- 13 Feb, 2017 24 commits
-
-
Shivasharan S authored
scsi: megaraid_sas: set pd_after_lb from MR_BuildRaidContext and initialize pDevHandle to MR_DEVHANDLE_INVALID Issue is limited for Syncro firmware where pd_after_lb is not set but is accidentally used. Not a functional issue, but results in low performance due to improper load balancing between two LUNs. Signed-off-by:
Kashyap Desai <kashyap.desai@broadcom.com> Signed-off-by:
Shivasharan S <shivasharan.srikanteshwara@broadcom.com> Reviewed-by:
Hannes Reinecke <hare@suse.com> Reviewed-by:
Tomas Henzl <thenzl@redhat.com> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Shivasharan S authored
Fetch the latest controller OCR capability from FW before sending MR_DCMD_CTRL_SHUTDOWN When application sends a shutdown DCMD (MR_DCMD_CTRL_SHUTDOWN), driver will fetch latest controller information from firmware. This is to ensure that driver always has latest OCR capability of controller before sending the DCMD. Signed-off-by:
-
Shivasharan S authored
Fix kernel warning for accessing unaligned memory access in driver. Signed-off-by:
-
Shivasharan S authored
Fix endiannes fixes for Ventura specific. Signed-off-by:
-
Shivasharan S authored
Fix if RDPQ mode enabled MR FW is deployed on big endian host machine, driver does not setup reply address correctly. Signed-off-by:
-
Shivasharan S authored
Change MR_TargetIdToLdGet return type from u8 to u16. ld id range check is added at two places in this patch - @megasas_build_ldio_fusion and @megasas_build_ld_nonrw_fusion. Previous driver code used different data type for lds TargetId returned from MR_TargetIdToLdGet. Prior to this change, above two functions was safeguarded due to function always return u8 and maximum value of ld id returned was 255. In below check, fw_supported_vd_count as of today is 64 or 256 and valid range to support is either 0-63 or 0-255. Ideally want to filter accessing raid map for ld ids which are not valid. With the u16 change, invalid ld id value is 0xFFFF and we will see kernel panic due to random memory access in MR_LdRaidGet. The changes will ensure we do not call MR_LdRaidGet if ld id is beyond size of ldSpanMap array. if (ld < instance->fw_supported_vd_count) >From firmware perspective,ld id 0xFF is invalid and even though current driver code forward such command, firmware fails with target not available. ld target id issue occurs mainly whenever driver loops to populate raid map (ea. MR_ValidateMapInfo). These are the only two places where we may see out of range target ids and wants to protect raid map access based on range provided by Firmware API. Signed-off-by: -
Shivasharan S authored
scsi: megaraid_sas: In validate raid map, raid capability is not converted to cpu format for all lds On a host, if an ld is deleted there is a hole in the ld array returned by the FW. But in MR_ValidateMapInfo we are not accounting for holes in the ld array and traverse only upto index num_lds. This patch takes care of converting the capability field of all the valid lds in the ld raid map. Signed-off-by:
-
Shivasharan S authored
Currently fusion context has fixed array load_balance_info. Use dynamic allocation. In few places, driver do not want physically contigious memory. Attempt to use vmalloc if physical contiguous memory is not available. Signed-off-by:
-
Shivasharan S authored
Signed-off-by:
-
Shivasharan S authored
Add additional logging from driver in OCR context. Add debug logs for partial completion of IOs is iodone context. Signed-off-by:
-
Shivasharan S authored
Fixing issue of not setting residual bytes correctly. Signed-off-by:
-
Shivasharan S authored
Avoid Host side PCI bandwidth bottleneck and hint FW to do Write buffering using RaidFlag MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT. Once IO is landed in FW with MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT, it will do single DMA from host and buffer the Write operation. On back end, FW will DMA same buffer to the Mirror and Data Arm. This will improve large block IO performance which bottleneck due to Host side PCI bandwidth limitation. Consistent ~4000MB T.P for 256K Block size is expected performance numbers. IOPS for small Block size should be on par with Disk performance. (E.g 42 SAS Disk in JBOD mode gives 3700MB T.P. Same Drive used in R1 WT mode, should give ~1800MB T.P) Using this patch 24 R1 VDs (HDD) gives below performance for Sequential Write. Without this patch, we cannot reach above 3200MB (Throughput is in MB.) Block Size 50% 256K and 50% 4K 100% 256K 4K 3100 2030 8K 3140 2740 16K 3140 3140 32K 3400 3240 64K 3500 3700 128K 3870 3870 256K 3920 3920 Signed-off-by:
