........
r3582 | vlnb | 2011-06-16 22:34:21 -0400 (Thu, 16 Jun 2011) | 3 lines
Docs update from Bart Van Assche <bvanassche@acm.org>
........
r3583 | vlnb | 2011-06-16 22:43:05 -0400 (Thu, 16 Jun 2011) | 16 lines
With NPIV initiators being destroyed and recreated, the loop_id and/or
the d_id of a virtual initiator may change. The current implementation not
always catches these changes, which leads to wrong virtual devices
being presented to the initiators. This can lead to i.e. corruption of LUN
content and multiple initiators writing to the same LUN, although the
LUN's access group only contains a single port.
Until a better solution is found, we schedule a session deletion for the
nexus when receiving an ELS port logout message. That way SCST
will re-initialize both loop_id and d_id when the initiator reconnects
before the actual deletion takes place. The performance impact is
minimal (few milliseconds) and currently necessary.
Signed-off-by: Jens-U. Mozdzen <jmozdzen@nde.ag>
........
r3584 | bvassche | 2011-06-17 06:58:46 -0400 (Fri, 17 Jun 2011) | 1 line
qla2x00t: Fix a checkpatch complaint about whitespace
........
r3585 | bvassche | 2011-06-17 13:53:06 -0400 (Fri, 17 Jun 2011) | 1 line
ib_srpt README: Add a paragraph about adding and removing LUNs dynamically
........
r3586 | vlnb | 2011-06-17 18:34:04 -0400 (Fri, 17 Jun 2011) | 3 lines
RHEL6 fix
........
r3587 | vlnb | 2011-06-17 18:35:12 -0400 (Fri, 17 Jun 2011) | 3 lines
Auto set thin provisioning for BLOCKIO devices
........
r3588 | vlnb | 2011-06-17 18:38:00 -0400 (Fri, 17 Jun 2011) | 3 lines
Unmap fixes
........
r3589 | bvassche | 2011-06-18 07:40:06 -0400 (Sat, 18 Jun 2011) | 1 line
ib_srpt: Update dynamic LUN documentation
........
r3590 | vlnb | 2011-06-18 17:00:14 -0400 (Sat, 18 Jun 2011) | 3 lines
Return max unmap LBA count in the blocks limits page as well
........
r3591 | vlnb | 2011-06-18 17:15:59 -0400 (Sat, 18 Jun 2011) | 15 lines
Several wait loops in the SCST source code are similar. Move the common code
into a new macro called wait_event_locked().
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
+
Micro-optimize wait_event_locked(). Switch from default_wake_function() to
autoremove_wake_function() (via DEFINE_WAIT()). The result is that
wait_queue_head_t.lock is locked once instead of twice for wakeup + wait queue
removal.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r3592 | bvassche | 2011-06-19 02:11:03 -0400 (Sun, 19 Jun 2011) | 1 line
Remove a recently introduced blank line
........
r3593 | bvassche | 2011-06-19 02:21:49 -0400 (Sun, 19 Jun 2011) | 1 line
scstadmin(sysfs): Make writing top-level non-key attributes work. Do not indent these attributes.
........
r3598 | bvassche | 2011-06-19 12:58:29 -0400 (Sun, 19 Jun 2011) | 1 line
scstadmin: Use 1 instead of TRUE where an integer exit code is returned
........
r3599 | bvassche | 2011-06-19 13:30:41 -0400 (Sun, 19 Jun 2011) | 1 line
scstadmin: Another return value consistency improvement
........
r3600 | bvassche | 2011-06-19 14:43:31 -0400 (Sun, 19 Jun 2011) | 1 line
scstadmin: Minor source code simplification
........
r3603 | vlnb | 2011-06-20 16:40:32 -0400 (Mon, 20 Jun 2011) | 3 lines
Fix crash on TP check, if file name is NULL
........
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3604 d57e44dd-8a1f-0410-8b47-8ef2f437770f
........
r3563 | vlnb | 2011-06-09 21:43:05 -0400 (Thu, 09 Jun 2011) | 3 lines
Web and docs updates
........
r3564 | vlnb | 2011-06-09 22:38:04 -0400 (Thu, 09 Jun 2011) | 36 lines
Currently there are three different init scripts inside the SCST project:
scst, iscsi-scst and qla2x00t. This makes it unnecessary hard for users to
run several target drivers simultaneously. Additionally, several variants of
these init scripts exist for different distributions. This patch unifies all
these init scripts into a single init script. Additionally, some new features
have been added:
- Kernel module parameters can now be specified per kernel module in
/etc/default/scst. An example:
ib_srpt_parameters="thread=1 srp_max_req_size=4200"
- When using the sysfs interface, compute the kernel module list from scst.conf.
