Ingo Molnar [Sat, 14 Jan 2006 21:21:30 +0000 (13:21 -0800)]
[PATCH] mark several functions __always_inline
Arjan van de Ven <arjan@infradead.org>
Mark a number of functions as 'must inline'. The functions affected by this
patch need to be inlined because they use knowledge that their arguments are
constant so that most of the function optimizes away. At this point this
patch does not change behavior, it's for documentation only (and for future
patches in the inline series)
Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ingo Molnar [Sat, 14 Jan 2006 21:21:29 +0000 (13:21 -0800)]
[PATCH] enable unit-at-a-time optimisations for gcc4
Allow gcc4 compilers to optimize unit-at-a-time.
This flag enables gcc to "see" the entire C file before making optimisation
decisions such as inline, which results in gcc making better decisions. One
of the immediate effects of this is that static functions that are used only
once now get inlined.
gcc 3.4 has this flag as well, however gcc 3.x have a problem with inlining
and stacks and as a result, enabling this flag there would cause excessive and
unacceptable stack use. This problem is fixed in the gcc 4.x series. The
x86-64 architecture already enables this feature so it's well tested already.
Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Arjan van de Ven <arjan@infradead.org> Acked-by: Jeff Garzik <jgarzik@pobox.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ingo Molnar [Sat, 14 Jan 2006 21:21:28 +0000 (13:21 -0800)]
[PATCH] Make __always_inline actually force always inlining
This patch is the first in a series that tries to optimize the kernel in terms
of size (and thus cache behavior, both cpu and pagecache).
This first patch changes __always_inline to be a forced inline instead of the
"regular" inline it was on everything except alpha. This forced inline
matches the intention of the define better as a matter of documentation.
There is no change in behavior by this patch, since "inline" currently is
mapped to a forced inline anyway.
Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
No need for a file argument. If we'd really need it it's in vma->vm_file
already. gbefb and sgivwfb used to set vma->vm_file to the file argument, but
the kernel alrady did that.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Antonino Daplas <adaplas@pol.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David Vrabel [Sat, 14 Jan 2006 21:21:23 +0000 (13:21 -0800)]
[PATCH] gx1fb: (try to) play nicer with various BIOSes
Seems that the CS5530A chip used in Geode GX1 systems has some crazy feature
that causes SMI traps when accessing the PCI configuration space of the video
device. Various GX1 BIOSes seem to use this 'feature' to hide the real BARs
of the device. This patch disables these traps (in an early PCI fixup) so
that Linux sees the real, physical BARs and not the virtual ones provided by
the BIOS.
This should allow the GX1 framebuffer driver to work on more systems that have
different BIOSes as the driver no longer guesses at what the virtual BARs
mean.
I'm not entirely sure it the correct solution as I can neither test regular
VGA console nor the X's 'cyrix' video driver so there might be some breakage
there -- probably best to get some more testers before applying it.
[PATCH] neofb: take existing display configuration as default
On a Dell Latitude CPi-A I noticed a strangeness wrt. the handling of an
external monitor by the neomagic framebuffer driver, namely when the laptop is
docked in a C/Dock II with the lid shut.
A cold boot would result in the BIOS configuring the video chip to use the
"external monitor only" mode, yet neofb would default to "internal LCD only".
An attempt for a quick fix by using the Fn-F8 keystroke to toggle the display
combination modes resulted in a reproductible hard lock, powering down being
the only solution.
The attached patch makes neofb probe the register for the current display
mode, using that value as a default if nothing was specified as kernel/module
parameter.
Randy Dunlap [Sat, 14 Jan 2006 21:21:19 +0000 (13:21 -0800)]
[PATCH] nlm kernel-parameters update
Add 2 lockd kernel parameters and spell 2 others correctly.
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net> Cc: Neil Brown <neilb@cse.unsw.edu.au> Cc: Olaf Kirch <okir@suse.de> Cc: <buraphalinuxserver@gmail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
cs89x0 inconsistently used 'int' and 'u32' for device register data.