-
Shivasharan S authored
This patch provide true fast path IO support. Driver creates PRP for NVME drives and send Fast Path for performance. Certain h/w requirement needs to be taken care in driver. Signed-off-by:
-
Shivasharan S authored
This patch fetch true values of NVME property from FW using New DCMD interface MR_DCMD_DEV_GET_TARGET_PROP Signed-off-by:
-
Shivasharan S authored
Adding detection logic for NVME device attached behind Ventura controller. Driver set HostPageSize in IOC_INIT frame to inform about page size for NVME devices. Firmware reports NVME page size to the driver. PD INFO DCMD provide new interface type NVME_PD. Driver set property of NVME device. Signed-off-by:
-
Shivasharan S authored
With the changes to remove checks for a valid request descriptor, issue_dcmd will now always return DCMD_SUCCESS. This patch changes return type of issue_dcmd to void and change all callers appropriately. Signed-off-by:
-
Shivasharan S authored
No functional change. Code clean up. Removing error code which is not valid scenario. In megasas_get_request_descriptor we can remove the error handling which is not required. With fusion controllers, if there is a valid message frame available, we are guaranteed to get a corresponding request descriptor. Signed-off-by:
-
Shivasharan S authored
Moving to use DID_REQUEUE return type for reliable unconditional retries. Driver wants unconditional re-queue, so replace DID_RESET with DID_REQUEUE Discussed below - https://www.spinics.net/lists/linux-scsi/msg102848.htmlSigned-off-by:
-
Shivasharan S authored
Signed-off-by:
-
Shivasharan S authored
Signed-off-by:
-
Shivasharan S authored
No functional change. Code refactor. megasas_fire_cmd_fusion can always use 32 bit descriptor write for ventura. No need to pass extra flag. Only IOC INIT required 64 bit Descriptor write. Signed-off-by:
-
Shivasharan S authored
No functional change. Code refactor. Remove function megasas_fpio_to_ldio as we never require to convert fpio to ldio because of frame unavailability. Grab extra frame of raid 1 write fast path before it creates first frame as Fast Path. Removed is_raid_1_fp_write flag as raid 1 write fast path command is decided using r1_alt_dev_handle only. Move resetting megasas_cmd_fusion fields at common function megasas_return_cmd_fusion. Signed-off-by:
-
Shivasharan S authored
No functional change. Code refactor. Signed-off-by:
-
Shivasharan S authored
This reverts commit "3e5eadb1" ("scsi: megaraid_sas: Enable or Disable Fast path based on the PCI Threshold Bandwidth") This patch was aimed to increase performance of R1 Write operation for large IO size. Since this method used timer approach, it turn on/off fast path did not work as expected. Patch 0013 describes new algorithm and performance number. Signed-off-by:
-
- 09 Feb, 2017 1 commit
-
-
Wei Yongjun authored
'conn_info' is malloced in qedi_iscsi_update_conn() and should be freed before leaving from the error handling cases, otherwise it will cause memory leak. Fixes: ace7f46b ("scsi: qedi: Add QLogic FastLinQ offload iSCSI driver framework.") Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Acked-by:
Manish Rangankar <Manish.Rangankar@cavium.com> Signed-off-by:
-
- 07 Feb, 2017 15 commits
-
-
Venkat Gopalakrishnan authored
Dump host state, power info and host/vendor specific registers on link failures. This provides useful info to debug the failures. Signed-off-by:
Venkat Gopalakrishnan <venkatg@codeaurora.org> Signed-off-by:
Subhash Jadavani <subhashj@codeaurora.org> Signed-off-by:
-
Venkat Gopalakrishnan authored
Change testbus default config, dump additional testbus registers along with other debug vendor specific registers. These additional info are useful in debugging link related failures. Signed-off-by:
-
subhashj@codeaurora.org authored
UFS clock scaling might start kicking in even before the device is running at the fastest interface speed which is undesirable. This change moves the clock scaling kick start only after the device is detected and running at the fastest interface speed. Signed-off-by:
-
subhashj@codeaurora.org authored