Or, editing the SCST_MODULES variable is no longer necessary.
- When using the procfs interface, the SCST handler kernel module list is
computed from /etc/scst.conf. The list of target driver kernel modules must
now be configured in /etc/default/scst.
Other changes:
- scstadmin output is shown if scstadmin fails.
- Running make -C scstadmin install does no longer cause the scst script to
be started at the next reboot - it is better to leave this as an explicit step.
- Made the error message generated if the lsb-core package has not yet been
installed more clear.
- Use proper init script path on Slackware.
- make uninstall: run remove_initd under chroot if $(DESTDIR) is not empty.
- Fixed a bug in "make -C scstadmin uninstall": the path specified for removing
/etc/init.d/scst was wrong.
- The command /etc/init.d/scst restart will only try to start SCST if stopping
succeeded.
- Replaced "mkdir -m 755 -p $(DESTDIR)$(INITDIR)" by
"install -d $(DESTDIR)$(INITDIR)" because of uniformity.
These changes have been tested on Ubuntu 11.04, Fedora 15, CentOS 5.6,
Scientific Linux 6.0, openSUSE 11.4, SLES11 SP1 and Slackware 13.37.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r3565 | bvassche | 2011-06-10 13:56:56 -0400 (Fri, 10 Jun 2011) | 1 line
Follow-up for r3564: add sample /etc/default/scst file
........
r3566 | bvassche | 2011-06-10 14:03:06 -0400 (Fri, 10 Jun 2011) | 1 line
Slackware: Do not suggest the user to run a script to activate /etc/init.d/scst on boot because such a script does not exist on Slackware.
........
r3567 | bvassche | 2011-06-10 14:03:11 -0400 (Fri, 10 Jun 2011) | 8 lines
More /ec/init.d/scst tuning:
- Works now if the SCST modules have been compiled into the kernel (=Y) instead of being compiled as modules (=M).
- Rename DAEMONS into SCST_DAEMONS.
- Use the same name for loading and removing the iSCSI kernel module (iscsi_scst).
- Set the path after having read /lib/lsb/init-functions instead of before.
- Remove the path from the killall command.
- When parsing /etc/default/scst, handle both parameters with and without embedded spaces correctly.
........
r3568 | bvassche | 2011-06-11 06:26:25 -0400 (Sat, 11 Jun 2011) | 1 line
/etc/init.d/scst: Relax parsing of /etc/scst.conf. Accept sequences of blanks and tabs instead of only spaces when looking for device handler and target driver names.
........
r3569 | bvassche | 2011-06-11 11:55:49 -0400 (Sat, 11 Jun 2011) | 1 line
/etc/init.d/scst: Improved /etc/scst.conf parsing robustness for manually edited scst.conf files
........
r3570 | bvassche | 2011-06-13 08:19:01 -0400 (Mon, 13 Jun 2011) | 2 lines
ib_srpt: Restore timeout handler (reverts r3537).
........
r3571 | bvassche | 2011-06-13 08:22:33 -0400 (Mon, 13 Jun 2011) | 2 lines
ib_srpt: Make srpt_rdy_to_xfer() handle IB send queue overflow correctly.
........
r3572 | bvassche | 2011-06-13 08:26:51 -0400 (Mon, 13 Jun 2011) | 1 line
ib_srpt: Export RDMA channel state to sysfs
........
r3573 | bvassche | 2011-06-13 08:32:10 -0400 (Mon, 13 Jun 2011) | 1 line
ib_srpt: Disable IB completion notifcations after having received the LAST_WQE notification / show how many SCST commands are still in progress while waiting for session unregistration
........
r3574 | bvassche | 2011-06-13 08:35:41 -0400 (Mon, 13 Jun 2011) | 1 line
ib_srpt: Convert srpt_rdma_ch.sq_wr_avail from an atomic into a spinlock-protected integer
........
r3575 | bvassche | 2011-06-13 08:38:18 -0400 (Mon, 13 Jun 2011) | 1 line
ib_srpt: Internal change to the credit algorithm. No externally observable behavior is changed.
........
