If an mlx4 device with default FW (which gives a UAR BAR size of 8 MB)
is used in a system with 64 KB pages, then there are only 8192/64==128
UAR pages available.  However, the first 128 UAR pages are reserved
for use with event queue doorbells, so no UAR pages are available to
do anything else with, which means that the driver cannot work.

The current driver fails with a fairly cryptic "Failed to allocate
driver access region, aborting" message in this situation.  Fix the
driver to detect the problem earlier and print out a clearer
description of the problem and a suggestion of how to fix it (use a
new firmware image).
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Add support for the following operations to mlx4 when device firmware
supports them:

 - Send with invalidate and local invalidate send queue work requests;
 - Allocate/free fast register MRs;
 - Allocate/free fast register MR page lists;
 - Fast register MR send queue work requests;
 - Local DMA L_Key.
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

MTT entries are allocated with a buddy allocator, which just keeps
bitmaps for each level of the buddy table.  However, all free space
starts out at the highest order, and small allocations start scanning
from the lowest order.  When the lowest order tables have no free
space, this can lead to scanning potentially millions of bits before
finding a free entry at a higher order.

We can avoid this by just keeping a count of how many free entries
each order has, and skipping the bitmap scan when an order is
completely empty.  This provides a nice performance boost for a
negligible increase in memory usage.
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Add ICM_ERROR firmware status code.  In mapping to errnos, -ENFILE
seems closest.

This is in preparation for providing more detailed log info using
mlx4_err() in low-level driver when a non-zero status is returned.
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Make the struct name consistent with other WQE segment struct types
defined in <linux/mlx4/qp.h>.
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Add a module parameter "enable_qos" to mlx4_core.  If this param is
set, enable support for QoS in the INIT_HCA command.  By default, the
parameter is set to 0 (disabled).
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

When a kernel was rebuilt, the previous Module.markers was not cleared.
It caused markers with different format strings to appear as duplicates
when a markers was changed.  This problem is present since
scripts/mod/modpost.c started to generate Module.markers, commit
b2e3e658

It therefore applies to 2.6.25, 2.6.26 and linux-next.

I merely merged the patches from Roland, Wenji and Takashi here.

Credits to
Roland McGrath <roland@redhat.com>
Wenji Huang <wenji.huang@oracle.com>
and
Takashi Nishiie <t-nishiie@np.css.fujitsu.com>

for providing the individual fixes.

- Changelog :
  - Integrated Takashi's Makefile modification to clear Module.markers upon
    make clean.
Signed-off-by: NMathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Roland McGrath <roland@redhat.com>
Cc: Wenji Huang <wenji.huang@oracle.com>
Cc: Takashi Nishiie <t-nishiie@np.css.fujitsu.com>
Cc: <stable@kernel.org>		[2.6.25.x, 2.6.26.x]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add missing cond_syscall() entry for compat_sys_epoll_pwait.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: <stable@kernel.org>		[2.6.25.x, 2.6.26.x]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NRajiv Andrade <srajiv@linux.vnet.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

struct pagemap_walk was placed on stack, some hooks are initialized, the
rest (->pgd_entry, ->pud_entry, ->pte_entry) are valid but junk.
Reported-by: NEric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Cc: "Vegard Nossum" <vegard.nossum@gmail.com>
Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix wrong domain attr updates, or we will always update the first sched
domain attr.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Cc: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org>	[2.6.26.x]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Two small updates to the pcf857x driver: (a) the max732[89] chips are
also second sources for the pcf8574/a, and (b) add a mutex to prevent
trashing the cached state.  Adding the lock is effectively a bugfix,
although it seems unlikely that anyone would have run into the issue it
protects against.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Maxim's MAX7301 is an SPI GPIO expander with 28 GPIOs.  Note: MAX7301's
interrupt feature is not supported yet.

[akpm@linux-foundation.org: coding-style fixes]
[g.liakhovetski@pengutronix.de: Fix inaccuracies in comments, check spi_setup()
return code, mask off high byte in max7301_read()]
Signed-off-by: NJuergen Beisert <j.beisert@pengutronix.de>
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@pengutronix.de>
Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The Linux kernel puts the filename argument of execve() into the new
address space.  Many developers are surprised to learn this.  Those who
know and could use it, object "But it's not documented."

Those who want to use it dislike the expression
  (char *)(1+ strlen(env[-1+ n_env]) + env[-1+ n_env])
because it requires locating the last original environment variable,
and assumes that the filename follows the characters.

This patch documents the insertion of the filename, and makes it easier
to find by adding a new tag AT_EXECFN in the ElfXX_auxv_t; see <elf.h>.

In many cases readlink("/proc/self/exe",) gives the same answer.  But if
all the original pages get unmapped, then the kernel erases the symlink
for /proc/self/exe.  This can happen when a program decompressor does a
good job of cleaning up after uncompressing directly to memory, so that
the address space of the target program looks the same as if compression
had never happened.  One example is http://upx.sourceforge.net .

