- 02 Mar, 2022 3 commits
-
-
Christoph Hellwig authored
Now that each scsi_request is backed by a scsi_cmnd, there is no need to indirect the CDB storage. Change all submitters of SCSI passthrough requests to store the CDB information directly in the scsi_cmnd, and while doing so allocate the full 32 bytes that cover all Linux supported SCSI hosts instead of requiring dynamic allocation for > 16 byte CDBs. On 64-bit systems this does not change the size of the scsi_cmnd at all, while on 32-bit systems it slightly increases it for now, but that increase will be made up by the removal of the remaining scsi_request fields. Link: https://lore.kernel.org/r/20220224175552.988286-4-hch@lst.deReviewed-by:
Bart Van Assche <bvanassche@acm.org> Reviewed-by:
John Garry <john.garry@huawei.com> Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Christoph Hellwig authored
Replace the big fat memset that requires saving and restoring various fields with just initializing those fields that need initialization. All the clearing to 0 is moved to scsi_prepare_cmd() as scsi_ioctl_reset() alreadly uses kzalloc() to allocate a pre-zeroed command. This is still conservative and can probably be optimized further. Link: https://lore.kernel.org/r/20220224175552.988286-3-hch@lst.deReviewed-by:
-
Christoph Hellwig authored
Copy directly from the se_cmd CDB to the one in the scsi_request. This temporarily limits the pscsi backend to supporting only up to 16 byte CDBs, but this restriction will be lifted later in this series. Link: https://lore.kernel.org/r/20220224175552.988286-2-hch@lst.deSigned-off-by:
-
- 28 Feb, 2022 16 commits
-
-
John Garry authored
Function queue_work() returns a bool, so use a bool to hold this value for the return code from callers, which should make the code a tiny bit more clear. Also take this opportunity to condense the code of the those callers, such as sas_queue_work(), as suggested by Damien. Link: https://lore.kernel.org/r/1645786656-221630-3-git-send-email-john.garry@huawei.comReviewed-by:
Damien Le Moal <damien.lemoal@opensource.wdc.com> Signed-off-by:
-
John Garry authored
Nobody checks the return codes, so make them return void. Indeed, if the LLDD cannot send an event, nothing much can be done in the LLDD about it. Also remove prototype for sas_notify_phy_event() in sas_internal.h, which should not be there. Link: https://lore.kernel.org/r/1645786656-221630-2-git-send-email-john.garry@huawei.comReviewed-by:
Xiang Chen <chenxiang66@hisilicon.com> Reviewed-by:
-
Xingui Yang authored
In case of SSP underflow allow the response frame IU to be examined for setting the response stat value rather than always setting SAS_DATA_UNDERRUN. This will mean that we call sas_ssp_task_response() in those scenarios and may send sense data to upper layer. Such a condition would be for bad blocks were we just reporting an underflow error to upper layer, but now the sense data will tell immediately that the media is faulty. Link: https://lore.kernel.org/r/1645703489-87194-7-git-send-email-john.garry@huawei.comSigned-off-by:
Xingui Yang <yangxingui@huawei.com> Signed-off-by:
Qi Liu <liuqi115@huawei.com> Signed-off-by:
-
Xiang Chen authored
Add a file operation for "cnt" file under bist directory, so users can only read "cnt" or clear "cnt" to zero, but cannot randomly modify. Link: https://lore.kernel.org/r/1645703489-87194-6-git-send-email-john.garry@huawei.comSigned-off-by:
-
Qi Liu authored
To avoid doubt, rename the error labels to indicate the action they will take. Link: https://lore.kernel.org/r/1645703489-87194-5-git-send-email-john.garry@huawei.comSigned-off-by:
-
Qi Liu authored