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3576 d57e44dd-8a1f-0410-8b47-8ef2f437770f
........
r3384 | bvassche | 2011-04-15 14:21:05 -0400 (Fri, 15 Apr 2011) | 1 line
Kernel version updates.
........
r3385 | vlnb | 2011-04-16 11:17:19 -0400 (Sat, 16 Apr 2011) | 3 lines
Minor fixes
........
r3386 | vlnb | 2011-04-16 11:24:43 -0400 (Sat, 16 Apr 2011) | 3 lines
Possibility to trace reservation conflicts added
........
r3387 | vlnb | 2011-04-16 11:30:33 -0400 (Sat, 16 Apr 2011) | 3 lines
Autogenerated SCST target names made deprecated
........
r3388 | vlnb | 2011-04-16 11:36:32 -0400 (Sat, 16 Apr 2011) | 3 lines
HCAs GUIDs should be used as target names
........
r3389 | vlnb | 2011-04-16 11:45:13 -0400 (Sat, 16 Apr 2011) | 3 lines
Minor cleanup
........
r3392 | vlnb | 2011-04-16 12:18:48 -0400 (Sat, 16 Apr 2011) | 2 lines
Version changed to 3.0
........
r3393 | bvassche | 2011-04-16 13:54:54 -0400 (Sat, 16 Apr 2011) | 1 line
Reverted because untested.
........
r3395 | vlnb | 2011-04-19 18:56:07 -0400 (Tue, 19 Apr 2011) | 3 lines
Docs update
........
r3397 | vlnb | 2011-04-20 18:51:51 -0400 (Wed, 20 Apr 2011) | 3 lines
HCAs GUIDs should be used as target names
........
r3399 | vlnb | 2011-04-22 15:10:08 -0400 (Fri, 22 Apr 2011) | 2 lines
Minor logging fix
........
r3400 | vlnb | 2011-04-23 17:13:50 -0400 (Sat, 23 Apr 2011) | 3 lines
Full buffers must always be processed, otherwise bad things could happen
........
r3401 | vlnb | 2011-04-23 20:12:01 -0400 (Sat, 23 Apr 2011) | 7 lines
Attached is a patch to make the names scst_get_full_buf and
scst_put_full_buf consistent with the naming convention. That is,
change them to scst_{get,put}_buf_full.
Signed-off-by: Richard Sharpe <realrichardsharpe@gmail.com>
........
r3402 | bvassche | 2011-04-24 02:40:52 -0400 (Sun, 24 Apr 2011) | 1 line
Removed a superfluous #include directive.
........
r3403 | bvassche | 2011-04-25 05:05:11 -0400 (Mon, 25 Apr 2011) | 1 line
Linux kernel 2.6.39 build fixes.
........
r3404 | bvassche | 2011-04-25 05:42:03 -0400 (Mon, 25 Apr 2011) | 1 line
RHEL / CentOS 5.6 build fix.
........
r3405 | bvassche | 2011-04-25 06:01:18 -0400 (Mon, 25 Apr 2011) | 1 line
Follow-up for r3403: builds again against kernel 2.6.27.
........
r3406 | bvassche | 2011-04-25 06:52:44 -0400 (Mon, 25 Apr 2011) | 1 line
Kernel version updates.
........
r3407 | bvassche | 2011-04-25 13:24:50 -0400 (Mon, 25 Apr 2011) | 1 line
Reverted r3404 because it's not a complete fix for RHEL / CentOS 5.6.
........
r3408 | bvassche | 2011-04-25 13:33:00 -0400 (Mon, 25 Apr 2011) | 1 line
RHEL/CentOS 5.6 build fix.
........
r3409 | bvassche | 2011-04-25 14:02:58 -0400 (Mon, 25 Apr 2011) | 1 line
Bug fix for RHEL 6.