One notable use of the underlying concept (what path containED the
executable) is glibc expanding $ORIGIN in DT_RUNPATH.  In practice for
the near term, it may be a good idea for user-mode code to use both
/proc/self/exe and AT_EXECFN as fall-back methods for each other.
/proc/self/exe can fail due to unmapping, AT_EXECFN can fail because it
won't be present on non-new systems.  The auxvec or {AT_EXECFN}.d_val
also can get overwritten, although in nearly all cases this would be the
result of a bug.

The runtime cost is one NEW_AUX_ENT using two words of stack space.  The
underlying value is maintained already as bprm->exec; setup_arg_pages()
in fs/exec.c slides it for stack_shift, etc.
Signed-off-by: NJohn Reiser <jreiser@BitWagon.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Jakub Jelinek <jakub@redhat.com>
Cc: Ulrich Drepper <drepper@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (100 commits)
  usb-storage: revert DMA-alignment change for Wireless USB
  USB: use reset_resume when normal resume fails
  usb_gadget: composite cdc gadget fault handling
  usb gadget: minor USBCV fix for composite framework
  USB: Fix bug with byte order in isp116x-hcd.c fio write/read
  USB: fix double kfree in ipaq in error case
  USB: fix build error in cdc-acm for CONFIG_PM=n
  USB: remove board-specific UP2OCR configuration from pxa27x-udc
  USB: EHCI: Reconciling USB register differences on MPC85xx vs MPC83xx
  USB: Fix pointer/int cast in USB devio code
  usb gadget: g_cdc dependso on NET
  USB: Au1xxx-usb: suspend/resume support.
  USB: Au1xxx-usb: clean up ohci/ehci bus glue sources.
  usbfs: don't store bad pointers in registration
  usbfs: fix race between open and unregister
  usbfs: simplify the lookup-by-minor routines
  usbfs: send disconnect signals when device is unregistered
  USB: Force unbinding of drivers lacking reset_resume or other methods
  USB: ohci-pnx4008: I2C cleanups and fixes
  USB: debug port converter does not accept more than 8 byte packets
  ...

This patch (as1110) reverts an earlier patch meant to help with
Wireless USB host controllers.  These controllers can have bulk
maxpacket values larger than 512, which puts unusual constraints on
the sizes of scatter-gather list elements.  However it turns out that
the block layer does not provide the support we need to enforce these
constraints; merely changing the DMA alignment mask doesn't help.
Hence there's no reason to keep the original patch.  The Wireless USB
problem will have to be solved a different way.

In addition, there is a reason to get rid of the earlier patch.  By
dereferencing a pointer stored in the ep_in array of struct
usb_device, the current code risks an invalid memory access when it
runs concurrently with device removal.  The members of that array are
cleared before the driver's disconnect method is called, so it should
not try to use them.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1109b) makes USB-Persist more resilient to errors.  With
the current code, if a normal resume fails, it's an unrecoverable
error.  With the patch, if a normal resume fails (and if the device is
enabled for USB-Persist) then a reset-resume is tried.

This fixes the problem reported in Bugzilla #10977.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

These two fixes ensure the new "CDC Composite Device" gadget
fails cleanly when it's loaded on hardware that can't support
this particular gadget driver.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Fill in a reserved/unused device qualifier field to ensure that
the USBCV tests will always pass.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

URB payload data are transfered in wrong byte order on a big endinan
architecture (AVR32).
Signed-off-by: NJulien May <mailinglist@miromico.ch>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

in the error case the ipaq driver leaves a dangling pointer to already
freed memory that will be freed again.
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Here's the fix. cdc-wdm has the same problem. The fix is the same.
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Remove the board-specific UP2OCR configuration from the
pxa27x-udc driver.
Signed-off-by: NDaniel Ribeiro <drwyrm@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

A couple of USB register initializations had to be changed on MPC85xx
platforms.  This is due to the internal SoC buses being different on
MPC83xx SoCs vs MPC85xx SoCs.

We currently handle this via an ifdef since 83xx and 85xx are mutually
exclusive kernel builds.
Signed-off-by: NSrikanth Srinivasan <srikanth.srinivasan@freescale.com>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Fix pointer/int cast in USB devio code, and thus avoid a compiler warning.

A void* data argument passed to bus_find_device() and thence to match_devt()
is used to carry a 32-bit datum.  However, casting directly between a u32 and
a pointer is not permitted - there must be an intermediate cast via (unsigned)
long.

This was introduced by the following patch:

	commit 94b1c9fa060ece2c8f080583beb6cc6008e41413
	Author: Alan Stern <stern@rowland.harvard.edu>
	Date:   Tue Jun 24 14:47:12 2008 -0400

	    usbfs: simplify the lookup-by-minor routines

	    This patch (as1105) simplifies the lookup-by-minor-number code in
	    usbfs.  Instead of passing the minor number to the callback, which
	    must then reconstruct the entire dev_t value, the patch passes the
	    dev_t value directly.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

g_cdc needs to depend on NET, otherwise net-related build errors happen:

ERROR: "netif_carrier_on" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "netif_carrier_off" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "netif_rx" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "alloc_etherdev_mq" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "free_netdev" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "register_netdev" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "print_mac" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "dev_kfree_skb_any" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "__alloc_skb" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "eth_type_trans" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "unregister_netdev" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "__netif_schedule" [drivers/usb/gadget/g_cdc.ko] undefined!
ERROR: "skb_put" [drivers/usb/gadget/g_cdc.ko] undefined!
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Copy the OHCI/EHCI PM callbacks of the PCI implementation since
they work equally well on Au1xxx hardware.