If the driver probe fails to request the channel IRQ or fatal IRQ, the driver will free the IRQ vectors before freeing the IRQs in free_irq(), and this will cause a kernel BUG like this: ------------[ cut here ]------------ kernel BUG at drivers/pci/msi.c:369! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Call trace: free_msi_irqs+0x118/0x13c pci_disable_msi+0xfc/0x120 pci_free_irq_vectors+0x24/0x3c hisi_sas_v3_probe+0x360/0x9d0 [hisi_sas_v3_hw] local_pci_probe+0x44/0xb0 work_for_cpu_fn+0x20/0x34 process_one_work+0x1d0/0x340 worker_thread+0x2e0/0x460 kthread+0x180/0x190 ret_from_fork+0x10/0x20 ---[ end trace b88990335b610c11 ]--- So we use devm_add_action() to control the order in which we free the vectors. Link: https://lore.kernel.org/r/1645703489-87194-4-git-send-email-john.garry@huawei.comSigned-off-by:
-
Xiang Chen authored
The time of phyup not only depends on the controller but also the type of disk connected. As an example, from experience, for some SATA disks the amount of time from reset/power-on to receive the D2H FIS for phyup can take upto and more than 10s sometimes. According to the specification of some SATA disks such as ST14000NM0018, the max time from power-on to ready is 30s. Based on this the current timeout of phyup at 2s which is not enough. So set the value as HISI_SAS_WAIT_PHYUP_TIMEOUT (30s) in hisi_sas_control_phy(). For v3 hw there is a pre-existing workaround for a HW bug, being that we issue a link reset when the OOB occurs but the phyup does not. The current phyup timeout is HISI_SAS_WAIT_PHYUP_TIMEOUT. So if this does occur from when issuing a phy enable or similar via hisi_sas_control_phy(), the subsequent HW workaround linkreset processing calls hisi_sas_control_phy(), but this will pend the original phy reset timing out, so it is safe. Link: https://lore.kernel.org/r/1645703489-87194-3-git-send-email-john.garry@huawei.comSigned-off-by:
-
Xiang Chen authored
Currently the permission of parameter prot_mask is 0x0, which means that the member does not appear in sysfs. Change it as other module parameters to 0444 for world-readable. [mkp: s/v3/v2/] Link: https://lore.kernel.org/r/1645703489-87194-2-git-send-email-john.garry@huawei.com Fixes: d6a9000b ("scsi: hisi_sas: Add support for DIF feature for v2 hw") Reported-by:
Yihang Li <liyihang6@hisilicon.com> Signed-off-by:
-
Changcheng Deng authored
Remove unneeded variable used to store return value. Link: https://lore.kernel.org/r/20220224090735.1967816-1-deng.changcheng@zte.com.cnReported-by:
Zeal Robot <zealci@zte.com.cn> Signed-off-by:
Changcheng Deng <deng.changcheng@zte.com.cn> Signed-off-by:
-
Randy Dunlap authored
__setup() handlers should return 1 if the command line option is handled and 0 if not (or maybe never return 0; doing so just pollutes init's environment with strings that are not init arguments/parameters). Return 1 from aha152x_setup() to indicate that the boot option has been handled. Link: lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru Link: https://lore.kernel.org/r/20220223000623.5920-1-rdunlap@infradead.org Cc: "Juergen E. Fischer" <fischer@norbit.de> Cc: "James E.J. Bottomley" <jejb@linux.ibm.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Reported-by:
Igor Zhbanov <i.zhbanov@omprussia.ru> Signed-off-by:
Randy Dunlap <rdunlap@infradead.org> Signed-off-by:
-
Ajish Koshy authored
Firmware expects host driver to clear scratchpad rsvd 0 register after non-fatal error is found. This is done when firmware raises fatal error interrupt and indicates non-fatal error. At this point firmware updates scratchpad rsvd 0 register with non-fatal error value. Here host has to clear the register after reading it during non-fatal errors. Rename: - MSGU_HOST_SCRATCH_PAD_6 to MSGU_SCRATCH_PAD_RSVD_0 - MSGU_HOST_SCRATCH_PAD_7 to MSGU_SCRATCH_PAD_RSVD_1 Link: https://lore.kernel.org/r/20220222092618.108198-1-Ajish.Koshy@microchip.comReviewed-by:
