mirror of
https://github.com/SCST-project/scst.git
synced 2026-09-04 07:06:57 +00:00
Merged revisions 5875,5878-5895,5903-5905,5910,5912-5914,5928-5929,5931-5991 via svnmerge from
svn+ssh://yanb123@svn.code.sf.net/p/scst/svn/trunk
........
r5875 | bvassche | 2014-11-16 19:58:07 +0200 (Sun, 16 Nov 2014) | 1 line
nightly build: Update kernel versions
........
r5878 | bvassche | 2014-11-19 02:17:41 +0200 (Wed, 19 Nov 2014) | 1 line
srpt/Makefile: Add double quotes around a path
........
r5879 | bvassche | 2014-11-19 02:20:20 +0200 (Wed, 19 Nov 2014) | 1 line
scripts/generate-release-archive: Accept an optional list of file names
........
r5880 | bvassche | 2014-11-22 13:12:29 +0200 (Sat, 22 Nov 2014) | 1 line
nightly build: Update kernel versions
........
r5881 | bvassche | 2014-11-24 19:59:14 +0200 (Mon, 24 Nov 2014) | 4 lines
ib_srpt: Add support for HCA's that do not support SRQ
Based on a patch provided by Parav Pandit <Parav.Pandit@Emulex.Com>
........
r5882 | vlnb | 2014-11-26 09:02:17 +0200 (Wed, 26 Nov 2014) | 3 lines
Update for kernels 3.17.x
........
r5883 | bvassche | 2014-11-26 10:05:09 +0200 (Wed, 26 Nov 2014) | 1 line
Add kernel 3.17 build infrastructure
........
r5884 | bvassche | 2014-11-26 10:07:08 +0200 (Wed, 26 Nov 2014) | 1 line
nightly build: Add kernel 3.17
........
r5885 | bvassche | 2014-11-26 10:16:44 +0200 (Wed, 26 Nov 2014) | 6 lines
Fix kernel 3.17 checkpatch warnings about 'long long unsigned'
Avoid that checkpatch reports the following warning:
WARNING: type 'long long unsigned' should be specified in 'unsigned long long' order.
........
r5886 | bvassche | 2014-11-26 15:38:52 +0200 (Wed, 26 Nov 2014) | 1 line
Build fixes for RHEL 6.6 kernel 2.6.32-504
........
r5887 | bvassche | 2014-11-26 16:39:51 +0200 (Wed, 26 Nov 2014) | 1 line
ib_srpt: Make the send queue full messages more informational
........
r5888 | bvassche | 2014-11-26 18:25:57 +0200 (Wed, 26 Nov 2014) | 1 line
scripts/specialize-patch: Support blanks around numbers inside parentheses
........
r5889 | bvassche | 2014-11-26 21:42:10 +0200 (Wed, 26 Nov 2014) | 1 line
scripts/specialize-patch: Reduce noise in nightly build output
........
r5890 | vlnb | 2014-11-27 06:36:33 +0200 (Thu, 27 Nov 2014) | 3 lines
Cleanup
........
r5891 | bvassche | 2014-11-27 17:18:58 +0200 (Thu, 27 Nov 2014) | 1 line
scst.h: Add uintptr_t
........
r5892 | bvassche | 2014-11-27 17:19:21 +0200 (Thu, 27 Nov 2014) | 1 line
ib_srpt: Add support for immediate data
........
r5893 | bvassche | 2014-11-27 17:24:17 +0200 (Thu, 27 Nov 2014) | 1 line
ib_srpt: Log reject reason
........
r5894 | bvassche | 2014-11-27 17:29:29 +0200 (Thu, 27 Nov 2014) | 1 line
ib_srpt: Rework the max_sge computation changes from r5795
........
r5895 | bvassche | 2014-11-28 11:16:37 +0200 (Fri, 28 Nov 2014) | 1 line
scst: Add scripts/rebuild-rhel-kernel-rpm to the SCST release archive
........
r5903 | bvassche | 2014-12-03 13:50:06 +0200 (Wed, 03 Dec 2014) | 4 lines
scripts/rebuild-rhel-kernel-rpm: Fix an error message
Reported-by: Hiroyuki Sato <hiroysato@gmail.com>
........
r5904 | bvassche | 2014-12-03 19:06:57 +0200 (Wed, 03 Dec 2014) | 1 line
iscsi-scst/kernel/patches/rhel/put_page_callback-2.6.32-504.patch: Add
........
r5905 | bvassche | 2014-12-03 19:07:31 +0200 (Wed, 03 Dec 2014) | 1 line
scripts/rebuild-rhel-kernel-rpm: Add support for RHEL 6.6
........
r5910 | bvassche | 2014-12-04 13:50:58 +0200 (Thu, 04 Dec 2014) | 1 line
scripts/generate-kernel-patch: Swap two filters
........
