- 23 Mar, 2017 35 commits
-
-
Brian King authored
This patch closes up some potential race conditions observed in the error handling paths in ipr while debugging an issue resulting in a hang with SATA error handling. These patches ensure we are holding the correct lock when adding and removing commands from the free and pending queues in some error scenarios. Signed-off-by:
Brian King <brking@linux.vnet.ibm.com> Reviewed-by:
Wendy Xiong <wenxiong@linux.vnet.ibm.com> Tested-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Brian King authored
This fixes a race condition in the error handlomg paths of ipr. While a command is outstanding to the adapter, it is placed on a pending queue for the hrrq it is associated with, while holding the HRRQ lock. When a command is completed, it is removed from the pending queue, under HRRQ lock, and placed on a local list. This list is then iterated through without any locks and each command's done function is invoked, inside of which, the command gets returned to the free list while grabbing the HRRQ lock. This fixes two race conditions when commands have been removed from the pending list but have not yet been added to the free list. Both of these changes fix race conditions that could result in returning success from eh_abort_handler and then later calling scsi_done for the same request. The first race condition is in ipr_cancel_op. It looks through each pending queue to see if the command to be aborted is still outstanding or not. Rather than looking on the pending queue, reverse the logic to check to look for commands that are NOT on the free queue. The second race condition can occur when in ipr_wait_for_ops where we are waiting for responses for commands we've aborted. Signed-off-by:
-
Brian King authored
Removes some code in __ipr_eh_dev_reset which was modifying the ipr_cmd done function. This should have already been setup at command allocation time and if its since been changed, it means we are in the ipr_erp* functions and need to wait for them to complete and don't want to override that here. Signed-off-by:
-
Brian King authored
Following a command abort or device reset, ipr's EH handlers wait for the commands getting aborted to get sent back from the adapter prior to returning from the EH handler. This fixes up some cases where the completion handler was not getting called, which would have resulted in the EH thread waiting until it timed out, greatly extending EH time. Signed-off-by:
-
Xiaofei Tan authored
Add helper function is_sata_phy_v2_hw() to judge whether the attached device is SATA disk for a root PHY. Signed-off-by:
Xiaofei Tan <tanxiaofei@huawei.com> Signed-off-by:
John Garry <john.garry@huawei.com> Signed-off-by:
-
Xiang Chen authored
When SMP IO is sent, sas_protocol_ata couldn't judge whether the disk is SATA or SAS disk. So use dev_is_sata to identify SATA or SAS disk. Signed-off-by:
Xiang Chen <chenxiang66@hisilicon.com> Signed-off-by:
-
John Garry authored
Unless we actually get some sort of failure in hisi_sas_lu_reset(), don't print a message. Signed-off-by:
-
Xiang Chen authored
When an SMP task timeouts, it will call lldd_abort_task to release the associated slot, and then will release the sas_task. Currently in lldd_abort_task, if we fail to internally abort IO, then the slot of SMP IO is not released, but sas_task will still be later released, so the slot's sas_task is NULL, which will cause NULL pointer when hisi_sas_slot_task_free happens later. To resolve, check the return value of internal abort, and release the slot if it failed. Signed-off-by:
-
John Garry authored
Add function for upper-layer to reset controller when all else fails. Signed-off-by:
-
John Garry authored
For consistency, remove the "hisi_sas_" prefix. Signed-off-by:
-
Xiaofei Tan authored
Handle the situation that PHY UP and DOWN irq happen simultaneously. There is no mechanism of SoC HW to ensure this situation will never happen. So, we add this handle just in case. Signed-off-by:
-
John Garry authored
This patch includes: (1) Disable transport layer retry (2) Support CQ time and count interrupt coal (3) fix link FIFO full issue Signed-off-by:
