SCSI updates for 2.6.14
PATCH 1/1 SCSI updates for 2.6.14
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
The short changelog is
Alan Stern:
o sd: Fix refcounting
Alex Aizman:
o open-iscsi/linux-iscsi-5 Initiator: Initiator code
o open-iscsi/linux-iscsi-5 Initiator: Kconfig update
o open-iscsi/linux-iscsi-5 Initiator: Transport class update for iSCSI
o open-iscsi/linux-iscsi-5 Initiator: Initiator header
o open-iscsi/linux-iscsi-5 Initiator: Header files
o open-iscsi/linux-iscsi-5 Initiator: Makefile Updates
Andrew Vasquez:
o Add an 'Issue LIP' device attribute in fc_transport class
o qla2xxx: fix compile warning
o qla2xxx: remove eh_active checks in qla2xxx error handling
o qla2xxx: Update license
o qla2xxx: Use midlayer's int_to_scsilun() function
o qla2xxx: Add support to dynamically enable/disable ZIO
o qla2xxx: Correct issue where fcport is prematurely marked DEAD
Christoph Hellwig:
o mptsas: white space fixes and version bump
o mptsas: add support for PHY resets
o sas: add support for PHY resets
o sas: add flag for locally attached PHYs
o mptsas: support link error attributes
o scsi_transport_sas: support link error attributes
Guennadi Liakhovetski:
o dc395x: atomic_kmap for PIO
Jack Hammer:
o ips: Fix initialization bug with kdump
o ips: Fix up for correct scatter/gather processing
James Bottomley:
o Fix ips.c compile
o Merge by Hand
o lpfc: Fix for "command completion for iotax x?? not found"
o remove broken driver cpqfc
o fix up mismerge in osst
o qla2xxx: put back label erroneously removed by eh_active patch
o qla2xxx: fix unnecessary activation of blk tag queue
o move the mid-layer printk's over to shost/starget/sdev_printk
o avoid overflows in disk size calculations
o qla2xxx: Resync with latest released ISP23xx/63xx firmware --
3.03.18
o Merge HEAD from ../scsi-misc-2.6-old
o aacraid: Use DMA mask defines
o move iscsi to a better place in Kconfig
o iscsi_tcp: make iscsi compile again after recent netlink changes
o Merge HEAD from ../scsi-iscsi-2.6
o Merge ../linux-2.6
o iscsi: fix 64 bit compile warning
James Smart:
o lpfc: Change version number to 8.1.0
o lpfc 8.1.0 : Add owner field to struct pci_driver
o lpfc: Fix eh_ return codes for commands
o lpfc: Remove unneeded IOCB_t * cast
o lpfc: Adjust lpfc_scsi_buf allocation
o lpfc: Replace lpfc_sli_issue_iocb_wait_high_priority
o lpfc: Remove RPI hash from the driver
o lpfc: Restore HEX safe bahavior of the sysfs xxx_store functions
o lpfc: Fix for "Unknown IOCB command Data: x0 x3 x0 x0"
o lpfc: Fix comments for nodev_tmo
o lpfc: Add range checking for attributes passed as options at load
time
o lpfc: Return -EINVAL, -EPERM, and -EIO instead of 0 from sysfs
callbacks
o lpfc: Update to Emulex hba model names
o lpfc: Cleanup code in lpfc_get_stats()
o update fc_transport for removal of block/unblock functions
Jayachandran C:
o sr: remove dead code
o Fix issue reported by coverity in drivers/scsi/scsi_ioctl.c
Jeff Garzik:
o use scmd_id(), scmd_channel() throughout code
o use sfoo_printk() in drivers
o use {sdev,scmd,starget,shost}_printk in generic code
o introduce sfoo_printk, sfoo_id, sfoo_channel helpers
o kill unused scsi_scan_single_target()
Mark Haverkamp:
o aacraid: Newer adapter communication iterface support
o aacraid: remove compiler warning
o aacraid: fix struct element cpu order
o aacraid: fix inquiry page
Matthew Wilcox:
o ncr53c8xx: Cleanup namespace collision with ktimers
Mike Christie:
o iscsi: add module version
o iscsi: fix nop-in handling
o iscsi: rename some proto defs
o iscsi: add newline to sysfs output
o iscsi: fix ahs len
o iscsi: update some iscsi proto defs
o iscsi: handle nonlinear skbs
o iscsi: preemt fix and cleanup
o iscsi: nodelay fix
Ralf Bächle:
o sgiwd93: small fixes
o dec_esp: Use the right address space macro
o dec_esp: Use physical addresses
o dec_esp: Fix mapping of ESP
And the diffstat (large qla firmware update again, I'm afraid):
b/Documentation/scsi/LICENSE.qla2xxx | 45 b/drivers/message/fusion/mptbase.h | 16 b/drivers/message/fusion/mptsas.c | 323 b/drivers/scsi/3w-9xxx.c | 4 b/drivers/scsi/3w-xxxx.c | 4 b/drivers/scsi/53c700.c | 89 b/drivers/scsi/53c700.h | 8 b/drivers/scsi/Kconfig | 39 b/drivers/scsi/Makefile | 4 b/drivers/scsi/NCR5380.c | 21 b/drivers/scsi/NCR53C9x.c | 16 b/drivers/scsi/NCR53c406a.c | 2 b/drivers/scsi/a100u2w.c | 2 b/drivers/scsi/aacraid/README | 2 b/drivers/scsi/aacraid/TODO | 1 b/drivers/scsi/aacraid/aachba.c | 86 b/drivers/scsi/aacraid/aacraid.h | 27 b/drivers/scsi/aacraid/commctrl.c | 6 b/drivers/scsi/aacraid/comminit.c | 25 b/drivers/scsi/aacraid/commsup.c | 86 b/drivers/scsi/aacraid/dpcsup.c | 115 b/drivers/scsi/aacraid/linit.c | 37 b/drivers/scsi/aacraid/rkt.c | 172 b/drivers/scsi/aacraid/rx.c | 157 b/drivers/scsi/aacraid/sa.c | 38 b/drivers/scsi/aha152x.c | 3 b/drivers/scsi/aha1542.c | 3 b/drivers/scsi/aha1740.c | 2 b/drivers/scsi/aic7xxx/aic79xx_osm.c | 39 b/drivers/scsi/aic7xxx/aic7xxx_osm.c | 32 b/drivers/scsi/atp870u.c | 29 b/drivers/scsi/ch.c | 4 b/drivers/scsi/constants.c | 5 b/drivers/scsi/dc395x.c | 13 b/drivers/scsi/dec_esp.c | 21 b/drivers/scsi/eata.c | 56 b/drivers/scsi/eata_pio.c | 21 b/drivers/scsi/fd_mcs.c | 4 b/drivers/scsi/fdomain.c | 4 b/drivers/scsi/hosts.c | 10 b/drivers/scsi/ibmmca.c | 5 b/drivers/scsi/ide-scsi.c | 2 b/drivers/scsi/imm.c | 2 b/drivers/scsi/in2000.c | 2 b/drivers/scsi/ipr.h | 5 b/drivers/scsi/ips.c | 153 b/drivers/scsi/iscsi_tcp.c | 3642 ++++++++ b/drivers/scsi/iscsi_tcp.h | 322 b/drivers/scsi/lpfc/lpfc.h | 4 b/drivers/scsi/lpfc/lpfc_attr.c | 201 b/drivers/scsi/lpfc/lpfc_crtn.h | 21 b/drivers/scsi/lpfc/lpfc_ct.c | 14 b/drivers/scsi/lpfc/lpfc_disc.h | 1 b/drivers/scsi/lpfc/lpfc_els.c | 36 b/drivers/scsi/lpfc/lpfc_hbadisc.c | 124 b/drivers/scsi/lpfc/lpfc_init.c | 69 b/drivers/scsi/lpfc/lpfc_mbox.c | 7 b/drivers/scsi/lpfc/lpfc_nportdisc.c | 21 b/drivers/scsi/lpfc/lpfc_scsi.c | 399 b/drivers/scsi/lpfc/lpfc_sli.c | 525 - b/drivers/scsi/lpfc/lpfc_sli.h | 22 b/drivers/scsi/lpfc/lpfc_version.h | 2 b/drivers/scsi/megaraid/megaraid_sas.c | 5 b/drivers/scsi/ncr53c8xx.c | 30 b/drivers/scsi/nsp32.c | 8 b/drivers/scsi/osst.c | 6 b/drivers/scsi/pcmcia/nsp_cs.c | 10 b/drivers/scsi/pcmcia/sym53c500_cs.c | 2 b/drivers/scsi/ppa.c | 2 b/drivers/scsi/psi240i.c | 2 b/drivers/scsi/qla2xxx/ql2100.c | 9 b/drivers/scsi/qla2xxx/ql2100_fw.c | 22 b/drivers/scsi/qla2xxx/ql2200.c | 9 b/drivers/scsi/qla2xxx/ql2200_fw.c | 22 b/drivers/scsi/qla2xxx/ql2300.c | 9 b/drivers/scsi/qla2xxx/ql2300_fw.c |13851 +++++++++++++++---------------- b/drivers/scsi/qla2xxx/ql2322.c | 7 b/drivers/scsi/qla2xxx/ql2322_fw.c |14483 ++++++++++++++++----------------- b/drivers/scsi/qla2xxx/ql6312.c | 7 b/drivers/scsi/qla2xxx/ql6312_fw.c |13882 +++++++++++++++---------------- b/drivers/scsi/qla2xxx/qla_attr.c | 111 b/drivers/scsi/qla2xxx/qla_dbg.c | 18 b/drivers/scsi/qla2xxx/qla_dbg.h | 23 b/drivers/scsi/qla2xxx/qla_def.h | 36 b/drivers/scsi/qla2xxx/qla_fw.h | 30 b/drivers/scsi/qla2xxx/qla_gbl.h | 31 b/drivers/scsi/qla2xxx/qla_gs.c | 18 b/drivers/scsi/qla2xxx/qla_init.c | 99 b/drivers/scsi/qla2xxx/qla_inline.h | 19 b/drivers/scsi/qla2xxx/qla_iocb.c | 63 b/drivers/scsi/qla2xxx/qla_isr.c | 25 b/drivers/scsi/qla2xxx/qla_mbx.c | 18 b/drivers/scsi/qla2xxx/qla_os.c | 112 b/drivers/scsi/qla2xxx/qla_rscn.c | 23 b/drivers/scsi/qla2xxx/qla_settings.h | 22 b/drivers/scsi/qla2xxx/qla_sup.c | 23 b/drivers/scsi/qla2xxx/qla_version.h | 22 b/drivers/scsi/qlogicfas408.c | 4 b/drivers/scsi/scsi.c | 26 b/drivers/scsi/scsi_debug.c | 17 b/drivers/scsi/scsi_error.c | 48 b/drivers/scsi/scsi_ioctl.c | 14 b/drivers/scsi/scsi_lib.c | 37 b/drivers/scsi/scsi_scan.c | 37 b/drivers/scsi/scsi_transport_fc.c | 476 - b/drivers/scsi/scsi_transport_iscsi.c | 1394 ++- b/drivers/scsi/scsi_transport_sas.c | 75 b/drivers/scsi/scsi_transport_spi.c | 24 b/drivers/scsi/sd.c | 114 b/drivers/scsi/sg.c | 14 b/drivers/scsi/sgiwd93.c | 5 b/drivers/scsi/sr.c | 17 b/drivers/scsi/st.c | 9 b/drivers/scsi/sym53c416.c | 2 b/drivers/scsi/sym53c8xx_defs.h | 13 b/drivers/scsi/tmscsim.c | 12 b/drivers/scsi/u14-34f.c | 47 b/include/scsi/iscsi_if.h | 245 b/include/scsi/iscsi_proto.h | 589 + b/include/scsi/scsi_device.h | 22 b/include/scsi/scsi_host.h | 6 b/include/scsi/scsi_transport_fc.h | 33 b/include/scsi/scsi_transport_iscsi.h | 202 b/include/scsi/scsi_transport_sas.h | 27 drivers/scsi/cpqfcTS.h | 19 drivers/scsi/cpqfcTSchip.h | 238 drivers/scsi/cpqfcTScontrol.c | 2231 ----- drivers/scsi/cpqfcTSi2c.c | 493 - drivers/scsi/cpqfcTSinit.c | 2096 ---- drivers/scsi/cpqfcTSioctl.h | 94 drivers/scsi/cpqfcTSstructs.h | 1530 --- drivers/scsi/cpqfcTStrigger.c | 33 drivers/scsi/cpqfcTStrigger.h | 8 drivers/scsi/cpqfcTSworker.c | 6516 -------------- drivers/scsi/sgiwd93.h | 24 135 files changed, 29959 insertions(+), 37041 deletions(-)
James
Quoted 4 lines · James Bottomley
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
No it's not.