r5912 | bvassche | 2014-12-04 14:19:56 +0200 (Thu, 04 Dec 2014) | 4 lines
/etc/init.d/scst: Exit with status code 0 upon 'start' if already running
Reported-by: Dimitar Tanev <dimitar@linuxdevgroup.org>
........
r5913 | vlnb | 2014-12-05 01:41:52 +0200 (Fri, 05 Dec 2014) | 3 lines
FORMAT commands should be strictly serialized
........
r5914 | vlnb | 2014-12-05 01:43:51 +0200 (Fri, 05 Dec 2014) | 3 lines
Oops, fix for the previous commit
........
r5928 | vlnb | 2014-12-06 07:02:27 +0200 (Sat, 06 Dec 2014) | 3 lines
Web updates
........
r5929 | bvassche | 2014-12-09 14:33:16 +0200 (Tue, 09 Dec 2014) | 1 line
rpm build: Add support for qla2x00t driver in QLogic git repository
........
r5931 | vlnb | 2014-12-11 06:27:17 +0200 (Thu, 11 Dec 2014) | 3 lines
Docs update
........
r5932 | vlnb | 2014-12-11 06:34:36 +0200 (Thu, 11 Dec 2014) | 8 lines
scst_vdisk: Increase virtual device name length
This change makes integration with OpenStack easier since OpenStack GUIDs
are 36 characters long: 32 hex characters and four dashes.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5933 | vlnb | 2014-12-11 06:38:04 +0200 (Thu, 11 Dec 2014) | 11 lines
vdisk_blockio: Report invalid scatterlists
It is possible for a target driver to pass a scatterlist via
scst_cmd_set_tgt_sg() that is valid for the vdisk_fileio handler
but not for the vdisk_blockio handler. Complain loudly if an invalid
scatterlist is passed to vdisk_blockio because such scatterlists
cause silent data corruption with most Linux block drivers.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5934 | bvassche | 2014-12-11 14:31:03 +0200 (Thu, 11 Dec 2014) | 1 line
scst_vdisk: Follow-up for r5932
........
r5935 | bvassche | 2014-12-11 14:37:02 +0200 (Thu, 11 Dec 2014) | 1 line
ib_srpt: Log P_Key during login
........
r5936 | bvassche | 2014-12-12 11:29:42 +0200 (Fri, 12 Dec 2014) | 1 line
scripts/generate-kernel-patch: Include scst_pg.sgml instead of sgv_cache.sgml
........
r5937 | bvassche | 2014-12-12 11:34:55 +0200 (Fri, 12 Dec 2014) | 1 line
doc/scst_pg.sgml: Remove trailing whitespace
........
r5938 | bvassche | 2014-12-17 09:48:40 +0200 (Wed, 17 Dec 2014) | 1 line
nightly build: Update kernel versions
........
r5939 | vlnb | 2014-12-19 05:50:58 +0200 (Fri, 19 Dec 2014) | 3 lines
Fallback to the old qla driver if the git one not detected
........
r5940 | vlnb | 2014-12-19 05:55:14 +0200 (Fri, 19 Dec 2014) | 7 lines
Replace in cases, where sporadic failures are possible, HARDWARE ERROR
by INTERNAL TARGET FAILURE, which is retriable (some OS'es don't retry
HARDWARE ERROR)
Reported and suggested by Shahar Salzman <shahar.salzman@kaminario.com>
........
r5941 | vlnb | 2014-12-20 05:48:07 +0200 (Sat, 20 Dec 2014) | 7 lines
scst_vdisk: Only accept NAA IDs allowed by SPC
See also paragraph 7.8.6.6 NAA designator format in SPC-4.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5942 | vlnb | 2014-12-20 05:49:23 +0200 (Sat, 20 Dec 2014) | 11 lines
scst_vdisk: Remove superfluous llseek() calls
vfs_read() and vfs_write() ignore the file offset set by llseek().
Hence remove the llseek() calls that occur just before vfs_read() and
vfs_write(). See also the implementation in the Linux kernel of the
pread64() and pwrite64() system calls for examples of code that uses
vfs_read() and vfs_write().
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5943 | bvassche | 2014-12-22 14:28:13 +0200 (Mon, 22 Dec 2014) | 1 line
Source code spelling fix: Equivilant -> Equivalent
........
r5944 | bvassche | 2014-12-22 14:28:56 +0200 (Mon, 22 Dec 2014) | 1 line
Source code spelling fix: accesss -> access
........
r5945 | bvassche | 2014-12-22 14:29:51 +0200 (Mon, 22 Dec 2014) | 1 line
Source code spelling fix: addres -> address
........
r5946 | bvassche | 2014-12-22 14:31:08 +0200 (Mon, 22 Dec 2014) | 1 line
Source code spelling fix: authentification -> authentication
........
r5947 | bvassche | 2014-12-22 14:32:30 +0200 (Mon, 22 Dec 2014) | 1 line
Source code comment spelling fix: explicitely -> explicitly
........
