The git commit 794f5bfa
accidentally suffers from a previous typo in that file
(',' instead of ';' in end of line). Patch included.

Signed-off-by: Mika Kukkonen (mikukkon@iki.fi)
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The fixes for sparse warnings mixed in with the fixups for
the raw_srb handler resulted in a bug that showed up in the 32 bit
environments when trying to issue calls directly to the physical devices
that are part of the arrays (ioctl scsi passthrough).

Received from Mark Salyzyn at adaptec.
Applied comment from Christoph to remove cpu_to_le32(0)
Applied Mark S fix of missing memcpy.
It applies to the scsi-misc-2.6 git tree.
Signed-off-by: NMark Haverkamp <markh@osdl.org>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

There are many drivers that have been setting the generic driver
model level shutdown callback, and pci thus must not override it.

Without this patch we can have really bad data loss on various
raid controllers.
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Drivers need not implement a hook that returns FAILED, and does nothing
else, since the SCSI midlayer code will do that for us.
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The SCSI ->done() hook should not be called from inside a spinlock.

Drivers that do this are mostly cut-n-paste from 2.2.x-era.
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

__cfq_get_queue().  __cfq_get_queue() finds an existing queue (struct
cfq_queue) of the current process for the device and returns it.  If it's not
found, __cfq_get_queue() creates and returns a new one if __cfq_get_queue() is
called with __GFP_WAIT flag, or __cfq_get_queue() returns NULL (this means that
get_request() fails) if no __GFP_WAIT flag.

On the other hand, in __make_request(), get_request() is called without
__GFP_WAIT flag at the first time.  Thus, the get_request() fails when there is
no existing queue, typically when it's called for the first I/O request of the
process to the device.

Though it will be followed by get_request_wait() for general case,
__make_request() will just end the I/O with an error (EWOULDBLOCK) when the
request was for read-ahead.
Signed-off-by: NJens Axboe <axboe@suse.de>
Signed-off-by: NKiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: NJun'ichi Nomura <j-nomura@ce.jp.nec.com>

Cc: Greg KH <greg@kroah.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

It looks like logic for enabling hardware tapping in ALPS driver was
inverted and we enable it only if it was already enabled by BIOS or
firmware.

I have a confirmation from one user that the patch below fixes the problem
for him and it might be beneficial if we could get it into 2.6.12.
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This fixed a problem that showed up in the Fedora development tree a few
weeks before the Fedora Core 4 release, initially as slab corruption, later
as hard crashes on boot up, when slab debugging was disabled for the
release.  More details on the history at
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=158424

The problem is caused by sbp2's use of scsi_host->hostdata[0] to hold a
scsi_id, without explicitly requesting space for it.  Since hostdata is
declared as a zero-sized array, we don't get any such space by default, so
it must be explicitly requested.  The patch below implements just that.
Signed-off-by: NAlexandre Oliva <aoliva@redhat.com>
Cc: Jody McIntyre <scjody@modernduck.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

__elv_add_request().  rq.count[READ] + rq.count[WRITE] can increase
more than one if another thread has allocated a request after the
current request is allocated or in_flight could have changed resulting
in larger-than-one change of nrq, thus breaking the threshold
mechanism.
Signed-off-by: NJens Axboe <axboe@suse.de>
Signed-off-by: NTejun Heo <htejun@gmail.com>

nothing.
Signed-off-by: NJens Axboe <axboe@suse.de>
Signed-off-by: NTejun Heo <htejun@gmail.com>

Module needs a license to prevent kernel tainting.
Signed-off-by: NRandy Dunlap <rdunlap@xenotime.net>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fixed freeing of event memory in i2o_block_event()
Signed-off-by: NMarkus Lidel <Markus.Lidel@shadowconnect.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

It prints out x,x instead of x,y.
Signed-off-by: NJon Smirl <jonsmirl@gmail.com>
Cc: "Antonino A. Daplas" <adaplas@hotpop.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

On 64-bit machines, PCI_BASE_ADDRESS_MEM_MASK and other mask constants
passed to pci_size() are 64-bit (for example ~0x0fUL).  However, pci_size
does comparisons between the u32 arguments and the mask, which will fail
even though any result from pci_size is still just 32-bit.

Changing the mask argument to u32 seems the obvious thing to do, since all
arithmetic in the function is 32-bit and having a larger mask makes no
sense.

