- 25 Sep, 2014 40 commits
-
-
Arun Easi authored
ABTS path takes path similar to regular task management, but reset_count was not initialized when allocating mcmd causing the response to get dropped in qlt_xmit_tm_rsp. Fix this by initializing reset_count correctly. Signed-off-by:
Arun Easi <arun.easi@qlogic.com> Signed-off-by:
Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by:
Christoph Hellwig <hch@lst.de>
-
Arun Easi authored
A chip reset can occur after driver submits command to the stack. Abort command processing if a chip reset has occurred or in progress when you get a follow up for a command. Signed-off-by:
-
Arun Easi authored
Signed-off-by:
-
Himanshu Madhani authored
tcm_qla2xxx_get_task_tag incorrectly assume command is qla_tgt_cmd. Add check to see if it's a scsi cmd or task mgmt command. Signed-off-by:
Quinn Tran <quinn.tran@qlogic.com> Signed-off-by:
-
Saurav Kashyap authored
Signed-off-by:
Giridhar Malavali <giridhar.malavali@qlogic.com> Signed-off-by:
-
Arun Easi authored
Sequence level error recovery (aka FC Tape) is not really required for disk devices. On heavily loaded system, with slow turn around, a bunch of status enquiries using REC puts additional burden to the target, so just turn off SLER by default. Signed-off-by:
-
Arun Easi authored
Turning logging bits for target mode ON dumps quite a lot verbose messages, remove those and change some of the IO path logging to use IO bits. Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
Himanshu Madhani <himanshu.madhani@qlogic.com> Signed-off-by:
-
Arun Easi authored
Cc: <stable@vger.kernel.org> Signed-off-by:
-
Himanshu Madhani authored
This patch is to fix regression added by commit id 51a07f84. When allocating memory for new session original patch does not assign vha to op->vha resulting into NULL pointer access during qlt_create_sess_from_atio(). Cc: <stable@vger.kernel.org> Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
-
Saurav Kashyap authored
Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
Joe Carnuccio <joe.carnuccio@qlogic.com> Signed-off-by:
-
Sawan Chandak authored
Signed-off-by:
Sawan Chandak <sawan.chandak@qlogic.com> Signed-off-by:
-
Chad Dupuis authored
If we fail to load one of the segments of firmware fail the operation instead of trying to continue which will lead to a system crash since the pointers to the next segment will not be updated correctly. Signed-off-by:
Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by:
-
Chad Dupuis authored
Disable the PCI device during shutdown to prevent any races with other PCI code such as the AER handling code. Signed-off-by:
-
Chad Dupuis authored
Signed-off-by:
-
Himanshu Madhani authored
For FA-WWPN capable device, ISP2031 and ISP27XX, when loop dead is detected by a driver, restore WWPN from NVRAM. Signed-off-by:
-
Chad Dupuis authored
Parse the retry delay timer field from the FCP response data and if: - It is not zero - The SCSI status is busy or queue full return SCSI_MLQUEUE_TARGET_BUSY for the number of milliseconds specified in the retry delay timer field. Signed-off-by:
-
Alex Vechersky authored
Signed-off-by:
Alex Vechersky <alex.vechersky@qlogic.com> Signed-off-by:
-
Chad Dupuis authored
Signed-off-by:
-
Nigel Kirkland authored
Signed-off-by:
Nigel Kirkland <nigel.kirkland@qlogic.com> Signed-off-by:
-
Himanshu Madhani authored
Nameserver data on FC switch is not refreshed when qla2xxx driver is unloaded. Disabling laser for ISP2031 will force FC switch to rescan ports and clear fdmi entries from Nameserver. Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
-
Quinn Tran authored
Declaration error of mb array in qla2x00_iidma_fcport cause data to be written beyond the array. This ends up causing stack corruption. Signed-off-by:
-
Sawan Chandak authored
Signed-off-by:
-
Arun Easi authored
Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
-
Joe Carnuccio authored
Add support for the D-port (diagnostic port) fabric switch feature. (Fabric Switch initiates loopback style port testing) Signed-off-by:
-
Joe Carnuccio authored
Add support for the FA-WWN (fabric assigned wwn) fabric switch feature. (Fabric Assigned World Wide Port Name) Signed-off-by:
-
Himanshu Madhani authored
Add support for the FDMI-2 fabric switch feature. Since FDMI-2 uses code from FDMI-1, some of the existing code needed to be repaired to prevent fields from being overflowed. Signed-off-by:
-
Jan Kulich authored
Signed-off-by:
Jan Kulich <jan.kulich@qlogic.com> Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Joe Carnuccio authored
Semaphore register does not exist for ISPFx00. Signed-off-by:
-
Joe Carnuccio authored
Call scnprintf() instead of snprintf() since the latter may return an incorrect count in cases where the write is truncated to fit. scnprintf() returns the count of what was actually written; snprintf() returns the count of what would have been written. Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Joe Carnuccio authored
When creating request/response queues from qla25xx_setup_mode(), the shadow index register pointers were not being initialized to point at the registers. Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
