- 12 Nov, 2014 40 commits
-
-
Martin K. Petersen authored
The T10 SBC UNMAP command does not provide any hard guarantees that blocks will return zeroes on a subsequent READ. This is due to the fact that the device server is free to silently ignore all or parts of the request. The only way to ensure that a block consistently returns zeroes after being unmapped is to use WRITE SAME with the UNMAP bit set. Should the device be unable to unmap one or more blocks described by the command it is required to manually write zeroes to them. Until now we have preferred UNMAP over the WRITE SAME variants to accommodate thinly provisioned devices that predated the final SBC-3 spec. This patch changes the heuristic so that we favor WRITE SAME(16) or (10) over UNMAP if these commands are marked as supported in the Logical Block Provisioning VPD page. The patch also disables discard_zeroes_data for devices operating in UNMAP mode. Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com> Reviewed-by:
Paolo Bonzini <pbonzini@redhat.com> Signed-off-by:
Christoph Hellwig <hch@lst.de>
-
Christoph Hellwig authored
Signed-off-by:
Hannes Reinecke <hare@suse.de>
-
Christoph Hellwig authored
Signed-off-by:
-
Christoph Hellwig authored
Signed-off-by:
-
Christoph Hellwig authored
No need to verify the passthrough ioctls, the real handler will take care of that. Also make sure not to block for resets on O_NONBLOCK fds. Signed-off-by:
-
Christoph Hellwig authored
The calling conventions for this function are bad as it could return -ENODEV both for a device not currently online and a not recognized ioctl. Add a new scsi_ioctl_block_when_processing_errors function that wraps scsi_block_when_processing_errors with the a special case for the SG_SCSI_RESET ioctl command, and handle the SG_SCSI_RESET case itself in scsi_ioctl. All callers of scsi_ioctl now must call the above helper to check for the EH state, so that the ioctl handler itself doesn't have to. Reported-by:
Robert Elliott <Elliott@hp.com> Signed-off-by:
-
Christoph Hellwig authored
Pull the common code from the two callers into the function, and rename it to scsi_ioctl_reset. Signed-off-by:
-
Bart Van Assche authored
Modify scsi_find_tag() and scsi_host_find_tag() such that these functions can translate a tag generated by blk_mq_unique_tag(). Signed-off-by:
Bart Van Assche <bvanassche@acm.org> Reviewed-by:
-
Bart Van Assche authored
Allow a SCSI LLD to declare how many hardware queues it supports by setting Scsi_Host.nr_hw_queues before calling scsi_add_host(). Signed-off-by:
Sagi Grimberg <sagig@mellanox.com> Reviewed-by:
-
Bart Van Assche authored
The queuecommand() callback functions in SCSI low-level drivers need to know which hardware context has been selected by the block layer. Since this information is not available in the request structure, and since passing the hctx pointer directly to the queuecommand callback function would require modification of all SCSI LLDs, add a function to the block layer that allows to query the hardware context index. Signed-off-by:
Jens Axboe <axboe@kernel.dk> Reviewed-by:
-
Hannes Reinecke authored
There can be quite a lot of I/O error messages, even on smaller machines. So we need to ratelimit them to not overwhelm logging. Signed-off-by:
-
Hannes Reinecke authored
The .eh_abort_handler needs to return SUCCESS, FAILED, or FAST_IO_FAIL. So fixup all callers to adhere to this requirement. Reviewed-by:
-
Hannes Reinecke authored
scsi_try_to_abort_cmd() should only return SUCCESS, FAILED, or FAST_IO_FAIL. So document that in the function description and simplify the logging message. Signed-off-by:
Christoph Hellwig <hch@infradead.org> Reviewed-by:
-
Hannes Reinecke authored
The EH statistics are per host, so we should be using shost_printk() here. Signed-off-by:
-
Hannes Reinecke authored
Use the matching scope for logging messages to allow for better command tracing. Signed-off-by:
-
Hannes Reinecke authored
Simplify scsi_log_(send|completion) by externalizing scsi_mlreturn_string() and always print the command address. Signed-off-by:
-
Hannes Reinecke authored
Open-code scsi_print_result in sd.c, and cleanup logging to not print duplicate informations. Also remove the call to scsi_show_result() in ufshcd.c to be consistent with other callers of scsi_execute(). With that we can remove scsi_show_result in constants.c Signed-off-by:
-
Hannes Reinecke authored
Export functions for later use. Signed-off-by:
-
Hannes Reinecke authored
Calling scsi_print_command should not be necessary during abort; if the information is required one should enable scsi logging. Signed-off-by:
-
Hannes Reinecke authored
print_opcode_name() was only ever called with a '0' argument from LLDDs and ULDs which were _not_ supporting variable length CDBs, so the 'if' clause was never triggered. Instead we should be using the last argument to specify the cdb length to avoid accidental overflow when reading the cdb buffer. Signed-off-by:
-
Hannes Reinecke authored
Consolidate the CDB opcode lookup in scsi_opcode_sa_name(), so that we don't have to call several functions to figure out the CDB opcode string. Signed-off-by:
-
Hannes Reinecke authored
Instead of having two versions of print_opcode_name() we should be consolidating them into one version. Signed-off-by:
-
Hannes Reinecke authored
Implement a lookup array for SERVICE ACTION commands instead of hardcoding it in a large switch statement. Signed-off-by:
-
Hannes Reinecke authored
Last caller is gone, so we can remove it. Signed-off-by:
