- 14 Dec, 2023 19 commits
-
-
Martin K. Petersen authored
Justin Tee <justintee8345@gmail.com> says: Update lpfc to revision 14.2.0.17 This patch set contains bug fixes for the VMID feature. The patches were cut against Martin's 6.8/scsi-queue tree. Link: https://lore.kernel.org/r/20231207224039.35466-1-justintee8345@gmail.comSigned-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Justin Tee authored
Update lpfc version to 14.2.0.17 Signed-off-by:
Justin Tee <justin.tee@broadcom.com> Link: https://lore.kernel.org/r/20231207224039.35466-5-justintee8345@gmail.comSigned-off-by:
-
Justin Tee authored
If priority tagging is set in the service parameters of a FLOGI cmpl, then we update the vmid_flag. In the current logic, if a follow up FLOGI cmpl updates its service parameters such that priority tagging is no longer set, then the vmid_flag ends up keeping stale data. Fix by ensuring we clear the vmid_flag member during lpfc_reinit_vmid, and check the priority tagging service parameter after reinitialization of the vmid data structures. Signed-off-by:
-
Justin Tee authored
After a follow up FDISC cmpl, an NPIV's VMID data structures are not updated. Fix by calling lpfc_reinit_vmid and copying the physical port's vmid_flag to the NPIV's vmid_flag in the NPIV registration cmpl code path. Signed-off-by:
-
Justin Tee authored
VMID driver support is a load time configuration setting. Thus, change sysfs attributes to read only. Signed-off-by:
-
Martin K. Petersen authored
Karan Tilak Kumar <kartilak@cisco.com> says: Hi Martin, reviewers, This cover letter describes the feature: add support for multiqueue (MQ) to fnic driver. Background: The Virtual Interface Card (VIC) firmware exposes several queues that can be configured for sending IOs and receiving IO responses. Unified Computing System Manager (UCSM) and Intersight Manager (IMM) allows users to configure the number of queues to be used for IOs. The number of IO queues to be used is stored in a configuration file by the VIC firmware. The fNIC driver reads the configuration file and sets the number of queues to be used. Previously, the driver was hard-coded to use only one queue. With this set of changes, the fNIC driver will configure itself to use multiple queues. This feature takes advantage of the block multiqueue layer to parallelize IOs being sent out of the VIC card. Here's a brief description of some of the salient patches: - vnic_scsi.h needs to be in sync with VIC firmware to be able to read the number of queues from the firmware config file. A patch has been created for this. - In an environment with many fnics (like we see in our customer environments), it is hard to distinguish which fnic is printing logs. Therefore, an fnic number has been included in the logs. - read the number of queues from the firmware config file. - include definitions in fnic.h to support multiqueue. - modify the interrupt service routines (ISRs) to read from the correct registers. The numbers that are used here come from discussions with the VIC firmware team. - track IO statistics for different queues. - remove usage of host_lock, and only use fnic_lock in the fnic driver. - use a hardware queue based spinlock to protect io_req. - replace the hard-coded zeroth queue with a hardware queue number. This presents a bulk of the changes. - modify the definition of fnic_queuecommand to accept multiqueue tags. - improve log messages, and indicate fnic number and multiqueue tags for effective debugging. Even though the patches have been made into a series, some patches are heavier than others. But, every effort has been made to keep the purpose of each patch as a single-purpose, and to compile cleanly. This patchset has been tested as a whole. Therefore, the tested-by fields have been added only to two patches in the set. All the individual patches compile cleanly. However, I've refrained from adding tested-by to most of the patches, so as to not mislead the reviewer/reader. A brief note on the unit tests: 1. Increase number of queues to 64. Load driver. Run IOs via Medusa. 12+ hour run successful. 2. Configure multipathing, and run link flaps on single link. IOs drop briefly, but pick up as expected. 3. Configure multipathing, and run link flaps on two links, with a 30 second delay in between. IOs drop briefly, but pick up as expected. Repeat the above tests with 1 queue and 32 queues. All tests were successful. Please consider this patch series for the next merge window. Link: https://lore.kernel.org/r/20231211173617.932990-1-kartilak@cisco.comSigned-off-by:
-
Karan Tilak Kumar authored
Increment driver version for multiqueue (MQ). Reviewed-by:
Sesidhar Baddela <sebaddel@cisco.com> Reviewed-by:
Arulprabhu Ponnusamy <arulponn@cisco.com> Signed-off-by:
Karan Tilak Kumar <kartilak@cisco.com> Link: https://lore.kernel.org/r/20231211173617.932990-14-kartilak@cisco.comSigned-off-by:
-
Karan Tilak Kumar authored
Improve existing logs by adding fnic number, hardware queue, tag, and mqtag in the prints. Add logs with the above elements for effective debugging. Reviewed-by:
Hannes Reinecke <hare@suse.de> Signed-off-by:
-
Karan Tilak Kumar authored
Implement support for MQ in fnic driver: The block multiqueue layer issues IO to the fnic driver with an MQ tag. Use the mqtag and derive a tag from it. Derive the hardware queue from the mqtag and use it in all paths. Modify queuecommand to handle mqtag. Replace wq and cq indices to support MQ. Replace the zeroth queue with a hardware queue. Implement spin locks on a per hardware queue basis. Replace io_lock with per hardware queue spinlock. Implement out of range tag checks. Allocate an io_req_table to track status of the io_req. Test the driver by building it, loading it, and configuring 64 queues in UCSM. Issue IOs using Medusa on multiple fnics. Enable/disable links to exercise the abort and clean up path. Reported-by:
kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202310300032.2awCqkfn-lkp@intel.com/Reviewed-by:
-
Karan Tilak Kumar authored
Set map_queues in the fnic_host_template to fnic_mq_map_queues_cpus. Define fnic_mq_map_queues_cpus to set cpu assignment to fnic queues. Refactor code in fnic_probe to enable vnic queues before scsi_add_host. Modify notify set to the correct index. Reviewed-by:
John Garry <john.g.garry@oracle.com> Signed-off-by:
-
Karan Tilak Kumar authored
Remove usage of host_lock. Replace with fnic_lock, where necessary. fnic does not use host_lock. fnic uses fnic_lock. Use fnic lock to protect fnic members in fnic_queuecommand. Add log messages in error cases. Reviewed-by:
-
Karan Tilak Kumar authored
Define an array to track IOs for the different queues, print the IO stats in fnic get stats data. Reviewed-by:
-
Karan Tilak Kumar authored
Modify interrupt service routines for INTx, MSI, and MSI-x to support multiqueue. Modify parameter list of fnic_wq_copy_cmpl_handler to take cq_index. Modify fnic_cleanup function to use the new function call of fnic_wq_copy_cmpl_handler. Refactor code to set interrupt mode to MSI-x to a new function. Add a new stat for intx_dummy. Reported-by:
-
Karan Tilak Kumar authored
Refactor and re-define values in fnic.h to implement multiqueue (MQ) functionality. VIC firmware allows fnic to create up to 64 copy workqueues. Update the copy workqueue max to 64. Modify the interrupt index to be in sync with the firmware to support MQ. Add irq number to the MSIX entry. Define a software workqueue table to track the status of io_reqs. Define a base for the copy workqueue. Reviewed-by:
-
Karan Tilak Kumar authored
Get the copy workqueue count and interrupt mode from the configuration. The config can be changed via UCSM. Add logs to print the interrupt mode and copy workqueue count. Add logs to print the vNIC resources. Reviewed-by:
-
Karan Tilak Kumar authored
Rename wq_copy to hw_copy_wq to accurately describe the copy workqueue. This will also help distinguish this data structure from software data structures that can be introduced. Reviewed-by:
-
Karan Tilak Kumar authored
Add link related log messages in fnic_fcs.c, Improve log message in fnic_fcs.c, Add log message in vnic_dev.c. Reviewed-by:
-
Karan Tilak Kumar authored