As the cs89x0 is a 16-bit chip, change the I/O accessors over to 'u16'.
(Spotted by Deepak Saxena.)
Abhay Salunke [Sat, 14 Jan 2006 21:21:14 +0000 (13:21 -0800)]
[PATCH] dell_rbu: fix Bug 5854
This fixes http://bugzilla.kernel.org/show_bug.cgi?id=5854
Root cause:
The dell_rbu driver creates entries in /sys/class/firmware/dell_rbu/ by
calling request_firmware_nowait (without hotplug ) this function inturn
starts a kernel thread which creates the entries in
/sys/class/firmware/dell_rbu/loading , data and the thread waits on the
user action to return control back to the callback fucntion of dell_rbu.
The thread calls wait_on_completion which puts it in a D state until the
user action happens. If there is no user action happening the load average
goes up as the thread D state is taken in to account. Also after
downloading the BIOS image the enrties go away momentarily but they are
recreated from the callback function in dell_rbu. This causes the thread
to get recreated causing the load average to permenently stay around 1.
Fix:
The dell_rbu also creates the entry
/sys/devices/platform/dell_rbu/image_type at driver load time. The image
type by default is mono if required the user can echo packet to image_type
to make the BIOS update mechanism using packets. Also by echoing init in
to image_type the /sys/class/firmware/dell_rbu entries can be created.
The driver code was changed to not create /sys/class/firmware/dell_rbu
entries during load time, and also to not create the above entries from the
callback function. The entries are only created by echoing init to
/sys/devices/platform/dell_rbu/image_type The user now needs to create the
entries to download the image monolithic or packet. This fixes the issue
since the kernel thread only is created when ever the user is ready to
download the BIOS image; this minimizes the life span of the kernel thread
and the load average goes back to normal.
Signed off by Abhay Salunke <abhay_salunke@dell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Jackson [Sat, 14 Jan 2006 21:21:06 +0000 (13:21 -0800)]
[PATCH] cpuset oom lock fix
The problem, reported in:
http://bugzilla.kernel.org/show_bug.cgi?id=5859
and by various other email messages and lkml posts is that the cpuset hook
in the oom (out of memory) code can try to take a cpuset semaphore while
holding the tasklist_lock (a spinlock).
One must not sleep while holding a spinlock.
The fix seems easy enough - move the cpuset semaphore region outside the
tasklist_lock region.
This required a few lines of mechanism to implement. The oom code where
the locking needs to be changed does not have access to the cpuset locks,
which are internal to kernel/cpuset.c only. So I provided a couple more
cpuset interface routines, available to the rest of the kernel, which
simple take and drop the lock needed here (cpusets callback_sem).
Signed-off-by: Paul Jackson <pj@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Cornelia Huck [Sat, 14 Jan 2006 21:21:03 +0000 (13:21 -0800)]
[PATCH] s390: chps[] array too short
The chps[] array in struct channel_subsystem is one too short; therefore the
code doesn't realize the chpid ff is already known. When several devices on
chpid ff become available, the message "new_channel_path: could not register
ff" is displayed for every device but the first one.
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
finish_arch_switch needs to update the user cpu time as well, not just the
system cpu time. Otherwise the partial user cpu time of a process that is
stored in the lowcore will be (mis-)accounted to the next process.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Heiko Carstens [Sat, 14 Jan 2006 21:20:57 +0000 (13:20 -0800)]
[PATCH] s390: show_task oops
The show_task function walks the kernel stack backchain of processes assuming
that the processes are not running. Since this assumption is not correct
walking the backchain can lead to an addressing exception and therefore to a
kernel hang. So prevent the kernel hang (you still get incorrect results)
verity that all read accesses are within the bounds of the kernel stack before
performing them.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jan Glauber [Sat, 14 Jan 2006 21:20:53 +0000 (13:20 -0800)]
[PATCH] s390: des crypto code speedup
Provide ECB and CBC encrypt / decrypt functions to crypto API to speed up our
hardware accelerated DES implementation. This new functions allow the crypto
API to call ECB / CBC directly with large blocks in difference to the old
functions that were calles with algorithm block size (8 bytes for DES).