-
Hannes Reinecke authored
Convert scsi_normalize_sense() and friends to return 'bool' instead of an integer. Signed-off-by:
Yoshihiro Yunomae <yoshihiro.yunomae.ez@hitachi.com> Signed-off-by:
-
Hannes Reinecke authored
Currently we're only decoding sense extras for tape devices. And even there only for fixed format sense formats. As this is of rather limited use in the general case we should be stop trying to decode sense extras; the tape driver does its own decoding anyway. Signed-off-by:
-
Hannes Reinecke authored
If scsi_normalize_sense() fails we couldn't decode the sense buffer, and the scsi_sense_hdr fields are invalid. For those cases we should rather dump the sense buffer and not try to decode invalid fields. Signed-off-by:
-
Hannes Reinecke authored
The 53c700 driver would be using scsi_print_sense() in a debug statement, which was never compiled in. Plus the same information can get retrieved with logging. So remove it. Signed-off-by:
-
Hannes Reinecke authored
Update logging messages to use dev_printk() variants for correct device annotations. Signed-off-by:
-
Hannes Reinecke authored
fas216 returns DID_BAD_TARGET for an incomplete data transfer. The midlayer uses DID_BAD_TARGET to signal a non-existing or not reachable target. So we should rather be using DID_ERROR here. Signed-off-by:
-
Hannes Reinecke authored
Update acornscsi to use scsi_print_command() instead of the underscore version and use scmd_printk() in acornscsi_done(). This will add correct device annotations in the resulting message. And we should be using set_host_byte() for setting the final result. Signed-off-by:
-
Hannes Reinecke authored
We should be using the standard dev_printk() variants for sense code printing. [hch: remove __scsi_print_sense call in xen-scsiback, Acked by Juergen] [hch: folded bracing fix from Dan Carpenter] Signed-off-by:
-
Hannes Reinecke authored
Like scmd_printk(), but the device name is passed in as a string. Can be used by eg ULDs which do not have access to the scsi_cmnd structure. Signed-off-by:
-
Hannes Reinecke authored
Remove all uncommented debugging code and move all printk() statements over to dev_printk(). And while we're at it we should be doing a whitespace cleanup, too. Signed-off-by:
-
Hannes Reinecke authored
sd_done() was calling scsi_print_sense() for a sense code of 'NO_SENSE'. Signed-off-by:
-
Hannes Reinecke authored
Unused. Signed-off-by:
-
Markus Stockhausen authored
Some SES devices give non-unique Element Descriptors as part of the Element Descriptor diag page. Since we use these for creating sysfs entries, they need to be unique. The specification doesn't require these to be unique. Eg: $ sg_ses -p 7 /dev/sg0 FTS CORP TXS6_SAS20BPX12 0500 enclosure services device Element descriptor In diagnostic page: generation code: 0x0 element descriptor by type list Element type: Array device, subenclosure id: 0 Overall descriptor: ArrayDevicesInSubEnclsr0 Element 1 descriptor: ArrayDevice00 Element 2 descriptor: ArrayDevice01 Element 3 descriptor: ArrayDevice02 Element 4 descriptor: ArrayDevice03 Element 5 descriptor: ArrayDevice03 Element 6 descriptor: ArrayDevice03 Element 7 descriptor: ArrayDevice03 Element 8 descriptor: ArrayDevice03 Element 9 descriptor: ArrayDevice03 Element 10 descriptor: ArrayDevice03 Element 11 descriptor: ArrayDevice03 Element 12 descriptor: ArrayDevice03 Based on http://thread.gmane.org/gmane.linux.scsi/69289. This version implements James' ideas about the naming convention Signed-off-by:Markus Stockhausen <stockhausen@collogia.de> Acked-by:
James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by:
-
Laurence Oberman authored
This patch adds a debug_flag parameter that can be set on module load, and allows the DEBUG facility without a module recompile. Note that now DEBUG 1 is the default with this patch. Usage: modprobe st debug_flag=1 Signed-off-by:
Laurence Oberman <loberman@redhat.com> Acked-by:
Kai M??kisara <kai.makisara@kolumbus.fi> Signed-off-by:
-
Douglas Gilbert authored
Further to a January 2013 thread titled: "[PATCH] SG_SCSI_RESET ioctl should only perform requested operation" by Jeremy Linton a patch (v3) is presented that expands the existing ioctl to include "no_escalate" versions to the existing resets. This requires no changes to SCSI low level drivers (LLDs); it adds several more finely tuned reset options to the user space. For example: /* This call remains the same, with the same escalating semantics * if the device (LU) reset fail. That is: on failure to try a * target reset and if that fails, try a bus reset, and if that fails * try a host (i.e. LLD) reset. */ val = SG_SCSI_RESET_DEVICE; res = ioctl(<sg_or_block_fd>, SG_SCSI_RESET, &val); /* What follows is a new option introduced by this patch series. Only * a device reset is attempted. If that fails then an appropriate * error code is provided. N.B. There is no reset escalation. */ val = SG_SCSI_RESET_DEVICE | SG_SCSI_RESET_NO_ESCALATE; res = ioctl(<sg_or_block_fd>, SG_SCSI_RESET, &val); Signed-off-by:Douglas Gilbert <dgilbert@interlog.com> Reviewed-by:
Jeremy Linton <jlinton@tributary.com> Reviewed-by:
-
Christoph Hellwig authored
As sparse correctly pointed out, scsi_partsize should use get_unaligned_le32 to read PC partition tables from disk, as they are little endian. The result of this bug is that we returned incorrect geometries on big endian systems when using the scsicam variant. Which probably doesn't matter as only old x86 systems every cared about the geometry. Signed-off-by:
-
