- 14 Jan, 2023 13 commits
-
-
Asutosh Das authored
Modify completion path APIs and add completion queue entry. Co-developed-by:
Can Guo <quic_cang@quicinc.com> Signed-off-by:
Asutosh Das <quic_asutoshd@quicinc.com> Reviewed-by:
Bart Van Assche <bvanassche@acm.org> Reviewed-by:
Manivannan Sadhasivam <mani@kernel.org> Reviewed-by:
Stanley Chu <stanley.chu@mediatek.com> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Asutosh Das authored
Add support to find the hardware queue on which the request would be queued. Since the very first queue is to serve device commands, an offset of 1 is added to the index of the hardware queue. Co-developed-by:
-
Asutosh Das authored
Add support to send commands using multiple submission queues in MCQ mode. Modify the functions that use ufshcd_send_command(). Co-developed-by:
-
Asutosh Das authored
Enable shared tags for MCQ. For UFS, this should not have a huge performance impact. It however simplifies the MCQ implementation and reuses most of the existing code in the issue and completion path. Also add multiple queue mapping to map_queue(). Co-developed-by:
-
Asutosh Das authored
Runtime and operation registers are defined per Submission and Completion queue. The location of these registers is not defined in the spec; meaning the offsets and stride may vary for different HC vendors. Establish the stride, base address, and doorbell address offsets from vendor host driver and program it. Co-developed-by:
-
Asutosh Das authored
To read the bqueuedepth, the device descriptor is fetched in Single Doorbell Mode. This allocated memory may not be enough for MCQ mode because the number of tags supported in MCQ mode may be larger than in SDB mode. Hence, release the memory allocated in SDB mode and allocate memory for MCQ mode operation. Define the UFS hardware queue and Completion Queue Entry. Co-developed-by:
-
Asutosh Das authored
The UFS device defines the supported queuedepth by bqueuedepth which has a max value of 256. The HC defines MAC (Max Active Commands) that defines the max number of commands that in flight to the UFS device. Calculate and configure the nutrs based on both these values. Co-developed-by:
-
Asutosh Das authored
Define the MCQ resources and add support to ioremap the resource regions. Co-developed-by:
-
Asutosh Das authored
Multi-circular queue (MCQ) has been added in UFSHC v4.0 standard in addition to the Single Doorbell mode. The MCQ mode supports multiple submission and completion queues. Add support to allocate and configure the queues. Add module parameters support to configure the queues. Co-developed-by:
-
Asutosh Das authored
If MCQ support is present, enabling it after MCQ support has been configured would require reallocating tags and memory. It would also free up the already allocated memory in Single Doorbell Mode. So defer invoking scsi_add_host() until MCQ is configured. Co-developed-by:
-
Asutosh Das authored
Add support to check for MCQ capability in the UFSHC. Add a module parameter to disable MCQ if needed. Co-developed-by:
-
Asutosh Das authored
Task Tag is limited to 8 bits and this restricts the number of active I/Os to 255. In multi-circular queue mode, this may not be enough. The specification provides EXT_IID which can be used to increase the number of I/Os if the UFS device and UFSHC support it. This patch adds support to probe for EXT_IID support in UFS device and UFSHC. Co-developed-by:
Avri Altman <avri.altman@wdc.com> Reviewed-by:
-
Kees Cook authored
Zero-length arrays are deprecated[1]. Replace struct fc_bsg_host_vendor_reply's "vendor_rsp" 0-length array with a flexible array. Detected with GCC 13, using -fstrict-flex-arrays=3: drivers/scsi/qla2xxx/qla_isr.c: In function 'qla25xx_process_bidir_status_iocb.isra': drivers/scsi/qla2xxx/qla_isr.c:3117:54: warning: array subscript 0 is outside array bounds of '__u32[0]' {aka 'unsigned int[]'} [-Warray-bounds=] 3117 | bsg_reply->reply_data.vendor_reply.vendor_rsp[0] = rval; | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~ In file included from drivers/scsi/qla2xxx/qla_def.h:34, from drivers/scsi/qla2xxx/qla_isr.c:6: include/uapi/scsi/scsi_bsg_fc.h:219:15: note: while referencing 'vendor_rsp' 219 | __u32 vendor_rsp[0]; | ^~~~~~~~~~ [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org> Signed-off-by:Kees Cook <keescook@chromium.org> Reviewed-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Link: https://lore.kernel.org/r/20230105233042.never.913-kees@kernel.orgSigned-off-by:
-
- 12 Jan, 2023 27 commits
-
-
Kees Cook authored