master$ git-cat-file -t HEAD
gives
error: unable to find 39b7f1e25a412b0ef31e516cfc2fa4f40235f263 fatal: git-cat-file HEAD: bad file
and the reason seems to be:
master$ ll objects/39/
ls: objects/39/: Permission denied
Hmmph.
Linus
Quoted 16 lines · Linus Torvalds
Quoted 4 lines · James Bottomley
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
No it's not.
master$ git-cat-file -t HEAD
gives
error: unable to find 39b7f1e25a412b0ef31e516cfc2fa4f40235f263 fatal: git-cat-file HEAD: bad file
Quoted 7 lines
and the reason seems to be:
master$ ll objects/39/
ls: objects/39/: Permission denied
Hmmph.
Heh, they're all drwx--S---
I've had no end of strange trouble like this since I moved from my own version of git to the one installed on hera.
I think I've found and changed all the directories and verified the individual object permissions. Could you try again?
Thanks,
James
Quoted 7 lines · James Bottomley
Heh, they're all drwx--S---
I've had no end of strange trouble like this since I moved from my own version of git to the one installed on hera.
I think I've found and changed all the directories and verified the individual object permissions. Could you try again?
With the latest git...
Quoted 3 lines
[jgarzik@pretzel scsi-misc-2.6]$ git pull
fatal: unexpected EOF
Fetch failure: master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6.git
Do you have standard permissions (chmod -R og+rX) on your repo, and, are you using rsync to push to kernel.org?
I've attached my rsync-based push script. git people seem to dislike rsync, but this tends to work every time, for all users.
Jeff
Quoted 5 lines · Jeff Garzik
Do you have standard permissions (chmod -R og+rX) on your repo, and, are you using rsync to push to kernel.org?
I've attached my rsync-based push script. git people seem to dislike rsync, but this tends to work every time, for all users.
No, I push to my scsi-misc-2.6 repository and then clone that on hera for linus. the scsi-misc-2.6 permissions are fine, it was the clone -l on hera that caused the problems.
James
Quoted 10 lines · James Bottomley
Quoted 3 lines · Jeff Garzik
Do you have standard permissions (chmod -R og+rX) on your repo, and, are you using rsync to push to kernel.org?
I've attached my rsync-based push script. git people seem to dislike rsync, but this tends to work every time, for all users.
No, I push to my scsi-misc-2.6 repository and then clone that on hera for linus. the scsi-misc-2.6 permissions are fine, it was the clone -l on hera that caused the problems.
Well, _something_ is clearly screwed up in your scsi-misc-2.6 repo on master.kernel.org, as its the only repo that I cannot pull from (last successful pull was Oct 31, before I left for Portland).
Jeff
A message that is not in the archive (sent off-list or not yet received): [email protected]
A message that is not in the archive (sent off-list or not yet received): 1131216913.3614.15.camel@mulgrave
Quoted 6 lines · James Bottomley
Quoted 2 lines · Linus Torvalds
James: what's your umask on master? Maybe you've set it to 0077, from the default 0002?
I set it to 0022 ... could that be the problem?
No, 0022 should still be perfectly fine, and should result in drwxr-xr-x, not your strange things.
Looking at things, I can't imagine it being anything else than a cpio bug (or misfeature - I don't know where cpio _could_ get the permissions for the intermediate directories from). But that implies that the system git shouldn't have mattered, and if you used "git clone -l" with your old git it should have done exactly the same thing.