r5948 | bvassche | 2014-12-22 14:33:06 +0200 (Mon, 22 Dec 2014) | 1 line
Source code comment spelling fix: hander -> handler
........
r5949 | bvassche | 2014-12-22 14:33:37 +0200 (Mon, 22 Dec 2014) | 1 line
Source code comment spelling fix: loosing -> losing
........
r5950 | bvassche | 2014-12-22 14:35:00 +0200 (Mon, 22 Dec 2014) | 1 line
Spelling fix: occured -> occurred
........
r5951 | bvassche | 2014-12-22 14:35:51 +0200 (Mon, 22 Dec 2014) | 1 line
Source code comment spelling fix: refering -> referring
........
r5952 | bvassche | 2014-12-22 14:36:47 +0200 (Mon, 22 Dec 2014) | 1 line
Spelling fix: shrinked -> shrunk
........
r5953 | bvassche | 2014-12-22 15:08:34 +0200 (Mon, 22 Dec 2014) | 1 line
Spelling fix: choosen -> chosen
........
r5954 | bvassche | 2014-12-22 15:09:20 +0200 (Mon, 22 Dec 2014) | 1 line
Spelling fix: existant -> existent
........
r5955 | bvassche | 2014-12-22 15:10:41 +0200 (Mon, 22 Dec 2014) | 1 line
Update for kernel 3.18
........
r5956 | bvassche | 2014-12-22 15:15:55 +0200 (Mon, 22 Dec 2014) | 1 line
Spelling fix: immediatelly -> immediately
........
r5957 | bvassche | 2014-12-24 16:28:36 +0200 (Wed, 24 Dec 2014) | 1 line
nightly build: Add kernel 3.18
........
r5958 | bvassche | 2014-12-29 14:14:52 +0200 (Mon, 29 Dec 2014) | 1 line
scst_lib: Convert spaces into tabs (reported by checkpatch)
........
r5959 | bvassche | 2015-01-06 15:25:28 +0200 (Tue, 06 Jan 2015) | 1 line
scst_calc_block_shift: Log block shift and sector size upon mismatch
........
r5960 | bvassche | 2015-01-07 11:20:06 +0200 (Wed, 07 Jan 2015) | 4 lines
scst_local: Fix unique per session sas address
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
........
r5961 | bvassche | 2015-01-09 14:23:25 +0200 (Fri, 09 Jan 2015) | 4 lines
scst_sysfs: return EINVAL on too big LUN
Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
........
r5962 | bvassche | 2015-01-10 17:52:57 +0200 (Sat, 10 Jan 2015) | 1 line
nightly build: Update kernel versions
........
r5963 | bvassche | 2015-01-13 10:42:28 +0200 (Tue, 13 Jan 2015) | 10 lines
scst: Switch to thread context before executing a reservation command
Persistent reservation commands need thread context because
scst_pr_is_cmd_allowed() locks the PR mutex. Reservation commands
either need BH or thread context. Hence switch from atomic to
thread context before processing such commands.
Reported-by: Shahar Salzman <shahar.salzman@kaminario.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5964 | bvassche | 2015-01-13 10:51:08 +0200 (Tue, 13 Jan 2015) | 5 lines
scst_parse_unmap_descriptors(): Avoid using GFP_KERNEL in atomic context
Reported-by: Shahar Salzman <shahar.salzman@kaminario.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5965 | bvassche | 2015-01-13 10:55:46 +0200 (Tue, 13 Jan 2015) | 68 lines
qla2x00t: Copy entire SCST sense buffer to q2x ctio
There seems to be a bug in passing sense information to QLA HBAs, where
the last 2 bytes of the sense data (ASC, ASCQ) are not copied to the low
level sense buffer.
We encountered this in ESX, which relies on these 2 bytes to parse the
MISCOMPARE sense code (0xE1, 0x1D, 0x00).
Bellow is a simple test to recreate this issue, but during vMotion
operations (where VMs are moved from one host to another), this may
cause the operation to fail leaving the VM in an inconsistent state.
The test I ran to verify that we are indeed missing the bytes is the
following:
1. Create a SCST based device
2. Expose the device to 2 ESX hosts
3. Format the device as VMFS5, create a test directory
4. From both hosts, I start writing to this directory (no VMs involved,
just write normal files)
At this stage, both ESX hosts try to take access to the directory.
The VMFS filesystem contains a per-directory lock which is managed by
COMPARE AND WRITE command.
Each ESX will attempt to change the VMFS lock location from unlocked to
locked to create the new file.
Obviously there are bound to be failures (which are equivalent to
programming locking conflicts), these are reported by the MISCOMPARE
sense code.
Upon these MISCOMPARE errors, the host will re-try taking the lock until
it succeeds, and will then proceed to perform the write operation on the
directory.
Due to the bug in copying the sense buffer from the SCST core to the QLA
ctio, instead of the full sense code, only the key (0xE) is sent, and
ESX does not know how to handle it resulting in IO error.