Add fnic_num in fnic.h to identify fnic in a multi-fnic environment. Increment and set the fnic number during driver load in fnic_probe. Replace the host number with fnic number in debugfs. Reviewed-by:
-
Karan Tilak Kumar authored
VIC firmware has updated definitions. Modify structure and definitions to sync with the latest VIC firmware. Reviewed-by:
-
- 08 Dec, 2023 1 commit
-
-
Arnd Bergmann authored
The newly introduced error messages get multiple format strings wrong: size_t must be printed using the %z modifier rather than %l and dma_addr_t must be printed by reference using the special %pad pointer type: drivers/scsi/mpi3mr/mpi3mr_app.c: In function 'mpi3mr_build_nvme_prp': include/linux/kern_levels.h:5:25: error: format '%llx' expects argument of type 'long long unsigned int', but argument 4 has type 'dma_addr_t' {aka 'unsigned int'} [-Werror=format=] drivers/scsi/mpi3mr/mpi3mr_app.c:949:25: note: in expansion of macro 'dprint_bsg_err' 949 | dprint_bsg_err(mrioc, | ^~~~~~~~~~~~~~ include/linux/kern_levels.h:5:25: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'size_t' {aka 'unsigned int'} [-Werror=format=] drivers/scsi/mpi3mr/mpi3mr_app.c:1112:41: note: in expansion of macro 'dprint_bsg_err' 1112 | dprint_bsg_err(mrioc, | ^~~~~~~~~~~~~~ Fixes: 9536af61 ("scsi: mpi3mr: Support for preallocation of SGL BSG data buffers part-3") Signed-off-by:Arnd Bergmann <arnd@arndb.de> Link: https://lore.kernel.org/r/20231207142813.935717-1-arnd@kernel.orgAcked-by:
Randy Dunlap <rdunlap@infradead.org> Tested-by: Randy Dunlap <rdunlap@infradead.org> # build-tested Signed-off-by:
-
- 06 Dec, 2023 20 commits
-
-
Martin K. Petersen authored
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> says: Hello, this series converts all drivers below drivers/scsi to struct platform_driver::remove_new(). See commit 5c5a7680 ("platform: Provide a remove callback that returns no value") for an extended explanation and the eventual goal. All conversations are trivial, because all .remove() callbacks returned zero unconditionally. Best regards Uwe Link: https://lore.kernel.org/r/cover.1701619134.git.u.kleine-koenig@pengutronix.deSigned-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Link: https://lore.kernel.org/r/1d385231c23c2a1e6e7dc1968eb111327386d1f6.1701619134.git.u.kleine-koenig@pengutronix.deSigned-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). In the error path emit an error message replacing the (less useful) message by the core. Apart from the improved error message there is no change in behaviour. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Martin K. Petersen authored
Two driver updates from Chandrakanth patil at Broadcom: scsi: mpi3mr: Update driver version to 8.5.1.0.0 scsi: mpi3mr: Support for preallocation of SGL BSG data buffers part-3 scsi: mpi3mr: Support for preallocation of SGL BSG data buffers part-2 scsi: mpi3mr: Support for preallocation of SGL BSG data buffers part-1 scsi: mpi3mr: Fetch correct device dev handle for status reply descriptor scsi: mpi3mr: Block PEL Enable Command on Controller Reset and Unrecoverable State scsi: mpi3mr: Clean up block devices post controller reset scsi: mpi3mr: Refresh sdev queue depth after controller reset Signed-off-by: -
Chandrakanth patil authored
Update driver version to 8.5.1.0.0 Signed-off-by:
Chandrakanth patil <chandrakanth.patil@broadcom.com> Link: https://lore.kernel.org/r/20231205191630.12201-5-chandrakanth.patil@broadcom.comSigned-off-by:
-
Chandrakanth patil authored
The driver acquires the required NVMe SGLs from the pre-allocated pool. Co-developed-by:
Sathya Prakash <sathya.prakash@broadcom.com> Signed-off-by:
-
Chandrakanth patil authored
The driver acquires the required SGLs from the pre-allocated pool. Co-developed-by:
-
Chandrakanth patil authored
The driver now supports SGLs for BSG data transfer. Upon loading, the driver pre-allocates SGLs in chunks of 8k, results in a total of 256 * 8k, equal to 2MB. These pre-allocated SGLs are reserved for handling BSG commands and are deallocated during driver unload. Co-developed-by:
-
