- 07 Feb, 2017 13 commits
-
-
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:
Subhash Jadavani <subhashj@codeaurora.org> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
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:
Hannes Reinecke <hare@suse.com> 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:
-
- 03 Feb, 2017 24 commits
-
-
Raghava Aditya Renukunta authored
Update the driver version to 50740 Signed-off-by:
Dave Carroll <David.Carroll@microsemi.com> Reviewed-by:
-
Raghava Aditya Renukunta authored
Change the aacraid driver prefix from 1.2-1 to 1.2.1 Signed-off-by:
-
Raghava Aditya Renukunta authored
Added new copyright messages Signed-off-by:
-
Raghava Aditya Renukunta authored
Added a new ioctl interface to retrieve the host device information. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added a new ioctl interface to trigger an IOP or IWBR reset from ioctl. Primary used by management utility to trigger resets. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added a new IWBR soft reset type, reworked the IOP reset interface for a bit. Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch adds support to retrieve the unique identifier data (VPD page 83 type3) for Logical drives created on SmartIOC 2000 products. In addition added a sysfs device structure to expose the id information. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added task management command support to abort any timed out commands in case of a eh_abort call and to reset lun's in case of eh_reset call. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added support to send out task management commands. [mkp: removed // fibsize... ] Signed-off-by:
-
Raghava Aditya Renukunta authored
Added support to send direct pasthru srb commands from management utilty to the controller. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added support for drive hotplug add and removal Signed-off-by:
-
Raghava Aditya Renukunta authored
Added support to set qd of drives in slave_configure.This only works for HBA1000 attached drives. Signed-off-by:
-
Raghava Aditya Renukunta authored
Retrieved queue depth from fw and saved it for future use. Only applicable for HBA1000 drives. Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch adds a new functions that periodically sync the time of host to the adapter. In addition also informs the adapter that the driver is alive and kicking. Only applicable to the HBA1000 and SMARTIOC2000. Signed-off-by:
-
Raghava Aditya Renukunta authored
Reworked aac_command_thread into aac_process_events Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch processes Raw IO read medium errors. Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch enables the driver to actually process the I/O, or srb replies from adapter. In addition to any HBA1000 or SmartIOC2000 adapter events. Signed-off-by:
-
Raghava Aditya Renukunta authored
Make sure that the driver processes error conditions even in the fast response path for response from the adapter. Signed-off-by:
-
Raghava Aditya Renukunta authored
Moved the READ and WRITE switch cases to the top. Added a default case to the switch case and replaced duplicate scsi result value with a macro. The idea is that since most of scsi commands we care about performance wise are read or write, we need to process them first. Internally the compiler (GCC) converts a switch case into either a jump table or a bunch of if else conditions, so placing the often used read, write cases at the top is an effort in optimization. Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch adds support to retrieve the type of each adapter connected device. Applicable to HBA1000 and SmartIOC2000 products Signed-off-by:
-
Raghava Aditya Renukunta authored
sa_firmware adds the capability to differentiate the new SmartIOC family of adapters from the series 8 and below. Signed-off-by:
-
Raghava Aditya Renukunta authored
This patch lays the groundwork for supporting the new HBA-1000 controller family.A new INIT structure INIT_STRUCT_8 has been added which allows for a variable size for MSI-x vectors among other things, and is used for both Series-8, HBA-1000 and SmartIOC-2000. Signed-off-by:
-
Raghava Aditya Renukunta authored
Added aacraid.h include guard Signed-off-by:
-
Raghava Aditya Renukunta authored
Removed duplicate code that for acquiring and releasing irqs Signed-off-by:
-
- 01 Feb, 2017 3 commits
-
-
Lukas Herbolt authored
[mkp: whitespace fixes] Signed-off-by:
Lukas Herbolt <lherbolt@redhat.com> Signed-off-by:
-
Chaitra P B authored
Updated driver version to "15.100.00.00" Signed-off-by:
Chaitra P B <chaitra.basappa@broadcom.com> Signed-off-by:
Suganath Prabu S <suganath-prabu.subramani@broadcom.com> Reviewed-by:
-
Chaitra P B authored
Due existence of loop in the IO path our HBA will receive heavy IOs and also as driver is not updating the Reply Post Host Index frequently, So there will be a high chance that our Firmware unable to find any free entry in the Reply Post Descriptor Queue (i.e. Queue overflow occurs) and can observe 0x2100 firmware fault. So to fix this, we have defined a thresh hold value. After continuously processing this thresh hold number of reply descriptors driver will update the Reply Descriptor Host Index so that this thresh hold number of reply descriptors entries will be freed and these entries will be available for firmware and we won't observe this Firmware fault. We have defined this threshold value as 1/3rd of the hba queue depth. Signed-off-by:
-