Zhao Nenglong <zhaonenglong@hisilicon.com> Signed-off-by:
-
Xiang Chen authored
There are some rules to decide which error code has the high priority when errors happen together: (1) Error phase of CQ decides the error happens on RX or TX; (2) For TX error, when DMA/TRANS TX error happen simultaneously, the priority of DMA TX error is higher than TRANS TX error, so for the priority of TX error: DW2 (DMA TX part) > DW0; (3) For RX error, when TRANS/DMA/SIPC RX error happen simultaneously, the priority of TRANS RX error is higher than DMA and SIPC RX error, and we should also keep the rules (the priority of DW3 > DW2), so for the priority of RX error: DW1 > DW3 > DW2(SIPC RX part); (4) There are also a priority we should keep in the same error type. So, modify slot error code to handle this. In addition to this, some some error codes are modified according to recommendation from SoC designer. Signed-off-by: -
John Garry authored
When a TMF or internal abort times-out, do not free slot. We expect this to be done upon later escalated error handling. Signed-off-by:
-
John Garry authored
Some more locking needs to be added/modified for when read-modify-writing sas_task.task_state_flags. Note: since we can attempt to grab this lock in interrupt context we should use irq variant of spin_lock. Signed-off-by: -
Xiang Chen authored
After hardreset, we clear up IOs of remote disks, so we need to free those slots in LLDD. Signed-off-by:
-
John Garry authored
Check in slot_complete_v2_hw() for whether a task has already been completed by upper layer. Signed-off-by:
-
John Garry authored
When issuing an LU reset for a SATA target, issue an internal abort and a hard reset. Signed-off-by:
-
John Garry authored
Currently an internal abort is executed regardless of the result of the TMF. We should also check the result of the internal abort to see if we should free the slot. So change the status code STAT_IO_COMPLETE to TMF_RESP_FUNC_SUCC, meaning the slot has been successfully aborted. Signed-off-by:
-
Xiang Chen authored
For error codes which need abort-and-retry, simulate IO timeout and let SCSI+ATA layers process those errors. Previously for SSP, we should try to abort the IO in the LLDD and then pass back to upper layer, but sometimes this would also error. So Instead of adding special error handling for this scenario in the LLDD, allow the upper layer to handle completely. No performance hit is seen by taking this approach. Signed-off-by:
-
John Garry authored
We currently do a hard reset for a link reset. Change this to do a link reset only. Signed-off-by:
Xiang Chen <chenxiang@hisilicon.com> Signed-off-by:
-
John Garry authored
When sas_port is NULL, then return SAS_PHY_DOWN. In addition, when the sas_dev is gone then explicitly return SAS_PHY_DOWN. Signed-off-by:
-
John Garry authored
Currently when a root PHY is deformed from a asd_sas_port we try to release the slots in the LLDD, and fail. Regardless, it is not right to release this early. This patch removes the deformed function. As it was before, port deformation is still done in hisi_sas_phy_down(). It would be nice to actually remove the hisi_sas_port_{de}formed() pair, however we cannot as we need to know the asd_sas_port index libsas has associated with an asd_sas_phy. The hw does actually generate a port id for a PHY, but this seems to a random number, so ignored for this purpose. This patch also changes the code to link slots to the hisi_sas_device, and not hisi_sas_port. Signed-off-by: -
Xiang Chen authored
Add softreset to clear IO after internal abort device for SATA disk. The SATA error handling for the controller is based on device internal abort and softreset function. The controller does not support internal abort for single IO, so we need to execute internal abort for device. Signed-off-by:
-
John Garry authored
Relocate the PHY init code from LLDD hw init path to hisi_sas_scan_start(). Signed-off-by:
-
Xiang Chen authored
There are some scenarios that we need to warm-reset to reset registers of SAS controller. During reset we disable interrupts/DQs/PHYs, and after reset we re-init the hardware and rescan the topology to see if anything changed. Signed-off-by:
-
John Garry authored
Introduce function to get hisi_sas_port from asd_sas_port. Signed-off-by:
-
Satish Kharat authored
This is a bug introduced when they moved the fip subcodes to central place. Was sending FIP_SC_VL_NOTE in fip.fip_subcode for VLAN request in fnic_fcoe_send_vlan_req. Change is to use FIP_SC_VL_REQ instead. Signed-off-by:
Satish Kharat <satishkh@cisco.com> Signed-off-by:
Sesidhar Baddela <sebaddel@cisco.com> Signed-off-by:
-
Satish Kharat authored
The IO and Abort latency counter counts the time taken to complete the IO and abort command into broad buckets. This is not intended for performance measurement, just a debug statistic. current_max_io_time tries to keep track of the maximum time an IO has taken to complete if it is > 30sec. Signed-off-by:
-
Satish Kharat authored
Just a simple counter of number of check conditions encountered on that host. Signed-off-by:
-
Satish Kharat authored
If SCSI-ML has already issued abort on a command i.e FNIC_IOREQ_ABTS_PENDING is set and we get a IO completion, avoid this being flagged as out-of-order completion by setting the FNIC_IO_DONE flag in fnic_fcpio_icmnd_cmpl_handler Signed-off-by:
-
Satish Kharat authored
Fixing the IO stats update (Active IOs and IO completion) to prevent "Number of Active IOs" from becoming negative in the fnistats output. Signed-off-by:
-
Satish Kharat authored
Getting rid of else case to make the flow look bit simpler. Signed-off-by:
-
Satish Kharat authored
This is to avoid the log from being filled with vlan discovery messages when there is no vlan configured on the switch. Signed-off-by:
-
Christoph Hellwig authored
Not a full cleanup for the IRQ code, for that we'd need to know if the max number of the various CQ types is going to stay 1 forever. Signed-off-by:
Christoph Hellwig <hch@lst.de> Acked-by:
-
- 16 Mar, 2017 1 commit
-
-
Colin Ian King authored
Buffer is a pointer to the static char array event_buffer and therefore can never be null, so the check is redundant. Remove it. Detected by CoverityScan, CID#1077556 ("Logically Dead Code") Signed-off-by:Colin Ian King <colin.king@canonical.com> Acked-by:
Bradley Grove <bgrove@attotech.com> Signed-off-by:
-
- 15 Mar, 2017 4 commits
-
-
Dmitry Torokhov authored
Even if bus is not hot-pluggable, devices can be unbound from the driver via sysfs, so we should not be using __exit annotations on remove() methods. The only exception is drivers registered with platform_driver_probe() which specifically disables sysfs bind/unbind attributes. Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Charles authored
1. Add reboot notifier and register it in stex_probe for all supported device. 2. For all supported device in restart flow, we get a callback from notifier and set S6flag for stex_shutdown & stex_hba_stop to send restart command to FW. Signed-off-by:
Charles.Chiou <charles.chiou@tw.promise.com> Signed-off-by:
Paul.Lyu <paul.lyu@tw.promise.com> Signed-off-by:
-
Charles authored
Pegasus series is a RAID support product using Thunderbolt technology. The newest product, Pegasus 3(P3) supports Thunderbolt 3 technology with a different chip. 1. Change driver version. 2. Add P3 VID, DID and define it's device address. 3. P3 use msi interrupt, so stex_request_irq P3 type enable msi. 4. For hibernation, use msi_lock in stex_ss_handshake to prevent msi register write again when handshaking. 5. P3 doesn't need read() as flush. 6. In stex_ss_intr & stex_abort, P3 only clear interrupt register when getting vendor defined interrupt. Signed-off-by:
-
Elena Reshetova authored
refcount_t type and corresponding API should be used instead of atomic_t when the variable is used as a reference counter. This allows to avoid accidental refcounter overflows that might lead to use-after-free situations. Signed-off-by:
Elena Reshetova <elena.reshetova@intel.com> Signed-off-by:
Hans Liljestrand <ishkamiel@gmail.com> Signed-off-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
David Windsor <dwindsor@gmail.com> Acked-by:
Chris Leech <cleech@redhat.com> Signed-off-by:
-