Here are the errors as they appear on the command line:
/vmfs/volumes/54a297c4-ca5af1cc-7f94-002219d20f28/ats_test #
./open_close_test-esx2.sh
./open_close_test-esx2.sh: line 8: can't create
ats_fileoptest-esx2_1.txt: Input/output error
./open_close_test-esx2.sh: line 8: can't create
ats_fileoptest-esx2_21.txt: Input/output error
./open_close_test-esx2.sh: line 8: can't create
ats_fileoptest-esx2_110.txt: Input/output error
./open_close_test-esx2.sh: line 8: can't create
ats_fileoptest-esx2_111.txt: Input/output error
In the /var/log/vmkernel.log, we can see that the sense information is
missing (0xE, 0x0, 0x0) instead of (0xE, 0x1D, 0x0).
2014-12-30T12:13:20.714Z cpu6:33519)ScsiDeviceIO: 2338:
Cmd(0x412e84f957c0) 0x89, CmdSN 0x234d from world 519051 to dev
"eui.0024f400d5020007" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0xe 0x0 0x0.
2014-12-30T12:13:20.766Z cpu6:33519)ScsiDeviceIO: 2338:
Cmd(0x412e84f91d00) 0x89, CmdSN 0x2350 from world 519051 to dev
"eui.0024f400d5020007" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0xe 0x0 0x0.
2014-12-30T12:13:20.766Z cpu6:33519)ScsiDeviceIO: 2338:
Cmd(0x412e80449fc0) 0x89, CmdSN 0x234f from world 519051 to dev
"eui.0024f400d5020007" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0xe 0x0 0x0.
This patch fixes this issue, the test will run without a problem with the
fix (no IO errors, all the files are properly written to the directory).
Signed-off-by: Shahar Salzman <shahar.salzman@kaminario.com>
Reviewed-by: Eran Mann <eran.mann@kaminario.com>
[bvanassche: simplified implementation]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r5966 | bvassche | 2015-01-13 11:38:09 +0200 (Tue, 13 Jan 2015) | 5 lines
qla2x00t: Register for RSCNs in target mode
The QLogic firmware and qla2xxx do not register for RSCNs in
target-only mode, so do that explicitly.
........
r5967 | bvassche | 2015-01-14 10:06:12 +0200 (Wed, 14 Jan 2015) | 1 line
scst_targ: Use tabs instead of spaces for indentation (detected by checkpatch)
........
r5968 | bvassche | 2015-01-15 10:58:39 +0200 (Thu, 15 Jan 2015) | 4 lines
scst_targ: Avoid triggering a kernel panic if dev_user_parse() returns SCST_CMD_STATE_STOP
Reported-by: Ilan Steinberg <ilan.steinberg@kaminario.com>
........
r5969 | vlnb | 2015-01-16 03:21:10 +0200 (Fri, 16 Jan 2015) | 3 lines
Fix READ BUFFER and WRITE BUFFER commands
........
r5970 | vlnb | 2015-01-16 05:16:26 +0200 (Fri, 16 Jan 2015) | 3 lines
Follow up for r5968
........
r5971 | vlnb | 2015-01-16 05:53:29 +0200 (Fri, 16 Jan 2015) | 5 lines
Report during user devices unjam LUN NOT SUPPORTED sense
Reported-By: shahar.salzman <shahar.salzman@kaminario.com>
........
r5972 | bvassche | 2015-01-16 15:01:58 +0200 (Fri, 16 Jan 2015) | 2 lines
scst.spec.in: Rename variable kver into kversion
........
r5973 | bvassche | 2015-01-16 15:12:22 +0200 (Fri, 16 Jan 2015) | 2 lines
scst.spec.in: Pass kernel version via RPM-variable %{kversion} instead of shell variable ${KVER}
........
r5974 | bvassche | 2015-01-16 15:16:06 +0200 (Fri, 16 Jan 2015) | 6 lines
scst.spec.in: Determine version number correctly on a koji server
This patch has been tested on a koji build server and also on four
different RPM-based distributions (CentOS 7, Fedora 20, openSuSE 13.2
and SLES 11 SP3).
........
r5975 | bvassche | 2015-01-16 18:12:38 +0200 (Fri, 16 Jan 2015) | 1 line
scst.spec.in: Leave out kernel version from RPM name
........
r5976 | bvassche | 2015-01-16 18:20:10 +0200 (Fri, 16 Jan 2015) | 1 line
scst.spec.in: Add DKMS support
........
r5977 | vlnb | 2015-01-20 06:18:07 +0200 (Tue, 20 Jan 2015) | 3 lines
Revert r5964 as not needed
........
r5978 | vlnb | 2015-01-20 06:20:13 +0200 (Tue, 20 Jan 2015) | 3 lines
Revert r5963 as not needed
........
r5979 | bvassche | 2015-01-20 17:04:23 +0200 (Tue, 20 Jan 2015) | 13 lines
scst: Rework SCSI pass-through support for kernel versions >= 2.6.30
Changes in this patch:
- Rework the SCSI pass-through code such that for kernel versions
>= 2.6.30 the scst_exec_req_fifo patch is no longer needed.