........
r3410 | vlnb | 2011-04-25 19:16:53 -0400 (Mon, 25 Apr 2011) | 1 line
Revert r3403 and 3405, because they crash on scst_local with blockio initialization on 2.6.38
........
r3411 | vlnb | 2011-04-25 19:50:49 -0400 (Mon, 25 Apr 2011) | 3 lines
NOP-In/response timeouts improvements
........
r3412 | vlnb | 2011-04-25 20:01:29 -0400 (Mon, 25 Apr 2011) | 3 lines
Fix for crash on retry after double reset UA reported by Matteo Tescione <matteo@rmnet.it>
........
r3413 | vlnb | 2011-04-25 20:04:01 -0400 (Mon, 25 Apr 2011) | 3 lines
Correctly process being read cmd during abort
........
r3414 | vlnb | 2011-04-25 20:06:23 -0400 (Mon, 25 Apr 2011) | 3 lines
Rerference on the orig cmd should be taken during data out processing to correctly perform aborts
........
r3421 | vlnb | 2011-04-26 15:07:13 -0400 (Tue, 26 Apr 2011) | 5 lines
- More timers improvements
- Cleanups
........
r3422 | vlnb | 2011-04-27 16:46:17 -0400 (Wed, 27 Apr 2011) | 3 lines
Patch from Grigory Eykalis <Grigory.Eykalis@dothill.com> with some changes implementing pass-through functionality
........
r3423 | vlnb | 2011-04-27 19:29:46 -0400 (Wed, 27 Apr 2011) | 3 lines
Follow up commit for r3422
........
r3424 | vlnb | 2011-04-29 14:21:30 -0400 (Fri, 29 Apr 2011) | 3 lines
iSNS server should know only about enabled targets
........
r3425 | vlnb | 2011-04-29 15:54:13 -0400 (Fri, 29 Apr 2011) | 3 lines
Don't abort not yet received commands
........
r3428 | vlnb | 2011-04-29 17:53:19 -0400 (Fri, 29 Apr 2011) | 3 lines
Don't allow to create targets with the same names
........
r3429 | vlnb | 2011-04-29 19:50:59 -0400 (Fri, 29 Apr 2011) | 3 lines
Remove "unused variable" warning on old kernels
........
r3430 | bvassche | 2011-04-30 02:41:03 -0400 (Sat, 30 Apr 2011) | 1 line
Fix a compiler warning on pre-2.6.35 non-RHEL systems.
........
r3431 | bvassche | 2011-04-30 13:42:23 -0400 (Sat, 30 Apr 2011) | 2 lines
Bug fix: do not try to apply patch readahead-2.6.32.below11.patch on kernel 2.6.32.11 or any later 2.6.32 kernel.
........
r3432 | bvassche | 2011-05-01 03:43:18 -0400 (Sun, 01 May 2011) | 1 line
Removed the constant REPORT_DEVICE_IDENTIFIER from scst/include/scst_const.h because it is not used anywhere in the SCST source tree.
........
r3433 | bvassche | 2011-05-01 05:14:06 -0400 (Sun, 01 May 2011) | 1 line
Expand CONFIG_SCST_PROC too.
........
r3434 | bvassche | 2011-05-02 13:43:54 -0400 (Mon, 02 May 2011) | 2 lines
Added #include <linux/slab.h> for kmem_cache_free().
........
r3435 | vlnb | 2011-05-03 16:51:39 -0400 (Tue, 03 May 2011) | 3 lines
MultiP must be 1 and HiSup 0. Dejavu, I've already fixed it.
........
r3438 | vlnb | 2011-05-03 17:27:30 -0400 (Tue, 03 May 2011) | 2 lines
RHEL5/6 workaround
........
r3439 | vlnb | 2011-05-05 18:46:18 -0400 (Thu, 05 May 2011) | 3 lines
Fix incorrect errors reporting on deleting LUNs from groups in the proc interface
........
r3440 | vlnb | 2011-05-05 18:47:28 -0400 (Thu, 05 May 2011) | 3 lines
Cleanup
........
r3441 | bvassche | 2011-05-08 07:55:43 -0400 (Sun, 08 May 2011) | 1 line
scst.h: Avoid that including scst.h before <linux/slab.h> triggers a warning about a preprocessor symbol redefinition.
........
r3442 | bvassche | 2011-05-08 07:59:05 -0400 (Sun, 08 May 2011) | 1 line
Updated kernel versions.
........
r3443 | mgandalf | 2011-05-09 16:03:07 -0400 (Mon, 09 May 2011) | 1 line
- Enable all hardware targets before adding virtual ones.