Jack Wang <jinpu.wang@ionos.com> Signed-off-by:
Ajish Koshy <Ajish.Koshy@microchip.com> Signed-off-by:
Viswas G <Viswas.G@microchip.com> Signed-off-by:
-
Finn Thain authored
This driver doesn't use SCp.ptr to save a SCSI command data pointer which means "scsi pointer" is a complete misnomer here. Only a few members of struct scsi_pointer are used and the rest waste memory. Avoid the "struct foo { struct bar; };" silliness. Link: https://lore.kernel.org/r/3529a59873a7de8455a27af2528341afe5069adc.1645484982.git.fthain@linux-m68k.orgReviewed-by:Finn Thain <fthain@linux-m68k.org> Signed-off-by:
-
Finn Thain authored
This driver doesn't use SCp.ptr to save a SCSI command data pointer which means "scsi pointer" is a complete misnomer here. Only a few members of struct scsi_pointer are used and the rest waste memory. Avoid the "struct foo { struct bar; };" silliness. Link: https://lore.kernel.org/r/fbf930e64af5b15ca028dfe25b00fe933951f19b.1645484982.git.fthain@linux-m68k.orgReviewed-by: -
Khazhismel Kumykov authored
Most importantly: eh_timed_out() is not limited by scmd->allowed, and can reset timer forever. Fixes: c829c394 ("[SCSI] FC transport : Avoid device offline cases by stalling aborts until device unblocked") Link: https://lore.kernel.org/r/20220219001601.3534043-1-khazhy@google.comReviewed-by:
Khazhismel Kumykov <khazhy@google.com> Signed-off-by:
-
Christophe JAILLET authored
Just a few lines below this kzalloc() we have a mutex_lock() which can sleep. Moreover, the only way to call this function is when a delayed work is schedule. And delayed work can sleep: INIT_DELAYED_WORK(&fw_event->work, mptsas_firmware_event_work); --> mptsas_firmware_event_work() --> mptsas_send_link_status_event() --> mptsas_expander_add() So there is really no good reason to use GFP_ATOMIC here. Change it to GFP_KERNEL to give more opportunities to the kernel. Link: https://lore.kernel.org/r/eccb2179ce800529851ed4fabc9d3f95fbbf7d7f.1644906731.git.christophe.jaillet@wanadoo.frSigned-off-by:Christophe JAILLET <christophe.jaillet@wanadoo.fr> Signed-off-by:
-
Gustavo A. R. Silva authored
Use flexible-array members in struct fc_fdmi_attr_entry and fs_fdmi_attrs instead of one-element arrays, and refactor the code accordingly. Also, this helps with the ongoing efforts to globally enable -Warray-bounds and get us closer to being able to tighten the FORTIFY_SOURCE routines on memcpy(). https://github.com/KSPP/linux/issues/79 https://github.com/ClangBuiltLinux/linux/issues/1590 Link: https://lore.kernel.org/r/20220214223903.GA859464@embeddedorSigned-off-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by:
-
- 23 Feb, 2022 21 commits
-
-
Damien Le Moal authored
Make the driver messages more readable by adding a space after the message prefix ":" and removing the extra space between function name and line number. Link: https://lore.kernel.org/r/20220220031810.738362-32-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
Both pm8001_send_abort_all() and pm80xx_send_abort_all() are called only for a non null device with the NCQ_READ_LOG_FLAG set, so remove the device check on entry of these functions. Furthermore, setting the NCQ_ABORT_ALL_FLAG device id flag and clearing the NCQ_READ_LOG_FLAG is always done before calling these functions. Move these operations inside the functions. Link: https://lore.kernel.org/r/20220220031810.738362-31-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
The task argument of the pm8001_ccb_task_free() function can be inferred from the ccb argument ccb_task field. So there is no need to have this argument. Likewise, the ccb_index argument is always equal to the ccb tag field and is not needed either. Remove both arguments and update all call sites. The pm8001_ccb_task_free_done() helper is also modified to match this change. Link: https://lore.kernel.org/r/20220220031810.738362-30-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