- Modify the pass-through code such that blk_rq_append_bio() is only
called for kernel version 2.6.30. For later kernel versions
blk_make_request() is called instead.
- Rework scst_scsi_exec_async().
- Add debug tracing of SCSI pass-through result status.
- Add a lockdep_assert_held() call in scsi_end_async().
........
r5980 | bvassche | 2015-01-20 19:13:13 +0200 (Tue, 20 Jan 2015) | 1 line
nightly build: Update kernel versions
........
r5981 | vlnb | 2015-01-21 06:15:42 +0200 (Wed, 21 Jan 2015) | 3 lines
Follow up for r5979
........
r5982 | vlnb | 2015-01-21 06:20:53 +0200 (Wed, 21 Jan 2015) | 5 lines
Fix returning changeable values for caching mode page
Reported by Consus <consus@gmx.com>
........
r5983 | bvassche | 2015-01-21 15:11:56 +0200 (Wed, 21 Jan 2015) | 1 line
scst.h: Fix a sparse warning for kernels 2.6.29..2.6.31
........
r5984 | vlnb | 2015-01-22 07:03:17 +0200 (Thu, 22 Jan 2015) | 9 lines
[PATCH] scst_local: Fix bidirectional command support
scsi_setup_cmnd() sets sc_data_direction to DMA_TO_DEVICE for bidirectional
commands. Hence test SCpnt->request->next_rq instead of sc_data_direction
to figure out whether or not a command is bidirectional.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5985 | vlnb | 2015-01-22 07:06:45 +0200 (Thu, 22 Jan 2015) | 12 lines
[PATCH] scst_main: Suppress a checkpatch warning triggered by INIT_CACHEP{,_ALIGN}
Avoid that checkpatch v3.18 reports the following warning for these
two macros:
WARNING: Macros with flow control statements should be avoided
This patch does not change any functionality.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5986 | vlnb | 2015-01-22 07:09:17 +0200 (Thu, 22 Jan 2015) | 9 lines
scst_vdisk: Micro-optimize vdisk_caching_pg
This patch does not change any behavior but micro-optimizes
vdisk_caching_pg(). Declaring the array caching_pg[] const reduces
11 bytes from the assembler code of this function.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5987 | vlnb | 2015-01-22 07:10:42 +0200 (Thu, 22 Jan 2015) | 10 lines
scst: Suppress a smatch warning in vdisk_unmap_range()
Avoid that the static source code analysis tool 'smatch' reports
the following warning:
vdisk_unmap_range() warn: should 'blocks << cmd->dev->block_shift' be a 64 bit type?
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5988 | vlnb | 2015-01-22 07:13:59 +0200 (Thu, 22 Jan 2015) | 27 lines
scst_vdisk: Fix zero-copy read for tmpfs
For some filesystems, e.g. tmpfs, address_space.readpage is NULL.
Disable zero-copy reading for such filesystems. See also shmem_aops
in mm/shmem.c. See also inode_init_always() and empty_aops in fs/inode.c.
This patch avoids that the following call trace is triggered:
BUG: unable to handle kernel NULL pointer dereference at (null)
Call Trace:
[<ffffffffa0547d66>] prepare_read+0x106/0x1d0 [scst_vdisk]
[<ffffffffa0547f20>] fileio_alloc_data_buf+0xf0/0x330 [scst_vdisk]
[<ffffffffa046fc9b>] scst_prepare_space+0x9b/0x6e0 [scst]
[<ffffffffa047d4d5>] scst_process_active_cmd+0x545/0x840 [scst]
[<ffffffffa047dad2>] scst_cmd_init_done+0x302/0x5d0 [scst]
[<ffffffffa0563ab2>] scst_cmd_init_stage1_done.constprop.37+0x12/0x20 [iscsi_scst]
[<ffffffffa056a9ea>] scsi_cmnd_start+0x25a/0x550 [iscsi_scst]
[<ffffffffa056b4a8>] cmnd_rx_start+0x148/0x1a0 [iscsi_scst]
[<ffffffffa056e4f8>] process_read_io+0x3b8/0x800 [iscsi_scst]
[<ffffffffa056ea07>] scst_do_job_rd+0xc7/0x220 [iscsi_scst]
[<ffffffffa056efed>] istrd+0x16d/0x2e0 [iscsi_scst]
[<ffffffff81079efd>] kthread+0xed/0x110
[<ffffffff817227fc>] ret_from_fork+0x7c/0xb0
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5989 | vlnb | 2015-01-24 07:37:57 +0200 (Sat, 24 Jan 2015) | 5 lines
scst_local: Rework data direction detection code
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
........
r5990 | bvassche | 2015-01-26 13:32:32 +0200 (Mon, 26 Jan 2015) | 1 line
ib_srpt: Detect Mellanox OFED 2.3 correctly
........
r5991 | vlnb | 2015-01-28 07:07:46 +0200 (Wed, 28 Jan 2015) | 3 lines
Cleanups
........