This triggered on a PPC64 system here where an adapter (VGA, as it
happened) had a memory region base of 0xfe000000 and a sz of the same,
matching the if (max == maxbase ...) test at the bottom of pci_size but
failing the mask comparison.  Quite a corner case which I guess explains
why we haven't seen it until now.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Acked-by: NGreg KH <greg@kroah.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Following the go around over the SONY DVD that needs artificial limits,
this should be the correct code for all cases (minus the debugging
prints).
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Maxim Shchetynin <maxim@de.ibm.com>
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Heiko Carstens <heiko.carstens@de.ibm.com>

Fixes module parameter parsing for "device" parameter.  The original
module parameter was changed while parsing it.  This corrupted the
output in sysfs (/sys/module/zfcp/parameters/device).
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Fixes a race between zfcp_fsf_req_dismiss_all and
zfcp_qdio_reqid_check. During adapter shutdown it occurred that a
request was cleaned up twice. First during its normal
completion. Second when dismiss_all was called.  The fix is to
serialize access to fsf request list between zfcp_fsf_req_dismiss_all
and zfcp_qdio_reqid_check and delete a fsf request from the list if
its completion is triggered.  (Additionally a rwlock was replaced by a
spinlock and fsf_req_cleanup was eliminated.)
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Maxim Shchetynin <maxim@de.ibm.com>

Fixes a bug in zfcp_send_els_handler.  If D_ID assignments for ports
are changing between initiation of one ELS request and its completion
the wrong port might be accessed in the completion for that ELS
request. Thus a pointer to the port has to be passed for ELS requests
to identify the port structure if required.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Maxim Shchetynin <maxim@de.ibm.com>

Correct a bug in zfcp_fsf_send_fcp_command_handler.  An fsf request
was not marked as failed if an unknown status qualifier was returned.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Maxim Shchetynin <maxim@de.ibm.com>

Reopen a remote port only if the link-test fails. This avoids that a
port is unnecessarily reopened.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Maxim Shchetynin <maxim@de.ibm.com>

Extend the time for adapter initialization: In case of protocol
status HOST_CONNECTION_INITIALIZING for the exchange config data
command do a first retry in 1 second, then double the sleep time for
each following retry until recovery exceeds 2 minutes. The old
behaviour of allowing 6 retries with .5 seconds delay between retries
was insufficient and qdio queues were shut down too erarly.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Fixes the handling of failed requests for GID_PN nameserver command:
Set ZFCP_STATUS_PORT_INVALID_WWPN only if indicated by response
payload for GID_PN nameserver command and not if fsf request fails.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

From: Ralf Baechle <ralf@linux-mips.org>

There are archives of the old list at http://oss.sgi.com/archives/netdevSigned-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This makes the EHCI driver spin a bit longer before concluding that the
port reset failed.  "Obviously safe."

It allows some devices to enumerate that previously didn't.  We've seen
a bunch of these problem reports recently, this will make some go away.

As reported by Michael Zapf <Michael.Zapf@uni-kassel.de>, some EHCI
controllers seem to take forever to finish port resets and produce
"port N reset error -110" type errors.  Spinning a bit longer helps.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The pwc chainsaw session left some setups not working.  There is a
sanity check on compression buffers that simply isn't right any more as
we never allocate one.

This doesn't address the email and other changes.  I'll do those
tomorrow if I get time, but it is the minimal fix for the code and basic
feature set.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The current radeonfb memset's the framebuffer to 0 when loaded.  This
removes occasional artifacts but has the nasty side effect that if you
load radeonfb without framebuffer console, you destroy the VGA text
buffer, font, etc...  radeon must not touch the framebuffer content when
it doesn't "own" it.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

M68k: Mark Sun-3 NCR5380 SCSI broken until NCR5380_abort() and
NCR5380_bus_reset() are replaced with real new-style EH routines (the old EH
SCSI constants were removed in 2.6.12-rc3).
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

 - the eisa layer only probes when it's actually safe, no need for
   a driver option
 - store the id table directly in linux format instead of convering
   at runtime
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

especially the now dead scsi_cmnd overlay
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

there's absolutely no reason not to trust the driver private data
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Removes the rarely used "flags_dump" mechanism of zfcp.

Equivalent debug information will be provided with a reworking of
zfcp's s390dbf-facilities which is in preparation.
Signed-off-by: NAndreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Patch removes our homegrown DMA masks and uses the ones defined in the kernel.
This patch replaces the broken one I sent in earlier. It has been tested and works. Please discard the first submission.
Signed-off-by: NMike Miller <mike.miller@hp.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

pci layer handles this just fine for us
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>