Tested on Au1200.
Signed-off-by: NManuel Lauss <mano@roarinelk.homelinux.net>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

- Fold multiple probe/remove callbacks into one function;
- minor style fixes, no functional changes.

Tested on Au1200.
Signed-off-by: NManuel Lauss <mano@roarinelk.homelinux.net>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1107) fixes a small bug in the usbfs registration and
unregistration code.  It avoids leaving an error value stored in the
device's usb_classdev field and it avoids trying to unregister a NULL
pointer.  (It also fixes a rather extreme overuse of whitespace.)
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1106) fixes a race between opening and unregistering
device files in usbfs.  The current code drops its reference to the
device and then reacquires it, ignoring the possibility that the
device structure might have been removed in the meantime.  It also
doesn't check whether the device is already in the NOTATTACHED state
when the file is opened.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1105) simplifies the lookup-by-minor-number code in
usbfs.  Instead of passing the minor number to the callback, which
must then reconstruct the entire dev_t value, the patch passes the
dev_t value directly.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

USB device files are accessible in two ways: as files in usbfs and as
character device nodes.  The two paths are supposed to behave
identically, but they don't.  When the underlying USB device is
unplugged, disconnect signals are sent to processes with open usbfs
files (if they requested these signals) but not to processes with open
device node files.

This patch (as1104) fixes the bug by moving the disconnect-signalling
code into a common subroutine which is called from both paths.
Putting this subroutine in devio.c removes the only out-of-file
reference to struct dev_state, and so the structure's declaration can
be moved from usb.h into devio.c.

Finally, the new subroutine performs one extra action: It kills all
the outstanding async URBs.  (I'd kill the outstanding synchronous
URBs too, if there was any way to do it.)  In the past this hasn't
mattered much, because devices were unregistered from usbfs only
when they were disconnected.  But now the unregistration can also
occur whenever devices are unbound from the usb_generic driver.  At
any rate, killing URBs when a device is unregistered from usbfs seems
like a good thing to do.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as1024) takes care of a FIXME issue: Drivers that don't
have the necessary suspend, resume, reset_resume, pre_reset, or
post_reset methods will be unbound and their interface reprobed when
one of the unsupported events occurs.

This is made slightly more difficult by the fact that bind operations
won't work during a system sleep transition.  So instead the code has
to defer the operation until the transition ends.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Various cleanups and fixes to the i2c code in ohci-pnx4008:
* Delete empty isp1301_command. The i2c driver command implementation
  is optional, so there's no point in providing an empty
  implementation.
* Give a name to isp1301_driver. I'm surprised that i2c-core accepted
  to register this driver at all. I've chosen "isp1301_pnx" as the
  name, because it's not a generic ISP1301 driver (much like the
  isp1301_omap driver.) We might want to make the name even more
  specific (but "isp1301_ohci_pnx4008" doesn't fit.)
* The ISP1301 is definitely not a hardware monitoring device.
* Fix a memory leak on failure in isp1301_attach. If
  i2c_attach_client fails, the client is not registered so
  isp1301_detach is never called and the i2c_client memory is lost.
* Use strlcpy instead of strcpy.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Vitaly Wool <vitalywool@gmail.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

USB debug port only supports 8 byte rx/tx packets. Although spec implies that
"if a packet larger than eight bytes is received from the remote computer, the
device must break the larger packet into eight-byte packets before sending the
data to the Debug Port", the real PLX NET20DC device does not handle it right -
data is corrupted on debug port end if serial interface sends >8 byte urbs.
Patch below fixes the issue by limiting tx urb to 8 byte.

Signed off by: Aleks Gorelov <dared1st@yahoo.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

cdc-acm must give up secondary interfaces if the primary is disconnected
and vice versa. This wasn't done correctly.
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

usb serial decrements the pm counter even if an interface has been
disconnected. If it was a logical disconnect the interface may belong
already to another driver. This patch introduces a check for disconnected
interfaces.
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Cc: Stable <stable@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

this patch implements

- suspend/resume
- aggressive autosuspend for the cdc-wdm driver
- pre/post_reset
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

this patch saves power for cdc-acm devices that support remote wakeup
while the device is connected.

- request needs_remote_wakeup when needed
- delayed write while a device is autoresumed
- the device is marked busy when appropriate
Signed-off-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The 28xb, as documented in comments, has the same ID's as the 28x.
Remove the duplicated ID's from the device tables, and expand the
comment to document this.
Signed-off-by: NBen Collins <ben.collins@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>