There's nastyness in the way we currently handle barriers (and
discards): They're effectively filesystem commands, but they get
processed as BLOCK_PC commands. Unfortunately BLOCK_PC commands are
taken by SCSI to be SG_IO commands and the issuer expects to see and
handle any returned errors, however trivial. This leads to a huge
problem, because the block layer doesn't expect this to happen and any
trivially retryable error on a barrier causes an immediate I/O error
to the filesystem.
The only real way to hack around this is to take the usual class of
offending errors (unit attentions) and make them all retryable in the
case of a REQ_HARDBARRIER. A correct fix would involve a rework of
the entire block and SCSI submit system, and so is out of scope for a
quick fix.
Cc: Hannes Reinecke <hare@suse.de>
Cc: Stable Tree <stable@kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Some arrays are giving I/O errors with ext3 filesystems when
SYNCHRONIZE_CACHE gets a UNIT_ATTENTION. What is happening is that
these commands have no retries, so the UNIT_ATTENTION causes the
barrier to fail. We should be enable retries here to clear any
transient error and allow the barrier to succeed.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Cc: Stable Tree <stable@kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
In the scsi_debug driver, the virtual_gb option ignores the
sector_size, implicitly assuming that is 512 bytes. So if
'virtual_gb=1 sector_size=4096' the result is an 8 GB (virtual) disk.
Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
Cc: Stable Tree <stable@kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
This fixes a regression introduced with this commit:
commit d3305f3407
Author: Mike Christie <michaelc@cs.wisc.edu>
Date: Thu Aug 20 15:10:58 2009 -0500
[SCSI] libiscsi: don't increment cmdsn if cmd is not sent
in 2.6.32.
When I moved the hdr->cmdsn after init_task, I added
a bug when header digests are used. The problem is
that the LLD may calculate the header digest in init_task,
so if we then set the cmdsn after the init_task call we
change what the digest will be calculated by the target.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Stable Tree <stable@kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
blk_abort_request() expects queue lock to be held by the caller.
Grab it before calling the function.
Lack of this synchronization led to infinite loop on corrupt
q->timeout_list.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@kernel.org
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Error handling on advansys_board_found is fixed, because it's buggy in
the case we have an ASC_NARROW_BOARD set and failure happens on
AscInitAsc1000Driver step: it was freeing items of wrong struct in the
dvc_var union of struct asc_board, which could lead to an oops in the
case we set some of the fields in struct of narrow board as code was
choosing to always freeing wide board fields, and not everything was
being freed/released properly.
Signed-off-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Fix operator precedence warning (from sparse), which results in the
data value always being 0:
drivers/scsi/qla4xxx/ql4_mbx.c:470:66: warning: right shift by bigger than source value
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Ravi Anand <ravi.anand@qlogic.com>
Cc: David C Somayajulu <david.somayajulu@qlogic.com>
Cc: Karen Higgins <karen.higgins@qlogic.com>
Cc: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
bnx2i driver has to wait and cleanup all iscsi endpoints before
returning from bnx2i_stop(). This is to make sure all chip resources
are freed before chip is reset.
As the requirements for 1G and 10G chipsets is different, added
per-device 'hba_shutdown_tmo' parameter to adapter structure
If the connections are not torn down by the daemon within this timeout
period, 'cid's will be leaked in 10G device. 1G devices are more
flexible and do not leak any resources because the whole chip ports
gets reset when MTU is changed or ethtool selftest is run
fixed a minor issue in bnx2i_ep_poll() which unnecessarily forced
error return code when driver timed out waiting for TCP connect
request to complete
Signed-off-by: Anil Veerabhadrappa <anilgv@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
ibmvscsi uses dma_unmap_single() for buffers mapped via
dma_map_sg(). It works however it's the API violation. The DMA debug
facility complains about it:
http://marc.info/?l=linux-scsi&m=127018555013151&w=2
Reported-by: Sachin Sant <sachinp@in.ibm.com>
Tested-by: Sachin Sant <sachinp@in.ibm.com>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
This was introduced back in 2005 at the very start of the git era by:
df0ae2497d
[SCSI] allow sleeping in ->eh_host_reset_handler()
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
When I made this patch:
b64e77f70b
it was to solve a problem where we were already on the waitqueue
becuase a connection problem/logout caused us to be on there
when we were cleaning up the session. If we happen to get
on queue for more normal reasons like their just does not happen
to be any send space at the same time we are closing the connection
we hit a race and get stuck in the wait.