This is up to factor 10 faster than our old hardware implementation :)
Signed-off-by: Jan Glauber <jan.glauber@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Chuck Ebbert [Sat, 14 Jan 2006 21:20:52 +0000 (13:20 -0800)]
[PATCH] i386: fix stack dump loglevel
Recent changes caused part of stack traces from SysRq-T to print at
KERN_EMERG loglevel. Also, parts of stack dump during oops were failing to
print at that level when they should.
Randy Dunlap [Sat, 14 Jan 2006 21:20:51 +0000 (13:20 -0800)]
[PATCH] i386: put HOTPLUG_CPU under Processor type, not Bus options
Move the HOTPLUG_CPU option under "Processor type" instead of under "Bus
options". This makes it the same for i386 as most other processor types
(arm, ia64, parisc, ppc, s390, & x86_64; but not for powerpc). Besides, it
takes me too long to find it under Bus options. I can't be the only person
who has trouble finding it.
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Kumar Gala [Sat, 14 Jan 2006 21:20:50 +0000 (13:20 -0800)]
[PATCH] powerpc: Add support for the MPC83xx watchdog
Add support for the PowerPC MPC83xx watchdog. The MPC83xx has a simple
watchdog that once enabled it can not be stopped, has some simple timeout
range selection, and the ability to either reset the processor or take a
machine check.
Signed-off-by: Dave Updegraff <dave@cray.org> Signed-off-by: Kumar Gala <galak@kernel.crashing.org> Cc: Wim Van Sebroeck <wim@iguana.be> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Robin Holt [Sat, 14 Jan 2006 21:20:48 +0000 (13:20 -0800)]
[PATCH] Add tmpfs options for memory placement policies
Anything that writes into a tmpfs filesystem is liable to disproportionately
decrease the available memory on a particular node. Since there's no telling
what sort of application (e.g. dd/cp/cat) might be dropping large files
there, this lets the admin choose the appropriate default behavior for their
site's situation.
Introduce a tmpfs mount option which allows specifying a memory policy and
a second option to specify the nodelist for that policy. With the default
policy, tmpfs will behave as it does today. This patch adds support for
preferred, bind, and interleave policies.
The default policy will cause pages to be added to tmpfs files on the node
which is doing the writing. Some jobs expect a single process to create
and manage the tmpfs files. This results in a node which has a
significantly reduced number of free pages.
With this patch, the administrator can specify the policy and nodes for
that policy where they would prefer allocations.
This patch was originally written by Brent Casavant and Hugh Dickins. I
added support for the bind and preferred policies and the mpol_nodelist
mount option.
Signed-off-by: Brent Casavant <bcasavan@sgi.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Robin Holt <holt@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Some people apparently run CONFIG_NUMA without CONFIG_SWAP. The migration
code currently depends on swap. This patch provides a set of inline
fallback functions so that the kernel properly compiles. However, calls to
migration functions will fail.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Calin A. Culianu [Sat, 14 Jan 2006 21:20:46 +0000 (13:20 -0800)]
[PATCH] Watchdog: Winsystems EPX-C3 SBC
This is a 2.6 patch that adds support for the watchdog timer built into the
EPX-C3 single board computer manufactured by Winsystems, Inc.
Driver details:
This is for x86 only. This watchdog is pretty basic and simple. It is
only configurable via jumpers on the SBC, and it only has either a 1.5s or
200s interval. The watchdog can either be auto-configured to start as soon
as the machine powers up (bad idea for the 1.5s interval!) or it can be
enabled and disabled by writing to io port 0x1ee. Petting the watchdog
involves writing any value to io port 0x1ef.
The only unfortunate thing about this watchdog (and it is not at all
uncommmon in watchdogs that linux supports) is that it is not a PCI or
ISA-PNP device and as such it isn't at all probeable. Either the watchdog
exists as 2 bytes at 0x1ee, or it doesn't. Thus, using this driver on a
machine that doesn't have that watchdog can potentially hang/crash the
system, etc. So only use this driver if you in fact are on a Winsystems
EPX-C3 SBC.