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/iser@5993 d57e44dd-8a1f-0410-8b47-8ef2f437770f
This commit is contained in:
Yan Burman
+1
-1
@@ -8,7 +8,7 @@ RTFS = $(SRCS:.sgml=.rtf)
|
||||
|
||||
COMMAND=linuxdoc --backend=
|
||||
|
||||
all: txt pdf html
|
||||
all: pdf html
|
||||
|
||||
txt: $(TXTS)
|
||||
|
||||
|
||||
Binary file not shown.
|
Before Width: | Height: | Size: 28 KiB After Width: | Height: | Size: 42 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 38 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 42 KiB |
+1674
-1795
File diff suppressed because it is too large
Load Diff
@@ -1,335 +0,0 @@
|
||||
<!doctype linuxdoc system>
|
||||
|
||||
<article>
|
||||
|
||||
<title>
|
||||
SCST SGV cache description
|
||||
</title>
|
||||
|
||||
<author>
|
||||
<name>Vladislav Bolkhovitin</name>
|
||||
</author>
|
||||
|
||||
<date>Version 2.1.0</date>
|
||||
|
||||
<toc>
|
||||
|
||||
<sect>Introduction
|
||||
|
||||
<p>
|
||||
SCST SGV cache is a memory management subsystem in SCST. One can call it
|
||||
a "memory pool", but Linux kernel already have a mempool interface,
|
||||
which serves different purposes. SGV cache provides to SCST core, target
|
||||
drivers and backend dev handlers facilities to allocate, build and cache
|
||||
SG vectors for data buffers. The main advantage of it is the caching
|
||||
facility, when it doesn't free to the system each vector, which is not
|
||||
used anymore, but keeps it for a while (possibly indefinitely) to let it
|
||||
be reused by the next consecutive command. This allows to:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> Reduce commands processing latencies and, hence, improve performance;
|
||||
|
||||
<item> Make commands processing latencies predictable, which is essential
|
||||
for RT applications.
|
||||
|
||||
</itemize>
|
||||
|
||||
The freed SG vectors are kept by the SGV cache either for some (possibly
|
||||
indefinite) time, or, optionally, until the system needs more memory and
|
||||
asks to free some using the set_shrinker() interface. Also the SGV cache
|
||||
allows to:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> Cluster pages together. "Cluster" means merging adjacent pages in a
|
||||
single SG entry. It allows to have less SG entries in the resulting SG
|
||||
vector, hence improve performance handling it as well as allow to
|
||||
work with bigger buffers on hardware with limited SG capabilities.
|
||||
|
||||
<item> Set custom page allocator functions. For instance, scst_user device
|
||||
handler uses this facility to eliminate unneeded mapping/unmapping of
|
||||
user space pages and avoid unneeded IOCTL calls for buffers allocations.
|
||||
In fileio_tgt application, which uses a regular malloc() function to
|
||||
allocate data buffers, this facility allows ~30% less CPU load and
|
||||
considerable performance increase.
|
||||
|
||||
<item> Prevent each initiator or all initiators altogether to allocate too
|
||||
much memory and DoS the target. Consider 10 initiators, which can have
|
||||
access to 10 devices each. Any of them can queue up to 64 commands, each
|
||||
can transfer up to 1MB of data. So, all of them in a peak can allocate
|
||||
up to 10*10*64 = ~6.5GB of memory for data buffers. This amount must be
|
||||
limited somehow and the SGV cache performs this function.
|
||||
|
||||
</itemize>
|
||||
|
||||
<sect> Implementation
|
||||
|
||||
<p>
|
||||
From implementation POV the SGV cache is a simple extension of the kmem
|
||||
cache. It can work in 2 modes:
|
||||
|
||||
<enum>
|
||||
|
||||
<item> With fixed size buffers.
|
||||
|
||||
<item> With a set of power 2 size buffers. In this mode each SGV cache
|
||||
(struct sgv_pool) has SGV_POOL_ELEMENTS (11 currently) of kmem caches.
|
||||
Each of those kmem caches keeps SGV cache objects (struct sgv_pool_obj)
|
||||
corresponding to SG vectors with size of order X pages. For instance,
|
||||
request to allocate 4 pages will be served from kmem cache[2&rsqb, since the
|
||||
order of the of number of requested pages is 2. If later request to
|
||||
allocate 11KB comes, the same SG vector with 4 pages will be reused (see
|
||||
below). This mode is in average allows less memory overhead comparing
|
||||
with the fixed size buffers mode.