The main part of the pm8001_task_exec() function uses a do {} while(0) loop that is useless and only makes the code harder to read. Remove this loop. The unnecessary local variable t is also removed. Additionally, avoid repeatedly declaring "struct task_status_struct *ts" to handle error cases by declaring this variable for the entire function scope. This allows simplifying the error cases, and together with the addition of blank lines make the code more readable. Finally, handling of the running_req counter is fixed to avoid decrementing it without a corresponding incrementation in the case of an invalid task protocol. Link: https://lore.kernel.org/r/20220220031810.738362-29-damien.lemoal@opensource.wdc.comReviewed-by: -
Damien Le Moal authored
There is no need to pass a pointer to a struct inbound_queue_table to pm8001_mpi_build_cmd(). Passing the start index in the inbound queue table of the adapter is enough. This change allows avoiding the declaration of a struct inbound_queue_table pointer (circularQ variables) in many functions, simplifying the code. While at it, blank lines are added i(e.g. after local variable declarations) to make the code more readable. Link: https://lore.kernel.org/r/20220220031810.738362-28-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
Introduce the pm8001_ccb_alloc() and pm8001_ccb_free() helpers to replace the typical code patterns: res = pm8001_tag_alloc(pm8001_ha, &ccb_tag); if (res) ... ccb = &pm8001_ha->ccb_info[ccb_tag]; ccb->device = pm8001_ha_dev; ccb->ccb_tag = ccb_tag; ccb->task = task; ccb->n_elem = 0; and ccb->task = NULL; ccb->ccb_tag = PM8001_INVALID_TAG; pm8001_tag_free(pm8001_ha, tag); With the simpler function calls: ccb = pm8001_ccb_alloc(pm8001_ha, pm8001_ha_dev, task); if (!ccb) ... and pm8001_ccb_free(pm8001_ha, ccb); The pm8001_ccb_alloc() helper ensures that all fields of the ccb info structure for the newly allocated tag are all initialized, except the buf_prd field. The pm8001_ccb_free() helper clears the initialized fields and the ccb tag to ensure that iteration over the adapter ccb_info array detects ccbs that are in use. All call site of the pm8001_tag_alloc() function that use a ccb info associated with an allocated tag are converted to use the new helpers. Link: https://lore.kernel.org/r/20220220031810.738362-27-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
To detect if a command is NCQ, there is no need to test all possible NCQ command codes. Instead, use ata_is_ncq() to test the command protocol. Link: https://lore.kernel.org/r/20220220031810.738362-26-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
Replace the goto statement in the for loop with "break" and remove the ex_err label. Also fix long lines, identation and blank lines to make the code more readable. Link: https://lore.kernel.org/r/20220220031810.738362-25-damien.lemoal@opensource.wdc.comSigned-off-by:
-
Damien Le Moal authored
To detect if a command is NCQ, there is no need to test all possible NCQ command codes. Instead, use ata_is_ncq() to test the command protocol. Link: https://lore.kernel.org/r/20220220031810.738362-24-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
In pm8001_chip_fw_flash_update_build(), if pm8001_chip_fw_flash_update_build() fails, the struct fw_control_ex allocated must be freed. Link: https://lore.kernel.org/r/20220220031810.738362-23-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