Anyway this driver fits into the already-existing watchdog framework quite
nicely and I already tested it on my EPX-C3 and it works like a charm.
Signed-off-by: Calin A. Culianu <calin@ajvar.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Miklos Szeredi [Sat, 14 Jan 2006 21:20:44 +0000 (13:20 -0800)]
[PATCH] uml: fix symbol for mktime
LD .tmp_vmlinux1
/usr/lib/gcc-lib/i486-linux/3.3.4/../../../libc.a(mktime.o): In function `timelocal':
: multiple definition of `mktime'
kernel/built-in.o:kernel/time.c:604: first defined here
/usr/bin/ld: Warning: size of symbol `mktime' changed from 134 in kernel/built-in.o to 44 in /usr/lib/gcc-lib/i486-linux/3.3.4/../../../libc.a(mktime.o)
collect2: ld returned 1 exit status
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu> Cc: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Arjan van de Ven [Sat, 14 Jan 2006 21:20:43 +0000 (13:20 -0800)]
[PATCH] Unlinline a bunch of other functions
Remove the "inline" keyword from a bunch of big functions in the kernel with
the goal of shrinking it by 30kb to 40kb
Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jeff Garzik <jgarzik@pobox.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ingo Molnar [Sat, 14 Jan 2006 21:20:41 +0000 (13:20 -0800)]
[PATCH] sched: add new SCHED_BATCH policy
Add a new SCHED_BATCH (3) scheduling policy: such tasks are presumed
CPU-intensive, and will acquire a constant +5 priority level penalty. Such
policy is nice for workloads that are non-interactive, but which do not
want to give up their nice levels. The policy is also useful for workloads
that want a deterministic scheduling policy without interactivity causing
extra preemptions (between that workload's tasks).
Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Michael Kerrisk <mtk-manpages@gmx.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Patrick Gefre [Sat, 14 Jan 2006 21:20:40 +0000 (13:20 -0800)]
[PATCH] Altix: ioc3 serial support
Add driver support for a 2 port PCI IOC3-based serial card on Altix boxes:
This is a re-submission. On the original submission I was asked to
organize the code so that the MIPS ioc3 ethernet and serial parts could be
used with this driver. Stanislaw Skowronek was kind enough to provide the
shim layer for this - thanks Stanislaw. This patch includes the shim layer
and the Altix PCI ioc3 serial driver. The MIPS merged ioc3 ethernet and
serial support is forthcoming.
Signed-off-by: Patrick Gefre <pfg@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Alexander Viro [Sat, 14 Jan 2006 20:29:55 +0000 (15:29 -0500)]
[PATCH] Fix double decrement of mqueue_mnt->mnt_count in sys_mq_open
Fixed the refcounting on failure exits in sys_mq_open() and
cleaned the logics up. Rules are actually pretty simple - dentry_open()
expects vfsmount and dentry to be pinned down and it either transfers
them into created struct file or drops them. Old code had been very
confused in that area - if dentry_open() had failed either in do_open()
or do_create(), we ended up dentry and mqueue_mnt dropped twice, once
by dentry_open() cleanup and then by sys_mq_open().
Fix consists of making the rules for do_create() and do_open()
same as for dentry_open() and updating the sys_mq_open() accordingly;
that actually leads to more straightforward code and less work on
normal path.
Signed-off-by: Al Viro <aviro@redhat.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Evgeniy [Sat, 14 Jan 2006 08:42:06 +0000 (11:42 +0300)]
[PATCH] ufs cleanup
Here is update of ufs cleanup patch, brought on by the recently fixed
ubh_get_usb_second() bug that made some ugly code rather painfully
obvious. It also includes
- fix compilation warnings which appears if debug mode turn on
- remove unnecessary duplication of code to support UFS2
[SCSI] qla2xxx: Disable port-type RSCN handling via driver state-machine.
Given the semantic changes in both the device-model and
fc-transport APIs, the driver's handling of port-type RSCNs
via a series of ADISCs and PLOGIs can cause series of
badness ranging from unexpectedly device loss to devices not
being discovered.