........
r3444 | vlnb | 2011-05-09 19:07:23 -0400 (Mon, 09 May 2011) | 3 lines
Rework iSCSI aborts to make them more correct and remove recent workarounds (one more post-IET cleanup)
........
r3445 | vlnb | 2011-05-10 21:56:38 -0400 (Tue, 10 May 2011) | 3 lines
Let's make build state changes errors more clear
........
r3446 | vlnb | 2011-05-10 22:22:48 -0400 (Tue, 10 May 2011) | 22 lines
Add explicit ALUA support. On a setup with more than one target port
(sometimes called storage head) this patch makes it possible to tell the
initiator which port to use. Since that information can be specified per LUN,
with this patch it is not only possible to tell the initiator which path to
choose but also to balance the load over target ports. Target port attributes
can not only be defined for target ports that exist on the system on which
SCST is running but also for target ports present on other nodes of the same
HA setup. As far as I know this feature is unique among open source storage
target implementations.
Note: with this patch SCSI target port attributes can be defined for each LUN
defined on each SCST target individually. Such an SCST target is either a
physical entity (e.g. a HCA for the ib_srpt driver) or a logical entity (e.g.
an iSCSI target for the iscsi_scst driver). In the last case it is up to the
user to make sure that there is a one-to-one relationship between SCST target
and SCSI target port.
This patch is based on the specifications in the ANSI T10 spc4r30 document.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
........
r3447 | bvassche | 2011-05-11 07:00:01 -0400 (Wed, 11 May 2011) | 1 line
iscsi-scst/kernel/iscsi.c: fix a checkpatch complaint about whitespace.
........
r3448 | bvassche | 2011-05-11 07:01:32 -0400 (Wed, 11 May 2011) | 1 line
Update kernel versions.
........
r3449 | vlnb | 2011-05-11 12:28:41 -0400 (Wed, 11 May 2011) | 3 lines
Minor logging changes
........
r3450 | vlnb | 2011-05-11 21:31:01 -0400 (Wed, 11 May 2011) | 3 lines
TM handling fixes, part2, hopefully, final.
........
r3451 | bvassche | 2011-05-13 01:59:51 -0400 (Fri, 13 May 2011) | 1 line
Added missing scst_tg.o as reported by Novgorodov Igor.
........
r3452 | bvassche | 2011-05-13 02:13:52 -0400 (Fri, 13 May 2011) | 2 lines
iscsi-scst: Fixed a compilation error that only occurred with GENERATING_UPSTREAM_PATCH defined.
........
r3453 | bvassche | 2011-05-13 06:40:40 -0400 (Fri, 13 May 2011) | 1 line
scripts/run-regression-tests: echo "FAILED" if compilation failed.
........
r3454 | mgandalf | 2011-05-13 15:05:56 -0400 (Fri, 13 May 2011) | 1 line
- Add LUN 0 warning.
........
r3455 | mgandalf | 2011-05-13 15:21:16 -0400 (Fri, 13 May 2011) | 1 line
- Fix a small type.
........
r3456 | bvassche | 2011-05-16 14:25:33 -0400 (Mon, 16 May 2011) | 1 line
scstadmin/Makefile: verify whether the Perl module ExtUtils::MakeMaker has been installed before trying to install scstadmin itself. Based on a suggestion by Hiroyuki Sato.
........
r3457 | bvassche | 2011-05-17 15:28:56 -0400 (Tue, 17 May 2011) | 3 lines
Refined ALUA documentation:
- Added pointer to information about associating iSCSI targets with network interfaces.
- Added information about configuring ALUA multipath on a Linux initiator.
........
r3458 | vlnb | 2011-05-17 20:46:37 -0400 (Tue, 17 May 2011) | 3 lines
Fix for double scst_check_local_events() call crash on commands processing path
........
r3459 | vlnb | 2011-05-17 21:22:22 -0400 (Tue, 17 May 2011) | 3 lines
Regenerated scst_exec_req_fifo patches with fix for memory leak on errors processing path discovered and fixed by Steve Corbin <steve.corbin@virtualcomputer.com>
........
r3460 | vlnb | 2011-05-18 17:45:09 -0400 (Wed, 18 May 2011) | 3 lines
Some more TM cleanups/improvements
........
r3461 | vlnb | 2011-05-18 17:50:01 -0400 (Wed, 18 May 2011) | 3 lines
Blocked commands should not pass again through order checks
........
r3462 | vlnb | 2011-05-19 20:52:34 -0400 (Thu, 19 May 2011) | 3 lines
Update for 2.6.39
........
r3463 | bvassche | 2011-05-20 06:45:56 -0400 (Fri, 20 May 2011) | 1 line
Removed spaces in front of tabs.
