In a recent patch, all the ST_ typedefs were removed from st.h.  One of
the side effects of this makes osst no longer compile.  This patch
replaces all of the typedef uses from st.h with their struct
equivalents.
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

Just had a case with a gdth adapter not working in a 16GB box. Seems to
me that the SCSI bounce decision is completely screwy - it returns the
device dma mask, if it is set only and the box doesn't have an iommu. So
if we can't get the mask, we allow any io. Not so good. I think the
logic should be more as follows:

	if (isa)
		return bounce_above_isa;
	if (has_iommu)
		return bounce_nothing;

	if (we_know_device_dma_mask)
		return mask;
	else
		return default_to_32bit;

Patch corrects scsi_calculate_bounce_limit() to actually work partly
sanely.
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

This patch augments Christoph Hellwig's earlier patch by converting the local
typedefs to struct definitions. One debugging printk format is also fixed. The
patch is against 2.6.10-rc1, compiles and has been tested.
Signed-off-by: Kai Makisara <kai.makisara@kolumbus.fi>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

Any comments would be appreciated.

Description: Use msleep_interruptible() instead of
schedule_timeout() to guarantee the task delays as expected.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Maximilian Attems <janitor@sternwelten.at>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

Any comments would be appreciated.

Description: Use msleep_interruptible() instead of
schedule_timeout() to guarantee the task delays as expected.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Maximilian Attems <janitor@sternwelten.at>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

This bug was present as of 2.6.9-bk7, apparently introduced along with the
addition of SCSI targets into sysfs.  The code takes a reference to the
host for each initialized sdev but only drops a reference when each
target is released.  As you might expect, this causes a refcount leak
whenever a target has more than 0 LUNs.

The patch changes things so that the reference is acquired when the target
is initialized, not when the sdev is initialized.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

From: 	Matthew Wilcox <matthew@wil.cx>

sym2 2.1.18m:
 - Improve documentation
 - Remove host-level command queueing (Christoph Hellwig)
 - Add support for IU and QAS
 - sparse annotation (Randolph Chung)
 - Remove SYM_LINUX_BOOT_COMMAND_LINE_SUPPORT ifdefs
 - Switch to the new module_param interface
   - Rename tags to cmd_per_lun for consistency
   - Fix debug param
   - Split extended tags functionality into the new tag_ctrl parameter
   - rewrite how safe param works
   - reverse_probe param has had no effect for a while, remove it.
   - Add descriptions for all the parameters
 - Add MODULE_AUTHOR and MODULE_DESCRIPTION
 - Restructure sym_config_pqs() a little
 - Move hostid setup to a more sensible place
 - Do away with SYM_GLUE_C
 - Use a completion instead of a semaphore (Thomas Gleixner)
 - Use IDENTIFY() instead of M_IDENTIFY
 - Define messages in terms of those defined in <scsi/scsi.h>
 - Define PCI IDs in terms of those in <linux/pci_ids.h>
 - Define sym2's status codes in terms of the SAM ones defined in <scsi/scsi.h>

Added brown bag fix
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

The MODULE_DEVICE_TABLE() declaration, which is used by boot and hot
plug facilities to load the module when the hardware is present was
missed in the PCI API conversion.
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

From: Jan Dittmer <jdittmer@ppp0.net>

This patch converts the aic7xxx driver to use module_param instead of
MODULE_PARM and therefore gets rid of two warning in the non-modular build
case.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

From: "Sergey S. Kostyliov" <rathamahata@ehouse.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

'driver_setup' cannot be __initdata if slave_configure() can be called
after driver-init:
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000000fbf R_X86_64_PC32     .init.data+0x000000000000007b
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000000fc9 R_X86_64_PC32     .init.data+0x000000000000007e
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000000fea R_X86_64_PC32     .init.data+0x000000000000007b
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000000ff4 R_X86_64_PC32     .init.data+0x0000000000000080
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000001039 R_X86_64_PC32     .init.data+0x000000000000007b
Error: ./drivers/scsi/qla1280.o .text refers to 0000000000001043 R_X86_64_PC32     .init.data+0x0000000000000082
Signed-off-by: Randy Dunlap <rddunlap@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

Make all of the qlogic firmware pointers static
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

This marks sets change_bits to all ones when bringing up a hub, since
not all hubs seem to send change events for devices that were plugged in
when the hub was reset. 

David Miller confirms this fixes his boot-time lost keyboard/mouse
problem

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into kroah.com:/home/greg/linux/BK/usb-2.6

This updates debug messages, declaring that labels can be constant strings
and changing some old-school dbg() calls to driver model dev_dbg().
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

The recent patch to scrub out ep0 state earlier (to get rid of some of
the enumeration problems that started with about 2.6.6) requires EHCI
to handle endpoint_disable() calls in a slightly different context.