In the interim, disable (via a module-parameter) this
feature and allow RSCN management to continue to occur
within the driver's DPC thread.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] qla2xxx: Correct issue where portstate does not transition during loop-resync.
If the Get Port Database call fails during local-loop
update, then schedule the DPC routine to perform a rescan as
the firmware would have updated the Get ID List port-entries
of their new state.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] qla2xxx: Correct excessive delay during LOAD-RISC-RAM mailbox command.
Problem report (against 2.4.x driver) from Jeff Layton
<jlayton@redhat.com>:
An OEM noticed that the U6 qla2200 driver would hang for
around 2 minutes at boot time and then proceed normally. I
found that the delay was occurring when loading the new
firmware into the card, and was due to a
schedule_timeout(10) added to the bottom of the polling
loop.
Some testing showed that the load ram operation on the card
was very quick (on the order of a couple of jiffies), but
the sleep in the polling loop was making each operation take
around 25-30.
The attached patch corrects this by making it skip sleeping
during the load ram operation, since I believe we only do
that when the module is plugged in. It also skips sleeping
if the mbox_int flag got set during the current loop.
This corrected the hang on my test setup, and OEM also
confirmed that it corrected the problem for them.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] qla2xxx: Use msleep() as delay during ISP polling.
Mailbox commands are polled for completion during ISP
initialization. During potentially 'long' mailbox commands
(i.e. fabric login), we really don't want a busy-wait delay
to potentially trigger a (benign) soft-lockup BUG().
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] qla2xxx: Reference proper node/port names in fc_host class.
The initial-control-block references are not always correct
as the use-node-name qualifier during NVRAM configuration
will cause the firmware to use the portname as a base for
the nodename.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Mike Christie [Sat, 14 Jan 2006 00:05:56 +0000 (18:05 -0600)]
[SCSI] iscsi: use pageslab
From: FUJITA Tomonori <tomof@acm.org> and zhenyu.z.wang@intel.com:
We cannot handle filesystems like XFS becuase of the pages they
are sending us. We had thought page_count could be used to
work around this, but the correct test is for PageSlab.
The proper solution is to figure out what type of pages
filesystems can use so we do not have to add tests like
this or handle it in the block layer for all network block drivers
but the issue still has not been resolved on fs-devel
so we are sending this patch as a temporary fix.
This is last patch just in case it is Nakd with the explanation
that we need to push the correct fix through fs-devel, mm
or the block layer. The rest of the patchset can live without
the patch, but the driver will not work with filesystems like
XFS.
Signed-off-by: Alex Aizman <itn780@yahoo.com> Signed-off-by: Dmitry Yusupov <dmitry_yus@yahoo.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Mike Christie [Sat, 14 Jan 2006 00:05:53 +0000 (18:05 -0600)]
[SCSI] iscsi: fix 4k stack iscsi setups
When we run the xmit code from queuecomand the stack trace
gets too deep. The patch runs the xmit code from the scsi_host
work queue. This fixes 4k stack and xfs support and should
fix the st and sg stack usage bugs.
Signed-off-by: Alex Aizman <itn780@yahoo.com> Signed-off-by: Dmitry Yusupov <dmitry_yus@yahoo.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Mike Christie [Sat, 14 Jan 2006 00:05:50 +0000 (18:05 -0600)]
[SCSI] iscsi: seperate iscsi interface from setup functions
This is the second version of the patch to address Christoph's comments.
Instead of doing the lib, I just kept everything in scsi_trnapsort_iscsi.c
like the FC and SPI class. This was becuase the driver model and sysfs
class is tied to the session and connection setup so separating did not
buy very much at this time.
The reason for this patch was becuase HW iscsi LLDs like qla4xxx cannot
use the iscsi class becuase the scsi_host was tied to the interface and
class code. This patch just seperates the session from scsi host so
that LLDs that allocate the host per some resource like pci device
can still use the class.