We should not check if the waitqueue is active
because we could race with the network code. If
the network xmit code is just about to enter the
prepare to wait when we check for the waitqueue to
be active then we will miss each other and the
network code will fall into the wait and we will
not run wake_up.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Stanse found that one error path in qla24xx_bsg_timeout omits to
unlock ha->hardware_lock. Fix that.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Stanse found that two error paths in lpfc_bsg_rport_els_cmp and
lpfc_issue_ct_rsp_cmp omits to unlock phba->ct_ev_lock. It is
because they wrongly unlock phba->hbalock instead. Fix that.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Stanse found that one error path in mgmt_invalidate_icds omits to unlock
ctrl->mbox_lock. Fix that.
Added in 756d29c8c7 (Enable async mode for mcc rings)
where the spinlock was moved.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Jayamohan Kallickal <jayamohank@serverengines.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
adpt_i2o_delete_hba() calls kfree() so we have to save "pHba->next"
before calling it. Also inside adpt_i2o_delete_hba() itself, there
was another use after free bug which I fixed by moving the kfree()
down a line.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
The qla1280 driver acquires its firmware via udev. During boot the firmware
is located in the initrd. If, after root is mounted, the adapter needs to
reload firmware (host reset), the firmware load may fail if the root device
is on the adapter being reset. This patch modifies qla1280 to retain the
firmware loaded via the initial request_firmware() for use during error
recovery.
[jejb: fix up checkpatch issues]
Signed-off-by: Michael Reed <mdr@sgi.com>
Acked-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
On newer kernels users of advansys module are reporting system hang when
trying to load it without firmware files present. After looking closely
at description on https://qa.mandriva.com/show_bug.cgi?id=53220, I think
this is related to commit "[SCSI] advansys: use request_firmware". The
problem is that after switch to request_firmware, asc_dvc->err_code
isn't being set when firmware files aren't found or loading fails.
err_code is used by the driver to judge if there was a fatal error or
not, as can be seen for example on advansys_board_found, which will only
return -ENODEV when err_code is set. Because err_code isn't being set
when request_firmware fails, this is a change of behaviour of the code
before request_firmware addition, making it continue to load and it
fails later as the firmware wasn't really loaded.
Signed-off-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
The fix prevents application path from sending get-firmware-state mbx command
during as isp reset.
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
On specific platforms, MSI is unreliable on some of the QLA24xx chips, resulting
in fatal I/O errors under load, as reported in <http://bugs.debian.org/572322>
and by some RHEL customers.
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
The logic is changed to detect this condition based on following
1) both module parameters are off (ql2xmaxqueues and ql2xmultique_tag).
2) both module parameters are on (ql2xmaxqueues and ql2xmultique_tag).
3) The HBA does not support multi queue.
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Checks should only be done for NPIV-capable ISPs. Original code
could result in PORT_UPDATEs being missed on non-NPIV-capable
ISPs.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
We now enable/disable "Additional Receive Credits" in f/w based on nvram
parameter "Extended_BB_Credits" bit (i.e. Enhanced-Features: at offset 0x196).
This is applicable only for GEN2 CNAs.
Signed-off-by: Santosh Vernekar <santosh.vernekar@qlogic.com>
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
blk_end_request doesn't complete a bidi request
successfully
The unfinished request eventually triggers a panic in
timeout handling routine fc_bsg_job_timeout as
req->special is NULL
Use blk_end_request_all to end the request unconditionally
Signed-off-by: Lalit Chandivade <lalit.chandivade@qlogic.com>
Acked-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
We could be failing/stopping a connection due to libiscsi starting
recovery/cleanup, but the xmit path or scsi eh thread path
could be dropping the connection at the same time.
As a result the session->state gets set to failed instead of in
recovery. We end up not blocking the session
and so the replacement timeout never gets started and we only end up
failing the IO when scsi_softirq_done sees that the
cmd has been running for (cmd->allowed + 1) * rq->timeout secs.
We used to fail the IO right away so users are seeing a long
delay when using dm-multipath. This problem was added in
2.6.28.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: stable@kernel.org
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
This flag is not used, so best discarded.
Signed-off-by: NeilBrown <neilb@suse.de>
--
Hi Jens,
I came across this recently - these are the only two occurances
of "GENHD_FL_DRIVERFS" in the kernel, so it cannot be needed.
NeilBrown
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
There was a problem introduced in Jul 2008 by:
0e6f9d2708 pcmcia: use pcmcia_loop_config in scsi pcmcia drivers
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
The few lines below the kfree of hdr_buf may go to the label err_free
which will also free hdr_buf. The most straightforward solution seems to
be to just move the kfree of hdr_buf after these gotos.