Junio: I just confirmed this. Do a
git clone -l src dst
with any git, and you'll end up with
... drwx------ 2 torvalds torvalds 4096 Nov 5 11:22 de drwx------ 2 torvalds torvalds 4096 Nov 5 11:22 df drwx------ 2 torvalds torvalds 4096 Nov 5 11:22 e0 ... drwx------ 2 torvalds torvalds 4096 Nov 5 11:22 ff drwxrwxr-x 2 torvalds torvalds 4096 Nov 5 11:22 info drwxrwxr-x 2 torvalds torvalds 4096 Nov 5 11:22 pack
regardless of umask.
Happens with both cpio 2.5 and 2.6.
I _think_ it's this:
...
find objects -type f -print |
cpio -puamd$l "$D/.git/" || exit 1
...and I suspect that if you remove the "-type f" it will just work (because it will get the directory permissions from the source). But cpio is black magic to me (very un-unixy feel to it, I'd guess it comes from some odd other pre-unix OS).
Linus
Quoted 6 lines · James Bottomley
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
James, I note that my favorite masterpiece of ibmvscsis.c is not currently included. Does that fall into part of the "flood"?
Thanks
Dave B
Quoted 10 lines · Dave C Boutcher
Quoted 6 lines · James Bottomley
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
James, I note that my favorite masterpiece of
ibmvscsis.cis not currently included. Does that fall into part of the "flood"?
it shouldn't go in. we already have a version using the proper target framework, that should be imroved upon to get production quality and then merged.
Quoted 7 lines · Christoph Hellwig
Quoted 6 lines · Dave C Boutcher
Quoted 5 lines · James Bottomley
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
This patch is available from:
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
James, I note that my favorite masterpiece of
ibmvscsis.cis not currently included. Does that fall into part of the "flood"?it shouldn't go in. we already have a version using the proper target framework, that should be imroved upon to get production quality and then merged.
My only problem with that is the amount of time it is going to keep the driver out of the tree. I'm happy to commit to merging to a converged infrastructure as soon as it is available.
Signature · 1 line
Dave Boutcher
Quoted 16 lines · Dave C Boutcher
Quoted 14 lines · Christoph Hellwig
Quoted 11 lines · Dave C Boutcher
Quoted 6 lines · James Bottomley
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
James, I note that my favorite masterpiece of
ibmvscsis.cis not currently included. Does that fall into part of the "flood"?
Quoted 9 lines
Quoted 3 lines
it shouldn't go in. we already have a version using the proper target framework, that should be imroved upon to get production quality and then merged.
My only problem with that is the amount of time it is going to keep the driver out of the tree. I'm happy to commit to merging to a converged infrastructure as soon as it is available.
Your driver needs some work as it is. We are trying to kill scsi_request usage, your bio segments are wrong if cluster is not used (iscsi in mainline does not use clustering), and you could use bio_map_kern(). There are others, but I did not review it so closely.
Quoted 27 lines · Mike Christie
Quoted 25 lines · Dave C Boutcher
Quoted 20 lines · Christoph Hellwig
Quoted 5 lines · Dave C Boutcher
Quoted 3 lines · James Bottomley
This is Part I of the pending SCSI patches (I still need to go over all the flood that came in after 2.6.14 was declared).
Quoted 8 lines
Quoted 4 lines
This patch is available from:
master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6.git
James, I note that my favorite masterpiece of
ibmvscsis.cis not currently included. Does that fall into part of the "flood"?it shouldn't go in. we already have a version using the proper target framework, that should be imroved upon to get production quality and then merged.
Quoted 6 lines
Quoted 4 lines
My only problem with that is the amount of time it is going to keep the driver out of the tree. I'm happy to commit to merging to a converged infrastructure as soon as it is available.
Your driver needs some work as it is. We are trying to kill
scsi_requestusage, your bio segments are wrong if cluster is not used (iscsi in mainline does not use clustering), and you could usebio_map_kern(). There are others, but I did not review it so closely.
oh yeah we have a function to do the lun to integer stuff I beleive.