This is also fixes a couple refcount bugs that can be triggered
when users have a sysfs file open, close the session, then
read or write to the file.
Signed-off-by: Alex Aizman <itn780@yahoo.com> Signed-off-by: Dmitry Yusupov <dmitry_yus@yahoo.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
FUJITA Tomonori [Sat, 14 Jan 2006 00:05:44 +0000 (18:05 -0600)]
[SCSI] iscsi: data digest page cache usage fix
Users can write to a page while we are sending it and making
digest calculations. This ends up causing us to retry the command
when a digest error is later reported. By using sock_no_sendpage
when data digests are calculated we can avoid a lot of (not all but it
helps) the retries becuase sock_no_sendpage is not zero copy.
Signed-off-by: Alex Aizman <itn780@yahoo.com> Signed-off-by: Dmitry Yusupov <dmitry_yus@yahoo.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Moore, Eric [Fri, 13 Jan 2006 23:33:59 +0000 (16:33 -0700)]
[SCSI] scsi_transport_sas: mapping the rphy channel equal to the port identifier
We will be mapping the RAID volumes in mptsas to a reserved
channel that
is one larger than the anticapated number of ports on the direct
attached host
adapter.
Signed-off-by: Eric Moore <Eric.Moore@lsil.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
When James Smart fixed the issue of the userspace scan atributes
crashing the system with the FC transport class he added a patch to
let the transport class check if the parent is valid for a given
transport class.
When adding support for the integrated raid of fusion sas devices
we ran into a problem with that, as it didn't allow adding virtual
raid volumes without the transport class knowing about it.
So this patch adds a user_scan attribute instead, that takes over from
scsi_scan_host_selected if the transport class sets it and thus lets
the transport class control the user-initiated scanning. As this
plugs the hole about user-initiated scanning the target_parent hook
goes away and we rely on callers of the scanning routines to do
something sensible.
For SAS this meant I had to switch from a spinlock to a mutex to
synchronize the topology linked lists, in FC they were completely
unsynchronized which seems wrong.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Michael Reed [Fri, 13 Jan 2006 20:31:54 +0000 (14:31 -0600)]
[SCSI] mptfusion - fc transport attributes
Signed-off-by: Michael Reed <mdr@sgi.com> Signed-off-by: Eric Moore <Eric.Moore@lsil.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Adds hotplug support for SAS end devices. Unfortunately the fusion
firmware doesn't generate similar events for expanders addition/removal
so we can't support them yet. Eric has an idea about a clever scheme to
find out about expander changes so that'll be added later on.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Salyzyn, Mark [Thu, 12 Jan 2006 13:31:57 +0000 (08:31 -0500)]
[SCSI] I2O: move pci_request_regions() just behind pci_enable_device()
The problem in dpt_i2o could be the pci config space accesses it
triggers as it loads, dangerous to do if there is any I/O activity going
on in the other driver (probable if a boot driver I guess).
I approve this patch to dpt_i2o.c, and am applying it to the Adaptec
branch of the driver.
Thanks for the investigation Ryoji.
---
In linux 2.6.15, data transfer does hang when both dpt_i2o
and i2o_block drivers are loaded.
It seems that location of pci_request_regions() are wrong.
I moved it just behind pci_enable_device() like other drivers,
and it becomes fine.
Signed-off-by: Ryoji Kamei <kamei@miraclelinux.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Andreas Herrmann [Fri, 13 Jan 2006 01:26:11 +0000 (02:26 +0100)]
[SCSI] zfcp: transport class adaptations II
Replaced zfcp adapter attributes with fc_host attributes:
fc_topology by port_type, physical_wwpn by permanent_port_name.
Make use of fc_host attribute supported_speeds.
Removed zfcp adapter attribute physical_s_id.
Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Added host stats, removed superfluous get_starget_ functions,
removed some attributes from zfcp specific sysfs tree (e.g.
scsi_host_no, scsi_lun, wwnn and d_id).
Host stats are given for the physical adapter port not for the
virtual adapter. Reset stats is implemented in the device driver.
Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
projects / linux-2.6 / log