........
r3464 | bvassche | 2011-05-20 07:13:07 -0400 (Fri, 20 May 2011) | 1 line
Made it possible for generate-kernel-patch to evaluate all LINUX_VERSION_CODE statements.
........
r3465 | bvassche | 2011-05-20 12:38:57 -0400 (Fri, 20 May 2011) | 1 line
Added remaining files necessary for supporting 2.6.39 (these are copies of the 2.6.38 equivalents).
........
r3466 | vlnb | 2011-05-20 15:09:45 -0400 (Fri, 20 May 2011) | 3 lines
On NEED_THREAD current function should be immediately finished
........
r3471 | bvassche | 2011-05-22 05:52:53 -0400 (Sun, 22 May 2011) | 1 line
Added kernel version 2.6.39.
........
r3472 | vlnb | 2011-05-23 15:24:41 -0400 (Mon, 23 May 2011) | 3 lines
Docs about how to migrate from the deprecated ib_srpt_target_X layout updated
........
r3474 | vlnb | 2011-05-24 20:25:11 -0400 (Tue, 24 May 2011) | 2 lines
ChangeLogs updates
........
r3476 | bvassche | 2011-05-25 06:16:58 -0400 (Wed, 25 May 2011) | 1 line
Update kernel versions.
........
r3477 | bvassche | 2011-05-25 11:12:40 -0400 (Wed, 25 May 2011) | 1 line
Reverted r3476 because the relevant kernel patches haven't been mirrored yet.
........
r3478 | vlnb | 2011-05-25 15:59:38 -0400 (Wed, 25 May 2011) | 3 lines
Fix crash if /var/lib/scst/pr not found
........
r3480 | bvassche | 2011-05-26 14:26:41 -0400 (Thu, 26 May 2011) | 1 line
scstadmin/Makefile robustness improvements for the "install" target
........
r3483 | bvassche | 2011-05-27 06:15:41 -0400 (Fri, 27 May 2011) | 1 line
Reverted r3477.
........
r3484 | bvassche | 2011-05-27 06:33:59 -0400 (Fri, 27 May 2011) | 1 line
Removed a superfluous line.
........
r3488 | bvassche | 2011-05-28 06:37:46 -0400 (Sat, 28 May 2011) | 1 line
iSCSI-SCST builds again on RHEL 5.
........
r3489 | bvassche | 2011-05-28 06:59:27 -0400 (Sat, 28 May 2011) | 4 lines
scst_local/procfs bug fix: don't include <scst_const.h> before <scst.h> or
SCST_VERSION_STRING will be incorrect because it gets defined before
CONFIG_SCST_PROC is defined.
........
r3490 | bvassche | 2011-05-28 07:00:35 -0400 (Sat, 28 May 2011) | 4 lines
scst_local/procfs bug fix: don't include <scst_const.h> before <scst.h> or
SCST_VERSION_STRING will be incorrect because it gets defined before
CONFIG_SCST_PROC is defined.
........
r3492 | bvassche | 2011-06-01 02:32:47 -0400 (Wed, 01 Jun 2011) | 1 line
ib_srpt: Coding style fixes
........
r3493 | bvassche | 2011-06-01 02:39:49 -0400 (Wed, 01 Jun 2011) | 1 line
scst_local: Rearranged a preprocessor directive such that generate-kernel-patch can expand it.
........
r3494 | bvassche | 2011-06-01 02:57:50 -0400 (Wed, 01 Jun 2011) | 1 line
scst_sysfs.c: Convert spaces to tabs.
........
r3495 | bvassche | 2011-06-01 02:59:40 -0400 (Wed, 01 Jun 2011) | 1 line
Move spaces in long #warning/#error lines to the end of the previous line to shut up checkpatch.
........
r3496 | bvassche | 2011-06-01 03:00:48 -0400 (Wed, 01 Jun 2011) | 1 line
qla2x00t: get rid of checkpatch complaints about spaces at the start of a line.
........
r3497 | bvassche | 2011-06-01 03:19:12 -0400 (Wed, 01 Jun 2011) | 1 line
iscsi-scst: get rid of two "suspect code indent for conditional statements" checkpatch complaints.
........
r3498 | bvassche | 2011-06-01 04:03:32 -0400 (Wed, 01 Jun 2011) | 1 line
iSCSI-SCST: only define __packed for those kernels versions that do not have a definition for __packed.
