- 10 Apr, 2008 2 commits
-
-
James Smart authored
PCI Parity and EEH handling Fixes: - Under a PCI Data Parity Error, remove a completion routine callback that was on a command that we had already failed and released. - Under PCI parity error, we were not reinstalling the interrupt handler in the slot_reset callback, so we never became functional again. Signed-off-by:
James Smart <james.smart@emulex.com> Signed-off-by:
James Bottomley <James.Bottomley@HansenPartnership.com>
-
James Smart authored
Multiple Discovery Fixes: - Fix race on discovery due to link events coinciding with vport_delete. - Use NLP_FABRIC state to filter out switch-based pseudo initiators that reuse the same WWNs. - Correct erroneous setting of DID=0 in lpfc_matchdid() - Correct extra reference count that was in the lookup path for the remoteid from an unsolicited ELS. - Correct double-free bug in els abort path. - Correct FDMI server discovery logic for switch that return a WWN of 0. - Fix bugs in ndlp mgmt when a node changes address - Correct bug that did not delete RSCNs for vports upon link transitions - Fix "0216 Link event during NS query" error which pops up when vports are swapped to different switch ports. - Add sanity checks on ndlp structures - Fix devloss log message to dump WWN correctly - Hold off mgmt commands that were interferring with discovery mailbox cmds - Remove unnecessary FC_ESTABLISH_LINK logic. - Correct some race conditions in the worker thread, resulting in devloss: - Clear the work_port_events field before handling the work port events - Clear the deferred ring event before handling a deferred ring event - Hold the hba lock when waking up the work thread - Send an acc for the rscn even when we aren't going to handle it - Fix locking behavior that was not properly protecting the ACTIVE flag, thus allowing mailbox command order to shift. Signed-off-by:
-
- 07 Apr, 2008 38 commits
-
-
Andrew Vasquez authored
Signed-off-by:
Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by:
-
Seokmann Ju authored
Upon having configuration changes on vports only, the driver handles SCR regardless physical port state and, in turn, it results mailbox error as below: Mar 20 11:24:20 dl585 kernel: qla2x00_mailbox_command(9): **** FAILED. mbx0=4005, mbx1=1, mbx2=8100, cmd=70 **** With the changes, driver checks physical port loop_state and make sure the port is ready to take commands. Signed-off-by:
Seokmann Ju <seokmann.ju@qlogic.com> Signed-off-by:
-
Seokmann Ju authored
Changes are added to the driver so that it can behave properly upon having asynchronous events, for example, addition of target devices to the VPs. Signed-off-by:
-
Seokmann Ju authored
As there is no actual cable connection on vports, made change so that the driver checks DFLG_NO_CABLE against ha->device_flags only for physical port. Signed-off-by:
-
Seokmann Ju authored
There were several places where referencing ha structure of virtual ports for resources. Among those refereces, certain fields are get up-to-dated only on ha structure of physical port. Signed-off-by:
-
Andrew Vasquez authored
The member is not needed as there's no non-scatter-gather list I/Os submitted by the upper-layers. Signed-off-by:
-
Harihara Kadayam authored
Signed-off-by:
Ravi Anand <ravi.anand@qlogic.com> Additional cleanups and Signed-off-by:
-
Ravi Anand authored
Signed-off-by:
-
Andrew Vasquez authored
Now that infrastructure is present within the midlayer and there is a clear distinction between what is expected from a device and target reset, convert the current device-reset codes to a target-reset, and add codes to perform a proper device-reset (LUN reset). In the process of adding reset support, collapse and consolidate large sections of mailbox-command (TMF issuance) codes, generalize the two 'wait-for-commands-to-complete' functions, and add a generic-reset routine for use by midlayer reset functions. Signed-off-by:
-
Andrew Vasquez authored
Some switches return 0x09 (Command not supported) as the reason code for GPSC failure. Check for this code, and disable additional GPSC queries if found. Signed-off-by:
-
Andrew Vasquez authored
The Flash Descriptor Table (FDT) present on many recent HBAs encodes flash accessing characteristics of the flash-part used on the HBA. Use this information during flash manipulation (writes) rather than using specific hard-coded values based on queried manufacturer and device IDs. Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
-
Andrew Vasquez authored
Strip unused (DEBUG-ONLY) enabled functions, inlines, useless wrappers, and unused DPC flags from the code. Another step in the migration towards a cleaner (less-crusty) driver. Signed-off-by:
-
Andrew Vasquez authored
Recent ISPs have a region within FLASH which acts as a repository for the logging of serious hardware and software failures. Currently, the region is large enough to support up to 255 entries. Signed-off-by:
-
Andrew Vasquez authored
Supported events include LIP, LIP reset, RSCN, link up, and link down. To support AEN (and additional forthcoming features), we also introduce a simple deferred-work construct to manage events which require a non-atomic sleeping-capable context. This work-list is processed as part of the driver's standard DPC routine. Signed-off-by:
-
Andrew Vasquez authored
There's no need to use the heavier (albiet safer) *_irq[save|restore]() locking primitives within the driver's interrupt handlers, interrupts are guaranteed to be non-reentrant. Use lightweight spin_lock() and spin_unlock() primitives while acquiring the hardware_lock. Signed-off-by:
-
Andrew Vasquez authored
