, qla2x00_request_irqs(), and cleans-up the driver's probe() logic. Signed-off-by:
-
Andrew Vasquez authored
Collapse duplicate codes called during probe() and RISC-reset into qla2x00_setup_chip(). Signed-off-by:
-
Andrew Vasquez authored
- Drop loop-till-allocated structure of code within qla2x00_mem_alloc(). - Properly unwind deallcations of memory during failures. - Drop qla2x00_allocate_sp_pool() and qla2x00_free_sp_pool() functions as their implementations can easily be collapsed into the callers. - Defer DMA pool allocation of SFP data until requested. Signed-off-by:
-
Andrew Vasquez authored
To insure that there is no stale data present during EFT re-registration. Signed-off-by:
-
Andrew Vasquez authored
Refactor SRB-failure completion codes in the process. Also, signal the DPC routine to complete sooner as backend processing at shutdown-time is superflous. [jejb: resolve conflicts with pci_enable_device_bars removal] Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
-