One-element arrays (and multi-element arrays being treated as dynamically sized) are deprecated[1] and are being replaced with flexible array members in support of the ongoing efforts to tighten the FORTIFY_SOURCE routines on memcpy(), correctly instrument array indexing with UBSAN_BOUNDS, and to globally enable -fstrict-flex-arrays=3. Replace one-element arrays with flexible-array member in struct mvumi_msg_frame, struct mvumi_rsp_frame, and struct mvumi_hs_header, adjusting the explicit sizing calculations at the same time. This results in no functional differences in binary output. An explicit add is now folded into the size calculation: │ mov 0x1070(%r14),%eax │ - add $0x4,%eax │ - movabs $0xfffffffdc,%rbx │ + movabs $0xfffffffe0,%rbx │ add %rax,%rbx [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays Cc: "James E.J. Bottomley" <jejb@linux.ibm.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Kees Cook authored
One-element arrays (and multi-element arrays being treated as dynamically sized) are deprecated[1] and are being replaced with flexible array members in support of the ongoing efforts to tighten the FORTIFY_SOURCE routines on memcpy(), correctly instrument array indexing with UBSAN_BOUNDS, and to globally enable -fstrict-flex-arrays=3. Replace one-element arrays with flexible-array member in TW_Ioctl_Buf_Apache and TW_Param_Apache, adjusting the explicit sizing calculations at the same time. This results in no differences in binary output. [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays Cc: Adam Radford <aradford@gmail.com> Cc: "James E.J. Bottomley" <jejb@linux.ibm.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Martin K. Petersen authored
Justin Tee <justintee8345@gmail.com> says: Update lpfc to revision 14.2.0.10 This patch set contains fixes for bugs, kernel test robot, and introduces new attention type event handling. Link: https://lore.kernel.org/r/20230109233317.54737-1-justintee8345@gmail.comSigned-off-by:
-
Justin Tee authored
Update copyrights to 2023 for files modified in the 14.2.0.10 patch set. Signed-off-by:
Justin Tee <justin.tee@broadcom.com> Signed-off-by:
-
Justin Tee authored
Update lpfc version to 14.2.0.10 Signed-off-by:
-
Justin Tee authored
Define new FC Link ACQE with new attention types 0x8 (Link Activation Failure) and 0x9 (Link Reset Protocol Event). Both attention types are meant to be informational-only type ACQEs with no action required. 0x8 is reported for diagnostic purposes, while 0x9 is posted during a normal link up transition when activating BB Credit Recovery feature. As such, modify lpfc_sli4_async_fc_evt() logic to log the attention types according to its severity and early return when informational-only attention types are encountered. Signed-off-by:
-
Justin Tee authored
After enabling VMID, an issue LIP test was erasing fabric switch VMID information. Introduce a lpfc_reinit_vmid() routine, which reinitializes all VMID data structures upon FLOGI completion in fabric topology. Signed-off-by:
-
Justin Tee authored
During the sysfs firmware write process, a use-after-free read warning is logged from the lpfc_wr_object() routine: BUG: KFENCE: use-after-free read in lpfc_wr_object+0x235/0x310 [lpfc] Use-after-free read at 0x0000000000cf164d (in kfence-#111): lpfc_wr_object+0x235/0x310 [lpfc] lpfc_write_firmware.cold+0x206/0x30d [lpfc] lpfc_sli4_request_firmware_update+0xa6/0x100 [lpfc] lpfc_request_firmware_upgrade_store+0x66/0xb0 [lpfc] kernfs_fop_write_iter+0x121/0x1b0 new_sync_write+0x11c/0x1b0 vfs_write+0x1ef/0x280 ksys_write+0x5f/0xe0 do_syscall_64+0x59/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd The driver accessed wr_object pointer data, which was initialized into mailbox payload memory, after the mailbox object was released back to the mailbox pool. Fix by moving the mailbox free calls to the end of the routine ensuring that we don't reference internal mailbox memory after release. Signed-off-by:
-
Justin Tee authored
In a large SAN testing configuration, frequent target port toggle tests are occasionally resulting in missing lun path rediscoveries. An outstanding PRLI can be inflight when a target RSCN dissappearance occurs, causing the driver to retry PRLIs using invalid rpi contexts. Fix by verifying that an ndlp's state was not restarted from PRLI_ISSUE due to an intermediate RSCN. If not in a valid state, early exit PRLI completion handling. The last follow up RSCN indicating target reappearance retriggers PLOGI/PRLI with a valid rpi context and is expected to succeed in LUN path rediscovery. Signed-off-by:
-
Justin Tee authored
With faulty cables in PT2PT topology, an unintentional ndlp double kref decrement can occur. If a FLOGI request is outstanding before the link goes down, the missing FLOGI_ACC causes an F_Port ndlp to remain in the UNUSED state. During link down, lpfc_cleanup_rpis() is called and decrements an ndlp kref. Additionally, when the driver later decides to abort the FLOGI, the FLOGI completion handler decrements the ndlp kref a second time. Remove duplicate clean up logic in lpfc_cleanup_rpis() because the updated FLOGI completion handler already handles the ndlp kref decrement. Signed-off-by:
-
Justin Tee authored
The disable_vport() path calls the discovery state machine on all ndlps, puts them into NPR state, and then calls lpfc_cleanup_rpis() with the remove flag set. This unintentionally decrements an ndlp's kref twice and can result in premature release of an ndlp because lpfc_dev_loss_tmo_handler() triggers clean up of the ndlp again later. Remove redundant code in disable_vport() that sets all the ndlps to NPR, and change the call to lpfc_cleanup_rpis() to not remove the ndlps. lpfc_dev_loss_tmo_handler() will handle final removal of the ndlps. Signed-off-by:
-
Justin Tee authored
During I/O, the following warning message occasionally appears: DMA-API: lpfc 0000:04:00.0: mapping sg segment longer than device claims to support [len=131072] [max=65536] The HBA is capable of supporting 131,072 bytes, so notify DMA layer via the dma_set_max_seg_size() API during hba initialization. Signed-off-by:
-
Justin Tee authored
The local variables called curr_data are incremented, but not actually used for anything so they are removed. The return value of lpfc_sli4_poll_eq is not used anywhere and is not called outside of lpfc_sli.c. Thus, its declaration is removed from lpfc_crtn.h Also, lpfc_sli4_poll_eq's path argument is not used in the routine so it is removed along with corresponding macros. Signed-off-by:
-
Justin Tee authored
The kernel test robot pointed out non-NULL terminated string possibilities when using strncpy() in lpfc_xcvr_data_show() routine. Although we manually set the NULL character after strncpy(), strncpy() usage is outdated. Replace all strncpy() usages with the preferred strscpy() API. Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
-
Justin Tee authored
The kernel test robot detected inconsistent indentations for an if statement block in the lpfc_xcvr_data_show() routine. This patch reduces the extraneous tabs used for the if statement block in question. Reported-by:
-
Martin K. Petersen authored
Nilesh Javali <njavali@marvell.com> says: Please apply the qla2xxx driver enhancements to the SCSI tree at your earliest convenience. Link: https://lore.kernel.org/r/20221222043933.2825-1-njavali@marvell.comSigned-off-by:
-
Nilesh Javali authored
Signed-off-by:
Nilesh Javali <njavali@marvell.com> Signed-off-by:
-
Shreyas Deodhar authored
In current I/O path, Tx and Rx may not be processed on same CPU. This may lead to thrashing and optimum performance may not be achieved. Pick qpair such that Tx and Rx are processed on same CPU. Signed-off-by:
Shreyas Deodhar <sdeodhar@marvell.com> Signed-off-by:
-
Quinn Tran authored
clang warning: drivers/scsi/qla2xxx/qla_edif_bsg.h:93:12: warning: field remote_pid within 'struct app_pinfo_req' is less aligned than 'port_id_t' and is usually due to 'struct app_pinfo_req' being packed, which can lead to unaligned accesses [-Wunaligned-access] port_id_t remote_pid; ^ 2 warnings generated. Remove u32 field in remote_pid to silence warning. Reported-by:
Quinn Tran <qutran@marvell.com> Signed-off-by:
-
Quinn Tran authored
For edif, each I/O requires a secondary buffer to carry the FCP cmnd. During high traffic time, these buffers are cached in the qpair. As traffic dies down, these buffers will be trimmed as needed. If traffic is reduced to none over 2 consecutive intervals, then these buffers will be further trimmed. Free FCP cmnd buffers to reduce memory usage during slow I/O time. Signed-off-by:
-
Quinn Tran authored
For N2N, qla2x00_wait_for_sess_deletion call flushes a session which accidentally clear the scan_flag and thus prevents re-login to occur and causes session to stall. Use session delete to avoid the accidental clearing of scan_flag. Signed-off-by:
-
Quinn Tran authored
User experienced performance dip on measuring IOPS while EDIF enabled. During I/O time, driver uses dma_pool_zalloc() call to allocate a chunk of memory. This call contains a lock behind the scene which contribute to lock contention. Save the allocated memory for reuse and avoid the lock. Signed-off-by:
Himanshu Madhani <himanshu.madhani@oracle.com> Signed-off-by:
-
Quinn Tran authored
There is no functional change in this patch. VP map resource is renamed and relocated so it is not viewed as just a target mode resource. Signed-off-by:
-
Quinn Tran authored
Remove stale/unused code (GNN ID). Signed-off-by:
-
Quinn Tran authored
Remove stale unused code for GPNID. Signed-off-by:
-
Quinn Tran authored
Removing drport field and FCPORT_UPDATE_NEEDED signals. Signed-off-by:
-
Dan Carpenter authored
This code accidentally returns success instead of -ENOMEM. Fixes: 7cc7646b ("scsi: libsas: Factor out sas_ata_add_dev()") Link: https://lore.kernel.org/r/Y7asLxzVwQ56G+ya@kiliSigned-off-by:
Dan Carpenter <error27@gmail.com> Reviewed-by:
Jason Yan <yanaijie@huawei.com> Signed-off-by:
-