In pm8001_chip_set_dev_state_req(), pm8001_chip_fw_flash_update_req(), pm80xx_chip_phy_ctl_req() and pm8001_chip_reg_dev_req() add missing calls to pm8001_tag_free() to free the allocated tag when pm8001_mpi_build_cmd() fails. Similarly, in pm8001_exec_internal_task_abort(), if the chip ->task_abort method fails, the tag allocated for the abort request task must be freed. Add the missing call to pm8001_tag_free(). Link: https://lore.kernel.org/r/20220220031810.738362-22-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
In pm8001_send_abort_all(), make sure to free the allocated sas task if pm8001_tag_alloc() or pm8001_mpi_build_cmd() fail. Link: https://lore.kernel.org/r/20220220031810.738362-21-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
The function pm8001_tag_alloc() determines free tags using the function find_first_zero_bit() which can return 0 when the first bit of the bitmap being inspected is 0. As such, tag 0 is a valid tag value that should not be dismissed as invalid. Fix the functions pm8001_work_fn(), mpi_sata_completion(), pm8001_mpi_task_abort_resp() and pm8001_open_reject_retry() to not dismiss 0 tags as invalid. The value 0xffffffff is used for invalid tags for unused ccb information structures. Add the macro definition PM8001_INVALID_TAG to define this value. Link: https://lore.kernel.org/r/20220220031810.738362-20-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
The call to pm8001_ccb_task_free() at the end of pm8001_mpi_task_abort_resp() already frees the ccb tag. So when the device NCQ_ABORT_ALL_FLAG is set, the tag should not be freed again. Also change the hardcoded 0xBFFFFFFF value to ~NCQ_ABORT_ALL_FLAG as it ought to be. Link: https://lore.kernel.org/r/20220220031810.738362-19-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
In mpi_set_phy_profile_req() and in pm8001_set_phy_profile_single(), if pm8001_tag_alloc() fails to allocate a new tag, a warning message is issued but the uninitialized tag variable is still used to build a command. Avoid this by returning early in case of tag allocation failure. Also make sure to always return the error code returned by pm8001_tag_alloc() when this function fails instead of an arbitrary value. Link: https://lore.kernel.org/r/20220220031810.738362-18-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
In pm80xx_send_abort_all(), the n_elem field of the ccb used is not initialized to 0. This missing initialization sometimes lead to the task completion path seeing the ccb with a non-zero n_elem resulting in the execution of invalid dma_unmap_sg() calls in pm8001_ccb_task_free(), causing a crash such as: [ 197.676341] RIP: 0010:iommu_dma_unmap_sg+0x6d/0x280 [ 197.700204] RSP: 0018:ffff889bbcf89c88 EFLAGS: 00010012 [ 197.705485] RAX: dffffc0000000000 RBX: 0000000000000000 RCX: ffffffff83d0bda0 [ 197.712687] RDX: 0000000000000002 RSI: 0000000000000000 RDI: ffff88810dffc0d0 [ 197.719887] RBP: 0000000000000000 R08: 0000000000000000 R09: ffff8881c790098b [ 197.727089] R10: ffffed1038f20131 R11: 0000000000000001 R12: 0000000000000000 [ 197.734296] R13: ffff88810dffc0d0 R14: 0000000000000010 R15: 0000000000000000 [ 197.741493] FS: 0000000000000000(0000) GS:ffff889bbcf80000(0000) knlGS:0000000000000000 [ 197.749659] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 197.755459] CR2: 00007f16c1b42734 CR3: 0000000004814000 CR4: 0000000000350ee0 [ 197.762656] Call Trace: [ 197.765127] <IRQ> [ 197.767162] pm8001_ccb_task_free+0x5f1/0x820 [pm80xx] [ 197.772364] ? do_raw_spin_unlock+0x54/0x220 [ 197.776680] pm8001_mpi_task_abort_resp+0x2ce/0x4f0 [pm80xx] [ 197.782406] process_oq+0xe85/0x7890 [pm80xx] [ 197.786817] ? lock_acquire+0x194/0x490 [ 197.790697] ? handle_irq_event+0x10e/0x1b0 [ 197.794920] ? mpi_sata_completion+0x2d70/0x2d70 [pm80xx] [ 197.800378] ? __wake_up_bit+0x100/0x100 [ 197.804340] ? lock_is_held_type+0x98/0x110 [ 197.808565] pm80xx_chip_isr+0x94/0x130 [pm80xx] [ 197.813243] tasklet_action_common.constprop.0+0x24b/0x2f0 [ 