........
r3499 | bvassche | 2011-06-01 04:04:56 -0400 (Wed, 01 Jun 2011) | 1 line
scst_local: get rid of a checkpatch complaint for the procfs build.
........
r3500 | bvassche | 2011-06-01 06:40:07 -0400 (Wed, 01 Jun 2011) | 1 line
Use __packed instead of __attribute__((packed)).
........
r3503 | bvassche | 2011-06-01 12:55:15 -0400 (Wed, 01 Jun 2011) | 1 line
Corrected t10_dev_id and usn documentation.
........
r3504 | vlnb | 2011-06-01 12:58:06 -0400 (Wed, 01 Jun 2011) | 3 lines
- Let's always do extraclean on Makefiles modifications, because they can affect data layout in header files, which could not be detected by all target drivers/dev handlers makes
- Cleanups
........
r3506 | bvassche | 2011-06-01 13:26:09 -0400 (Wed, 01 Jun 2011) | 1 line
ib_srpt: removed thread start/stop messages.
........
r3507 | vlnb | 2011-06-01 14:22:52 -0400 (Wed, 01 Jun 2011) | 3 lines
Remove TAGS file on extraclean
........
r3510 | vlnb | 2011-06-01 19:14:04 -0400 (Wed, 01 Jun 2011) | 3 lines
Fix possible return 0 in sysfs write to threads_pool_type attribute functions
........
r3512 | bvassche | 2011-06-02 06:04:22 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Fixed srp_max_rsp_size kernel module parameter description.
........
r3513 | bvassche | 2011-06-02 09:27:10 -0400 (Thu, 02 Jun 2011) | 1 line
scst: Update ALUA documentation
........
r3514 | bvassche | 2011-06-02 10:27:17 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Eliminate srpt_rdma_ch.wait_queue
........
r3515 | bvassche | 2011-06-02 10:35:14 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Eliminate srpt_rdma_ch.processing_compl.
........
r3516 | bvassche | 2011-06-02 10:54:51 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: follow-up for r3515.
........
r3517 | bvassche | 2011-06-02 10:55:54 -0400 (Thu, 02 Jun 2011) | 1 line
Only #include <linux/version.h> if necessary.
........
r3518 | bvassche | 2011-06-02 12:04:13 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Builds again on RHEL 5.x
........
r3519 | bvassche | 2011-06-02 14:21:39 -0400 (Thu, 02 Jun 2011) | 1 line
Made ib_srpt build with OFED 1.5.3.1.
........
r3520 | bvassche | 2011-06-02 15:12:29 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Another follow-up for r3515.
........
r3521 | bvassche | 2011-06-02 15:13:06 -0400 (Thu, 02 Jun 2011) | 1 line
ib_srpt: Removed a superfluous if-statement. This is a micro-optimization.
........
r3522 | bvassche | 2011-06-03 03:03:31 -0400 (Fri, 03 Jun 2011) | 1 line
ib_srpt: Silence a compiler warning.
........
r3523 | vlnb | 2011-06-03 16:29:48 -0400 (Fri, 03 Jun 2011) | 3 lines
Cleanup
........
r3525 | bvassche | 2011-06-04 08:24:41 -0400 (Sat, 04 Jun 2011) | 1 line
Reverted r3521.
........
r3526 | bvassche | 2011-06-04 08:33:57 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Remove a superfluous break statement
........
r3527 | bvassche | 2011-06-04 08:36:03 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Introduce the function srpt_set_ch_state_to_draining()
........
r3528 | bvassche | 2011-06-04 08:37:09 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Revert r3522.
........
r3529 | bvassche | 2011-06-04 09:45:16 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Update test plan
........
r3530 | bvassche | 2011-06-04 10:06:12 -0400 (Sat, 04 Jun 2011) | 2 lines
ib_srpt: Eliminate a kmalloc(GFP_ATOMIC) call
........
r3531 | bvassche | 2011-06-04 10:27:45 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Rework channel closing
........
r3532 | bvassche | 2011-06-04 10:32:22 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Fix a (hard to trigger) race with regard to SRP login.
........
r3533 | bvassche | 2011-06-04 10:37:56 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Reduce number of context switches during I/O for mode thread=1.