Currently we are suspending clock scaling during clock gating which doesn't allow us to have clock gating timeout lower than clock scaling polling window. If clock gating timeout is smaller than the clock scaling polling window then we will mostly suspend the clock scaling before clock scaling polling window expires and we might get stuck in same state (scaled down or scaled up) for quite a long time. And for this reason, we have clock gating timeout (150ms) greater than clock scaling polling window (100ms). We would like to have aggressive clock gating timeout even lower than the clock scaling polling window hence this change is decoupling the clock scaling suspend/resume from clock gate/ungate. We will not suspend the clock scaling as part of clock gating instead clock scaling context will schedule scaling suspend work if there are no more pending transfer requests. Signed-off-by:
-
Gilad Broner authored
Whenever some UFS failure occurs the driver prints the UFS registers in order to help with analysis of the failure. However this may not be sufficient in some cases, so having the host controller state as it is represented and managed in the driver will contribute to analysis efforts. Added prints of various fields in the hba struct which may be of interest. Signed-off-by:
Gilad Broner <gbroner@codeaurora.org> Signed-off-by:
-
subhashj@codeaurora.org authored
UFS driver's load based clock scaling feature scales down the ufs related clocks in order to allow low power modes of chipsets. UniPro 1.6 supports maximum gear up to HS-G3 (High Speed Gear3) and some of the chipsets low power modes may not be allowed in HS-G3 hence this change adds support to scale gear between HS-G3 and HS-G1 based on same existing load based clock scaling logic. Signed-off-by:
-
Gilad Broner authored
Details printed for each request that is aborted can overload the target as there can be several requests that are aborted at once. This change will print full request details only for the first aborted request since the last link reset, and minimal details for other subsequent requests. Signed-off-by:
-
Gilad Broner authored
On certain error conditions request abort task itself might fail when aborting a request. In such case, subsequent request aborts should skip issuing the abort task as it is expected to fail as well, and device reset handler will be called next. Signed-off-by:
-
Arnd Bergmann authored
Shifting a dma_addr_t right by 32 bits causes a compile-time warning when that type is only 32 bit wide: drivers/scsi/aacraid/src.c: In function 'aac_src_start_adapter': drivers/scsi/aacraid/src.c:414:29: error: right shift count >= width of type [-Werror=shift-count-overflow] This changes the driver to use the predefined macros consistently, including one correct but open-coded upper_32_bits() instance. Fixes: d1ef4da8 ("scsi: aacraid: added support for init_struct_8") Fixes: 423400e6 ("scsi: aacraid: Include HBA direct interface") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Reviewed-by:
Johannes Thumshirn <jthumshirn@suse.de> Reviewed-by:
Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com> Signed-off-by:
-
Colin Ian King authored
aac_fib_send can return -ve error returns and hence rcode should be signed. Currently the rcode >= 0 check is always true and -ve errors are not being checked. Thanks to Dan Carpenter for spotting my original broken fix to this issue. Signed-off-by:
Colin Ian King <colin.king@canonical.com> Reviewed-by:
-
Christoph Hellwig authored
Signed-off-by:
Christoph Hellwig <hch@lst.de> Acked-by:
Jack Wang <jinpu.wang@profitbricks.com> Signed-off-by:
-
Christoph Hellwig authored
They are never called and just dispatch to methods of the same names in the FC and SRP transport classes that are never implemented. Signed-off-by:
-
Christoph Hellwig authored
Instead define the timeout behavior purely based on the host_template eh_timed_out method and wire up the existing transport implementations in the host templates. This also clears up the confusion that the transport template method overrides the host template one, so some drivers have to re-override the transport template one. Signed-off-by:
Tyrel Datwyler <tyreld@linux.vnet.ibm.com> Signed-off-by:
-
Christoph Hellwig authored
EH_NOT_HANDLED is the default case if no eh_timed_out method is provided, so there is no need to supply it. Signed-off-by:
-
Christoph Hellwig authored
These days we can specify an eh_timed_out handler in the host_template, so don't have a transport_template definition just for it. [mkp: fixed typo] Signed-off-by:
-