197.818785] __do_softirq+0x1b5/0x82d [ 197.822485] ? do_raw_spin_unlock+0x54/0x220 [ 197.826799] __irq_exit_rcu+0x17e/0x1e0 [ 197.830678] irq_exit_rcu+0xa/0x20 [ 197.834114] common_interrupt+0x78/0x90 [ 197.840051] </IRQ> [ 197.844236] <TASK> [ 197.848397] asm_common_interrupt+0x1e/0x40 Avoid this issue by always initializing the ccb n_elem field to 0 in pm8001_send_abort_all(), pm8001_send_read_log() and pm80xx_send_abort_all(). Link: https://lore.kernel.org/r/20220220031810.738362-17-damien.lemoal@opensource.wdc.com Fixes: c6b9ef57 ("[SCSI] pm80xx: NCQ error handling changes") Reviewed-by:
-
Damien Le Moal authored
NCQ NON DATA is an NCQ command with the DMA_NONE DMA direction and so a register-device-to-host-FIS response is expected for it. However, for an IO_SUCCESS case, mpi_sata_completion() expects a set-device-bits-FIS for any ata task with an use_ncq field true, which includes NCQ NON DATA commands. Fix this to correctly treat NCQ NON DATA commands as non-data by also testing for the DMA_NONE DMA direction. Link: https://lore.kernel.org/r/20220220031810.738362-16-damien.lemoal@opensource.wdc.com Fixes: dbf9bfe6 ("[SCSI] pm8001: add SAS/SATA HBA driver") Reviewed-by:
-
Damien Le Moal authored
In the pm8001_chip_sata_req() and pm80xx_chip_sata_req() functions, all tasks with a DMA direction of DMA_NONE (no data transfer) are initialized using the ATAP value 0x04. However, NCQ NON DATA commands, while being DMA_NONE commands are NCQ commands and need to be initialized using the value 0x07 for ATAP, similarly to other NCQ commands. Make sure that NCQ NON DATA command tasks are initialized similarly to other NCQ commands by also testing the task "use_ncq" field in addition to the DMA direction. While at it, reorganize the code into a chain of if - else if - else to avoid useless affectations and debug messages. Link: https://lore.kernel.org/r/20220220031810.738362-15-damien.lemoal@opensource.wdc.com Fixes: dbf9bfe6 ("[SCSI] pm8001: add SAS/SATA HBA driver") Reviewed-by:
-
Damien Le Moal authored
In pm8001_pci_resume(), the use of the u32 type for the local variable device_state causes a sparse warning: warning: incorrect type in assignment (different base types) expected unsigned int [usertype] device_state got restricted pci_power_t [usertype] current_state Since this variable is used only once in the function, remove it and use pdev->current_state directly. While at it, also add a blank line after the last local variable declaration. Link: https://lore.kernel.org/r/20220220031810.738362-14-damien.lemoal@opensource.wdc.comReviewed-by: -
Damien Le Moal authored
In mpi_set_phy_profile_req() and pm8001_set_phy_profile_single(), use cpu_to_le32() to initialize the ppc_phyid field of struct set_phy_profile_req. Furthermore, fix the definition of the reserved field of this structure to be an array of __le32, to match the use of cpu_to_le32() when assigning values. These changes remove several sparse type warnings. Link: https://lore.kernel.org/r/20220220031810.738362-13-damien.lemoal@opensource.wdc.comReviewed-by:
-
Damien Le Moal authored
Make sure that the __le32 fields of struct sata_cmd are manipulated after applying the correct endian conversion. That is, use cpu_to_le32() for assigning values and le32_to_cpu() for consulting a field value. In particular, make sure that the calculations for the 4G boundary check are done using CPU endianness and *not* little endian values. With these fixes, many sparse warnings are removed. While at it, fix some code identation and add blank lines after variable declarations and in some other places to make this code more readable. Link: https://lore.kernel.org/r/20220220031810.738362-12-damien.lemoal@opensource.wdc.com Fixes: 0ecdf00b ("[SCSI] pm80xx: 4G boundary fix.") Reviewed-by:
-