A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)
// <smpl>
@r@
identifier E;
expression E1;
iterator I;
statement S;
@@
*kfree(E);
... when != E = E1
when != I(E,...) S
when != &E
*kfree(E);
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
qlogicpti driver registers its irq with a name containing slash.
This results in
[ 71.049735] WARNING: at fs/proc/generic.c:316 __xlate_proc_name+0xa8/0xb8()
[ 71.132815] name 'Qlogic/PTI'
because proc_mkdir with the name of the irq fails. Fix it by just
removing the slash from irq name. Discovered and tested on real hardware
(Sun Ultra 1).
Signed-off-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: David S. Miller <davem@davemloft.net>
The issue occur while deleting 60 virtual ports through the sys
interface /sys/class/fc_vports/vport-X/vport_delete. It happen while in
a mistake each request sent twice for the same vport. This interface is
asynchronous, entering the delete request into a work queue, allowing
more than one request to enter to the delete work queue. The result is a
NULL pointer. The first request already delete the vport, while the
second request got a pointer to the vport before the device destroyed.
Re-create vport later cause system freeze.
Solution: Check vport flags before entering the request to the work queue.
[jejb: fixed int<->long problem on spinlock flags variable]
Signed-off-by: Gal Rosen <galr@storwize.com>
Acked-by: James Smart <james.smart@emulex.com>
Cc: Stable Tree <stable@kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
A pointer to a probe callback is passed to the core via
platform_driver_register and so the function must not disappear when the
.init sections are discarded. Otherwise (if also having HOTPLUG=y)
unbinding and binding a device to the driver via sysfs will result in an
oops as does a device being registered late.
An alternative to this patch is using platform_driver_probe instead of
platform_driver_register plus removing the pointer to the probe function
from the struct platform_driver.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: Dmitri Vorobiev <dmitri.vorobiev@movial.fi>
Cc: Henrik Kretzschmar <henne@nachtwindheim.de>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: peter fuerst <post@pfrst.de>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Upgrade the upstream driver from 2.0.0.0 to 2.1.2.1.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
When IO is completed with underrun and with good SCSI status, check if
the transferred bytes against scsi_cmnd->underflow, which is set to
minimum number of bytes that must be transferred for this command, if
is less than required minimum, complete the IO with DID_ERROR.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
MS module did not invoke fdmi offline in all cases,
call fdmi offline when ms module receives a port offline,
so that fdmi offline is from one place in the ms module.
Make changes to handle 10G speed in the conversion routine.
Replaced the usage of bfa_adapter_attr_s struct with specific API's.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Modified the portstats get/clear logic for port physical/FCoE/QoS stats.
Added more stats to FC
Fixed some issues with FCoE stats collection.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
bfa_ioc_attr_s is a big structure and some times could cause
stack overflow if defined locally, so add specific APIs that
are needed to replace the use of ioc_attr local var.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Rename pport structures to fcport in BFA FCS, to resolve confusion
about the port structures in the firmware, and make sure the SG page
is setup correctly.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Currently BFA was not checking for IOC down condition when issuing
getstats/clearstats
Add check to see if IOC is operational, before issuing
getstats/clearstats.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Use dummy interrupt handlers till chip initialization is complete.
Install real interrupt handlers after chip initialization.
Also removed msix installation code in bfa_iocfc_init().
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Fixed the issue of not acknowledging the command queue
full-to-non-full interrupt. Implemented separate acknowledging
functions for different ASIC and interrupt mode.
Fixed the case of missing CPE interrupt by always processing the
pending requests in the completion path.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
ioc_recover failed to work in fcmode.
Fixed the code to initialize the ioc_regs.err_set during the
notify_hbfail.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Replace enum types with int and rearrange the fields to fix some
alignment issue.
Local var ioc_attr is causing the stack to overflow, so removed the
usage of the local ioc_attr var and now invoking an API to return the
ioc_type.
Fix some AEN issues.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Introduce a link notification state machine to handle next incoming
link events while the current event is being delivered to the driver.
When the event has been processed by the driver, the link notification
state machine will queue the next event (if there is any) to the
driver.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Made FCS authentication related changes to state machines and header
files.
Made changes in FCS state machines to handle the case when secret
string is NULL.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Changed PCI VPD to incorporate specific OEM vendors.
Added FCoE specific interrupt latency and delay params.
Added some variables needed by FIP 2.0.
Added some new logging and tracing definitions.
Added reserved members to make the structures (IOC, IOCFC) 64bit aligned.
Changed the module identifiers, as some files were moved.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Copy fpma MAC when requested by user space application. Added FPMA
mac address to the lport attributes structure.
Signed-off-by: Krishna Gudipati <kgudipat@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
James Bottomley