|
||||
|
||||
</enum>
|
||||
|
||||
Consider how the SGV cache works in the set of buffers mode. When a
|
||||
request to allocate new SG vector comes, sgv_pool_alloc() via
|
||||
sgv_get_obj() checks if there is already a cached vector with that
|
||||
order. If yes, then that vector will be reused and its length, if
|
||||
necessary, will be modified to match the requested size. In the above
|
||||
example request for 11KB buffer, 4 pages vector will be reused and
|
||||
modified using trans_tbl to contain 3 pages and the last entry will be
|
||||
modified to contain the requested length - 2*PAGE_SIZE. If there is no
|
||||
cached object, then a new sgv_pool_obj will be allocated from the
|
||||
corresponding kmem cache, chosen by the order of number of requested
|
||||
pages. Then that vector will be filled by pages and returned.
|
||||
|
||||
In the fixed size buffers mode the SGV cache works similarly, except
|
||||
that it always allocate buffer with the predefined fixed size. I.e.
|
||||
even for 4K request the whole buffer with predefined size, say, 1MB,
|
||||
will be used.
|
||||
|
||||
In both modes, if size of a request exceeds the maximum allowed for
|
||||
caching buffer size, the requested buffer will be allocated, but not
|
||||
cached.
|
||||
|
||||
Freed cached sgv_pool_obj objects are actually freed to the system
|
||||
either by the purge work, which is scheduled once in 60 seconds, or in
|
||||
sgv_shrink() called by system, when it's asking for memory.
|
||||
|
||||
<sect> Interface
|
||||
|
||||
<sect1> sgv_pool *sgv_pool_create()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
struct sgv_pool *sgv_pool_create(
|
||||
const char *name,
|
||||
enum sgv_clustering_types clustered, int single_alloc_pages,
|
||||
bool shared, int purge_interval)
|
||||
</verb>
|
||||
|
||||
This function creates and initializes an SGV cache. It has the following
|
||||
arguments:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> <bf/name/ - the name of the SGV cache
|
||||
|
||||
<item> <bf/clustered/ - sets type of the pages clustering. The type can be:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> <bf/sgv_no_clustering/ - no clustering performed.
|
||||
|
||||
<item> <bf/sgv_tail_clustering/ - a page will only be merged with the latest
|
||||
previously allocated page, so the order of pages in the SG will be
|
||||
preserved
|
||||
|
||||
<item> <bf/sgv_full_clustering/ - free merging of pages at any place in
|
||||
the SG is allowed. This mode usually provides the best merging
|
||||
rate.
|
||||
|
||||
</itemize>
|
||||
|
||||
<item> <bf/single_alloc_pages/ - if 0, then the SGV cache will work in the set of
|
||||
power 2 size buffers mode. If >0, then the SGV cache will work in the
|
||||
fixed size buffers mode. In this case single_alloc_pages sets the
|
||||
size of each buffer in pages.
|
||||
|
||||
<item> <bf/shared/ - sets if the SGV cache can be shared between devices or not.
|
||||
The cache sharing allowed only between devices created inside the same
|
||||
address space. If an SGV cache is shared, each subsequent call of
|
||||
sgv_pool_create() with the same cache name will not create a new cache,
|
||||
but instead return a reference to it.
|
||||
|
||||
<item> <bf/purge_interval/ - sets the cache purging interval. I.e. an SG buffer
|
||||
will be freed if it's unused for time t purge_interval <= t <
|
||||
2*purge_interval. If purge_interval is 0, then the default interval
|
||||
will be used (60 seconds). If purge_interval <0, then the automatic
|
||||
purging will be disabled. Shrinking by the system's demand will also
|
||||
be disabled.
|
||||
|
||||
</itemize>
|
||||
|
||||
Returns the resulting SGV cache or NULL in case of any error.
|
||||
|
||||
<sect1> void sgv_pool_del()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void sgv_pool_del(
|
||||
struct sgv_pool *pool)
|
||||
</verb>
|
||||
|
||||
This function deletes the corresponding SGV cache. If the cache is
|
||||
shared, it will decrease its reference counter. If the reference counter
|
||||
reaches 0, the cache will be destroyed.
|
||||
|
||||
<sect1> void sgv_pool_flush()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void sgv_pool_flush(
|
||||
struct sgv_pool *pool)
|
||||
</verb>
|
||||
|
||||
This function flushes, i.e. frees, all the cached entries in the SGV
|
||||
cache.
|
||||
|
||||
<sect1> void sgv_pool_set_allocator()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void sgv_pool_set_allocator(
|
||||
struct sgv_pool *pool,
|
||||
struct page *(*alloc_pages_fn)(struct scatterlist *sg, gfp_t gfp, void *priv),
|
||||
void (*free_pages_fn)(struct scatterlist *sg, int sg_count, void *priv));
|
||||
</verb>
|
||||
|
||||
This function allows to set for the SGV cache a custom pages allocator. For
|
||||
instance, scst_user uses such function to supply to the cache mapped from
|
||||
user space pages.
