litmus-rt.git/drivers/scsi/aacraid, branch wip-default-clustering

drivers: scsi: use newly introduced hex_to_bin() method

2010-08-11T15:59:02+00:00

Signed-off-by: Andy Shevchenko 
Cc: Adaptec OEM Raid Solutions 
Cc: "James E.J. Bottomley" 
Cc: James Smart 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[SCSI] aacraid: Do not set DMA mask to 32 bit first if adapter only supports 31

2010-07-28T14:06:19+00:00

Signed-off-by: Rolf Eike Beer 
Acked-by: Achim Leubner 
Signed-off-by: James Bottomley

Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6

2010-05-27T17:28:11+00:00

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6:
  [SCSI] fix race in scsi_target_reap
  [SCSI] aacraid: Eliminate use after free
  [SCSI] arcmsr: Support HW reset for EH and polling scheme for scsi device
  [SCSI] bfa: fix system crash when reading sysfs fc_host statistics
  [SCSI] iscsi_tcp: remove sk_sleep check
  [SCSI] ipr: improve interrupt service routine performance
  [SCSI] ipr: set the data list length in the request control block
  [SCSI] ipr: fix a register read to use the correct address for 64 bit adapters
  [SCSI] ipr: include the resource path in the IOA status area structure
  [SCSI] ipr: implement fixes for 64 bit adapter support
  [SCSI] be2iscsi: correct return value in mgmt_invalidate_icds()

[SCSI] aacraid: Eliminate use after free

2010-05-25T15:59:07+00:00

The debugging code using the freed structure is moved before the kfree.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// 
@free@
expression E;
position p;
@@
kfree@p(E)

@@
expression free.E, subE<=free.E, E1;
position free.p;
@@

  kfree@p(E)
  ...
(
  subE = E1
|
* E
)
// 

Signed-off-by: Julia Lawall 
Signed-off-by: James Bottomley

Merge branch 'bkl/ioctl' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing

2010-05-24T15:01:10+00:00

* 'bkl/ioctl' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing:
  uml: Pushdown the bkl from harddog_kern ioctl
  sunrpc: Pushdown the bkl from sunrpc cache ioctl
  sunrpc: Pushdown the bkl from ioctl
  autofs4: Pushdown the bkl from ioctl
  uml: Convert to unlocked_ioctls to remove implicit BKL
  ncpfs: BKL ioctl pushdown
  coda: Clean-up whitespace problems in pioctl.c
  coda: BKL ioctl pushdown
  drivers: Push down BKL into various drivers
  isdn: Push down BKL into ioctl functions
  scsi: Push down BKL into ioctl functions
  dvb: Push down BKL into ioctl functions
  smbfs: Push down BKL into ioctl function
  coda/psdev: Remove BKL from ioctl function
  um/mmapper: Remove BKL usage
  sn_hwperf: Kill BKL usage
  hfsplus: Push down BKL into ioctl function

scsi: Push down BKL into ioctl functions

2010-05-17T03:27:04+00:00

Push down the bkl into ioctl functions on the scsi layer.

[jkacur: Forward declaration missing ';'.
Conflicting declaraction in megaraid.h changed
Fixed missing inodes declarations]

Signed-off-by: Arnd Bergmann 
Signed-off-by: John Kacur 
Signed-off-by: Frederic Weisbecker

[SCSI] aacraid: add an ifdef'd device delete case instead of taking the device offline

2010-05-17T02:58:27+00:00

Problem description:
--------------------

The problem reported by one of the customer was when a logical array
is deleted(from the SDK, from the GUI, from arcconf) then the
corresponding physical device (/dev/sdb, for example) is not removed
from the Linux namespace. So you end up with a "dead" device
entry. And some of the linux tools go slightly wonky.

Solution:
---------

Based on the notification from FW, the driver calls
"scsi_remove_device" for the DELETED drive. This call not only informs
the scsi device status to the SCSI mid layer and also it will remove
corresponding scsi device entries from the Linux sysfs.

Signed-off-by: Mahesh Rajashekhara 
Signed-off-by: James Bottomley

[SCSI] aacraid: prohibit access to array container space

2010-05-17T02:49:46+00:00

Problem description:
--------------------

The issue reported by one of the customer was able to read LBA beyond
the array reported size with "sg_read" utility. If N is the last block
address reported, then should not be able to read past N,
i.e. N+1. But in their case, reported last LBA=143134719.  So should
not have been able to read with LBA=143134720, but it is read without
failure, which means reported size to the OS is not correct and is
less than the actual last block address.

Solution:
---------

Firmware layer exposes lesser container capacity than the actual
one. It exposes [Actual size - Spitfire space(10MB)] to the OS, IO's
to the 10MB should be prohibited from the Linux driver. Driver checks
LBA boundary, if its greater than the array reported size then sets
sensekey to HARDWARE_ERROR and sends the notification to the MID
layer.

Signed-off-by: Mahesh Rajashekhara 
Signed-off-by: James Bottomley

[SCSI] aacraid: add support for handling ATA pass-through commands.

2010-05-17T02:40:40+00:00

There are two conditions for ATA pass thru command that falls into
'SRB_STATUS_ERROR' condition.

1. When the "CC" bit is set by the host in ATA pass-through CDB

   - Even for the successful completion, SCSI target shall generate
     check condition.

   - Driver returns a result code of SAM_STAT_CHECK_CONDITION, with a
     driver byte of DID_OK to the mid layer.

     Below is the snippet of existing code which fills a result code
     of SAM_STAT_CHECK_CONDITION:

	***********************************
	        if (le32_to_cpu(srbreply->scsi_status) == SAM_STAT_CHECK_CONDITION) {
                		int len;
	               		scsicmd->result |= SAM_STAT_CHECK_CONDITION;
			..........
	************************************

2. When the "CC" bit is reset by the host and if SCSI target generates
   a check condition when an error occurs.

   - Driver returns a result code of SAM_STAT_CHECK_CONDITION, with a
     driver byte of DID_ERROR to the mid layer.

Signed-off-by: Mahesh Rajashekhara 
Signed-off-by: James Bottomley

[SCSI] aacraid: expose physical devices for models with newer firmware

2010-05-17T02:35:54+00:00

The default driver setting is "expose_physicals=0", which means raw
physical drives are not exposed to OS.  If the user wants to expose
connected physical drives, enable "expose_physicals" module parameter.
With the new JBOD firmware, physical drives are not available for
"expose_physicals>0".  In function "aac_expose_phy_device", modified
to reset the appropriate bit in the first byte of inquiry data.  This
fix exposes the connected physical drives.

Signed-off-by: Mahesh Rajashekhara 
Signed-off-by: James Bottomley