This makes those calls work when an endpoint's QH may still be on the
async schedule, rather than already unlinked.  (The QH stays on the async
schedule for a few milliseconds after it's empty, since it's routine to
issue another request almost immediately.)
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This has minor PCI tweaks for the EHCI driver:

  - If the (nonfunctional) AMD 8111 controller shows up, ignore it.
    (Patch from Matt Dharm, with minor coding style tweaks).

  - Delete some code that interprets an EHCI capability code; it
    should be a PCI capability instead.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

New and Improved! "Woozy Numbat!"

Pick up the phone in the next two minutes and phone in your
order, and we'll throw in a second copy of Woozy Numbat,
ABSOLUTELY FREE! You too could start a Numbat farm!

lib/kobject_uevent.c:39: warning: `action_to_string' defined but not used
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

From: Keshavamurthy Anil S <anil.s.keshavamurthy@intel.com>

Remove cpu_run_sbin_hotplug() - use kobject_hotplug() instead.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Make kobject_hotplug() work even if the kobject's kset doesn't implement any
hotplug_ops.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

kobject_set_name takes a printf style argument list. There are many
callers that pass only one string, if this string contained a '%' character
than bad things would happen.  The fix is simple.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Here is a better fix (thanks Greg) that allows long names for character
device objects.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Signed-off-by: cw@f00f.org
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

into kroah.com:/home/greg/linux/BK/driver-2.6

arch/sh/tools/machgen.sh is no longer used as it has been replaced by
arch/sh/tools/gen-mach-types
Signed-off-by: Tom Rini <trini@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Uses msleep() instead of schedule_timeout() to guarantee the task delays
as expected.  Two of the changes are reworks of previous msleep() calls
which unnecessarily added a jiffy to the parameter.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

there was an uneeded packed attribute for a data structure.
Signed-off-by: Oliver Neukum <oliver@neukum.name>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

kaweth used its own synchronisation superseded by usb_unlink_urb().
Signed-off-by: Oliver Neukum <oliver@neukum.name>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This tweaks the USB input driver support, notably fixing a botched
dependency that makes all the USB drivers appear strangely in Kconfig.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

The following compile error seems to come from Linus' tree:


<--  snip  -->

...
  CC      drivers/usb/serial/bus.o
  CC      drivers/usb/serial/console.o
drivers/usb/serial/console.c: In function `usb_console_write':
drivers/usb/serial/console.c:221: warning: passing arg 3 of pointer to function makes integer from pointer without a cast
drivers/usb/serial/console.c:221: error: too many arguments to function
drivers/usb/serial/console.c:223: warning: passing arg 3 of `usb_serial_generic_write' makes integer from pointer without a cast
drivers/usb/serial/console.c:223: error: too many arguments to function `usb_serial_generic_write'
make[3]: *** [drivers/usb/serial/console.o] Error 1

<--  snip  -->


This was caused by the changed "write" in usb_serial_device_type.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This patch implements the Windows scheme for USB device initialization.
It also incorporates the change recently posted by David to scrub the
endpoint state following a SET-ADDRESS.  Other noteworthy changes:

	There are two new module parameters to control whether the
	old scheme or the new one is used first and whether the other
	scheme is tried if the first one fails.  Default settings are
	to use the new scheme only.

	hub_set_address() returns 0 immediately if the device is already
	in the USB_STATE_ADDRESS state.

	On the first attempt to read the device descriptor the code
	uses a short 1-second timeout.  This ought to help prevent
	full-speed devices with an 8- or 16-byte maxpacket from
	slowing the procedure down by NAKing the unexpectedly early
	status stage of the transfer.

	For debugging, the ep0 maxpacket value is printed.  It might
	be a good idea to validate it rather than just believing the
	device -- although I haven't heard of any device providing
	an incorrect value other than 0.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Description: Use msleep_interruptible() instead of schedule_timeout()
so that the task delays as expected.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Maximilian Attems <janitor@sternwelten.at>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Description: Use msleep_interruptible() instead of schedule_timeout() to
guarantee the task delays as expected.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Maximilian Attems <janitor@sternwelten.at>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

.. broken by the rename.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Chris Wright <chrisw@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This didn't show up in my "allmodconfig" tests, because the
console requires things to be built-in, not modules.

Posix timers preallocate siqueue structures during timer creation
and keep them for reuse.  This allocation happens in user context
with no locks held, however it's designated as an atomic allocation.
Loosen this restriction, and while we're at it let's do a bit of code
consolidation so signal sending uses same __sigqueue_alloc() helper.
Signed-off-by: Chris Wright <chrisw@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>