........
r3534 | bvassche | 2011-06-04 13:48:14 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Add more TRACE_ENTRY()/TRACE_EXIT() macros
........
r3535 | bvassche | 2011-06-04 13:50:19 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Remove "thread" kernel module parameter and keep thread=1 behavior - the other two modes could cause soft lockup complaints
........
r3536 | bvassche | 2011-06-04 13:50:54 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Relax I/O context locking
........
r3537 | bvassche | 2011-06-04 13:51:46 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Remove srpt_pending_cmd_timeout() because it is never invoked
........
r3538 | bvassche | 2011-06-04 14:10:45 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Build fix for non-RHEL pre-2.6.25 kernels.
........
r3539 | bvassche | 2011-06-04 14:17:01 -0400 (Sat, 04 Jun 2011) | 1 line
ib_srpt: Convert a PRINT_INFO() statement into a TRACE_DBG() statement
........
r3540 | bvassche | 2011-06-05 05:56:58 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Remove an unused variable
........
r3541 | bvassche | 2011-06-05 05:57:35 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Fix a recently introduced leak in an error path
........
r3542 | bvassche | 2011-06-05 05:58:03 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Add a BUG_ON() statement
........
r3543 | bvassche | 2011-06-05 05:58:32 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Fix a spelling error
........
r3544 | bvassche | 2011-06-05 05:59:17 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Simplify IB CM code
........
r3545 | bvassche | 2011-06-05 06:00:57 -0400 (Sun, 05 Jun 2011) | 1 line
ib_srpt: Fix a recently introduced RHEL 5 / CentOS 5 deadlock triggered by initiator logout
........
r3546 | vlnb | 2011-06-06 17:04:04 -0400 (Mon, 06 Jun 2011) | 4 lines
Set data_out_in_data_receiving for preliminary completed commands as
well to prevent releasing them too early in iscsi_check_tm_data_wait_timeouts()
........
r3549 | vlnb | 2011-06-06 17:32:16 -0400 (Mon, 06 Jun 2011) | 6 lines
- Return no residual if expected values not set instead of BUG_ON(),
because this is valid situation for some preliminary comleted commands.
- Cleanup
........
r3552 | vlnb | 2011-06-07 14:08:58 -0400 (Tue, 07 Jun 2011) | 3 lines
Modified patch from Bart Van Assche <bvanassche@acm.org> fixing errors reporting
........
r3557 | bvassche | 2011-06-08 14:05:44 -0400 (Wed, 08 Jun 2011) | 1 line
ib_srpt: Update test instructions.
........
r3558 | bvassche | 2011-06-08 14:13:30 -0400 (Wed, 08 Jun 2011) | 1 line
ib_srpt: Use a spinlock to protect ch->req_lim and ch->req_lim_delta changes instead of using atomics. Reduces the time spent in the ib_srpt_compl thread with a little bit more than 10%.
........
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3561 d57e44dd-8a1f-0410-8b47-8ef2f437770f
- Return no residual if expected values not set instead of BUG_ON(),
because this is valid situation for some preliminary comleted commands.
- Cleanup
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3551 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Set data_out_in_data_receiving for preliminary completed commands as
well to prevent releasing them too early in iscsi_check_tm_data_wait_timeouts()
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3548 d57e44dd-8a1f-0410-8b47-8ef2f437770f
- Let's always do extraclean on Makefiles modifications, because they can affect data layout in header files, which could not be detected by all target drivers/dev handlers makes
- Cleanups
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3505 d57e44dd-8a1f-0410-8b47-8ef2f437770f
- Don't allow to create targets with the same names
- Remove "unused variable" warning on old kernels
- Fix a compiler warning on pre-2.6.35 non-RHEL systems.
- Bug fix: do not try to apply patch readahead-2.6.32.below11.patch on kernel 2.6.32.11 or any later 2.6.32 kernel.
- Removed the constant REPORT_DEVICE_IDENTIFIER from scst/include/scst_const.h because it is not used anywhere in the SCST source tree.
- Expand CONFIG_SCST_PROC too.
- Added #include <linux/slab.h> for kmem_cache_free().
- MultiP must be 1 and HiSup 0
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3437 d57e44dd-8a1f-0410-8b47-8ef2f437770f
- More timers improvements and cleanups
- iSNS server should know only about enabled targets
- Don't abort not yet received commands
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3427 d57e44dd-8a1f-0410-8b47-8ef2f437770f
- Full buffers must always be processed, otherwise bad things could happen
- Make scst_get_full_buf and scst_put_full_buf consistent with the naming convention
git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/2.1.0.x@3415 d57e44dd-8a1f-0410-8b47-8ef2f437770f