|
||||
|
||||
<bf/alloc_pages_fn()/ has the following parameters:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> <bf/sg/ - SG entry, to which the allocated page should be added.
|
||||
|
||||
<item> <bf/gfp/ - the allocation GFP flags
|
||||
|
||||
<item> <bf/priv/ - pointer to a private data supplied to sgv_pool_alloc()
|
||||
|
||||
</itemize>
|
||||
|
||||
This function should return the allocated page or NULL, if no page was
|
||||
allocated.
|
||||
|
||||
|
||||
<bf/free_pages_fn()/ has the following parameters:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> <bf/sg/ - SG vector to free
|
||||
|
||||
<item> <bf/sg_count/ - number of SG entries in the sg
|
||||
|
||||
<item> <bf/priv/ - pointer to a private data supplied to the
|
||||
corresponding sgv_pool_alloc()
|
||||
|
||||
</itemize>
|
||||
|
||||
<sect1> struct scatterlist *sgv_pool_alloc()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
struct scatterlist *sgv_pool_alloc(
|
||||
struct sgv_pool *pool,
|
||||
unsigned int size,
|
||||
gfp_t gfp_mask,
|
||||
int flags,
|
||||
int *count,
|
||||
struct sgv_pool_obj **sgv,
|
||||
struct scst_mem_lim *mem_lim,
|
||||
void *priv)
|
||||
</verb>
|
||||
|
||||
This function allocates an SG vector from the SGV cache. It has the
|
||||
following parameters:
|
||||
|
||||
<itemize>
|
||||
|
||||
<item> <bf/pool/ - the cache to alloc from
|
||||
|
||||
<item> <bf/size/ - size of the resulting SG vector in bytes
|
||||
|
||||
<item> <bf/gfp_mask/ - the allocation mask
|
||||
|
||||
<item> <bf/flags/ - the allocation flags. The following flags are possible and
|
||||
can be set using OR operation:
|
||||
|
||||
<enum>
|
||||
|
||||
<item> <bf/SGV_POOL_ALLOC_NO_CACHED/ - the SG vector must not be cached.
|
||||
|
||||
<item> <bf/SGV_POOL_NO_ALLOC_ON_CACHE_MISS/ - don't do an allocation on a
|
||||
cache miss.
|
||||
|
||||
<item> <bf/SGV_POOL_RETURN_OBJ_ON_ALLOC_FAIL/ - return an empty SGV object,
|
||||
i.e. without the SG vector, if the allocation can't be completed.
|
||||
For instance, because SGV_POOL_NO_ALLOC_ON_CACHE_MISS flag set.
|
||||
|
||||
</enum>
|
||||
|
||||
<item> <bf/count/ - the resulting count of SG entries in the resulting SG vector.
|
||||
|
||||
<item> <bf/sgv/ - the resulting SGV object. It should be used to free the
|
||||
resulting SG vector.
|
||||
|
||||
<item> <bf/mem_lim/ - memory limits, see below.
|
||||
|
||||
<item> <bf/priv/ - pointer to private for this allocation data. This pointer will
|
||||
be supplied to alloc_pages_fn() and free_pages_fn() and can be
|
||||
retrieved by sgv_get_priv().
|
||||
|
||||
</itemize>
|
||||
|
||||
This function returns pointer to the resulting SG vector or NULL in case
|
||||
of any error.
|
||||
|
||||
<sect1> void sgv_pool_free()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void sgv_pool_free(
|
||||
struct sgv_pool_obj *sgv,
|
||||
struct scst_mem_lim *mem_lim)
|
||||
</verb>
|
||||
|
||||
This function frees previously allocated SG vector, referenced by SGV
|
||||
cache object sgv.
|
||||
|
||||
<sect1> void *sgv_get_priv(struct sgv_pool_obj *sgv)
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void *sgv_get_priv(
|
||||
struct sgv_pool_obj *sgv)
|
||||
</verb>
|
||||
|
||||
This function allows to get the allocation private data for this SGV
|
||||
cache object sgv. The private data are set by sgv_pool_alloc().
|
||||
|
||||
<sect1> void scst_init_mem_lim()
|
||||
|
||||
<p>
|
||||
<verb>
|
||||
void scst_init_mem_lim(
|
||||
struct scst_mem_lim *mem_lim)
|
||||
</verb>
|
||||
|
||||
This function initializes memory limits structure mem_lim according to
|
||||
the current system configuration. This structure should be latter used
|
||||
to track and limit allocated by one or more SGV caches memory.
|
||||
|
||||
|
||||
<sect> Runtime information and statistics.
|
||||
|
||||
<p>
|
||||
Runtime information and statistics is available in /sys/kernel/scst_tgt/sgv.
|
||||
|
||||
</article>
|
||||
Reference in New Issue
Block a user