Rather than using a duplicate/cached value stored in the driver's internal fcport structure. Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
-
Andrew Vasquez authored
Use generic FC transport #defines for speed values. Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
-
Andrew Vasquez authored
Signed-off-by:
-
James Bottomley authored
These are completely wrong because both outX and writeX do an automatic reverse of their arguments if necessary, so having an extra cpu_to_leX gives us the wrong ordering on BE platforms again. Acked-by:
Mark Salyzyn <Mark_Salyzyn@adaptec.com> Signed-off-by:
-
James Bottomley authored
Two functions in include/scsi/sas_ata.h don't have static inlines leading to problems if they're built in: On Thu, 2008-04-03 at 14:06 +0200, Toralf Förster wrote: > drivers/scsi/mvsas.o: In function `sas_ata_init_host_and_port': > mvsas.c:(.text+0x0): multiple definition of `sas_ata_init_host_and_port' > drivers/scsi/libsas/built-in.o:(.text+0x37f4): first defined here > drivers/scsi/mvsas.o: In function `sas_ata_task_abort': > mvsas.c:(.text+0x7): multiple definition of `sas_ata_task_abort' > drivers/scsi/libsas/built-in.o:(.text+0x37fb): first defined here > make[2]: *** [drivers/scsi/built-in.o] Error 1 > make[1]: *** [drivers/scsi] Error 2 > make: *** [drivers] Error 2 Add the correct static inline modifiers. Tested-by:
Toralf Förster <toralf.foerster@gmx.de> Signed-off-by:
-
James Bottomley authored
Every current transport class calls transport_container_release but ignores the return value. This is catastrophic if it returns an error because the containers are part of a global list and the next action of almost every transport class is to free the memory used by the container. Fix this by making transport_container_release a void, but making it BUG if attribute_container_release returns an error ... this catches the root cause of a system panic much earlier. If we don't do this, we get an eventual BUG when the attribute container list notices the corruption caused by the freed memory it's still referencing. Also made attribute_container_release __must_check as a reminder. Cc: Greg KH <greg@kroah.com> Signed-off-by: -
James Bottomley authored
Currently, if the barrier command fails, the error return isn't seen by the block layer and it proceeds on regardless. The problem is that SCSI always returns no error for REQ_TYPE_BLOCK_PC ... it expects the submitter to pick the errors out of req->errors, which the block barrier functions don't do. Since it appears that the way SG_IO and scsi_execute_request() work they discard the block error return and always use req->errors, the best fix for this is to have the SCSI layer return an error to block if one actually occurred (this also allows us to filter out spurious errors, like deferred sense). This patch is a bug fix that will need backporting to stable, but it's also quite a big change and in need of testing, so we'll incubate in the main kernel tree and backport at the -rc2 or so stage if no problems turn up. Acked-by:
Jens Axboe <jens.axboe@oracle.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Signed-off-by:
-
Martin Peschke authored
Add missing docbook-comments for functions forming zfcp's internal trace API. Signed-off-by:
Martin Peschke <mp3@de.ibm.com> Signed-off-by:
Christof Schmitt <christof.schmitt@de.ibm.com> Signed-off-by:
-
Martin Peschke authored
Remove line breaks that do not conform to coding style. Signed-off-by:
-
Martin Peschke authored
Saving on line breaks, improving readability, by shortening excessive function names and identifiers, by simplifying some functions call chains, and by simplifying nesting of some data structure. Signed-off-by:
-
Christof Schmitt authored
Unclutter the global zfcp_def.h header. Move everything required to call into the debug feature to a new header file. Signed-off-by:
-
Martin Peschke authored
Simplify usage of zfcp_dbf_tag() and calling functions. Signed-off-by:
-
Martin Peschke authored
Simplify usage of output function for hex dumps. Signed-off-by:
-
Martin Peschke authored
Remove obsolete output function. Signed-off-by:
-
Martin Peschke authored
Improve readability of code by using more convenient output function. Signed-off-by:
-
Martin Peschke authored
Change zfcp_dbf_timestamp() so that it just calculates timespec from timestamp. First step to be able to rip this code out of zfcp. Besides, this change makes it easier to rip out old-style debug view functions. Signed-off-by:
-
Adrian Bunk authored
- static functions in .c files shouldn't be marked inline - make needlessly global code static - remove the unused aic94xx_seq.c:asd_unpause_lseq() - #if 0 other unused code [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:
Adrian Bunk <bunk@kernel.org> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Harvey Harrison authored
Replace the global err array with ch_err. drivers/scsi/ch.c:271:6: warning: symbol 'err' shadows an earlier one drivers/scsi/ch.c:116:3: originally declared here Replace the temporary cmd buffer with ch_err to avoid shadowing the cmd function parameter. drivers/scsi/ch.c:724:11: warning: symbol 'cmd' shadows an earlier one drivers/scsi/ch.c:596:20: originally declared here Signed-off-by:
Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by:
-
Adrian Bunk authored
This patch makes the needlessly global scsi_end_bidi_request() static. Signed-off-by:
Boaz Harrosh <bharrosh@panasas.com> Signed-off-by:
-
James Bottomley authored
header.context is a 64 bit field, but it's deliberately split into context and context_hi32. Thus cpu_to_le64 assignments are wrong on this. Replace them with the correct settings of both the low and high words. Cc: HighPoint Linux Team <linux@highpoint-tech.com> Signed-off-by:
-
