Instead of using a "I kmalloced this" flag, we keep track of the kmalloced
strings and use that list to check if we need to kfree (in practice, the
list is very short).

This means that kparams can be const again, and plugs a leak.  This
is important for drivers/usb/gadget/nokia.c which gets modprobe/rmmod'ed
frequently on the N9000.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Cc: Artem Bityutskiy <dedekind1@gmail.com>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>

This allows us to generalize the KPARAM_KMALLOCED flag, by calling a function
on every parameter when a module is unloaded.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>

Ideally, we'd check that it was only the "set" function which was __init,
and that the permissions were r/o.  But that's a little hard.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Acked-by: NSam Ravnborg <sam@ravnborg.org>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>

Today's next 20091117 build failed on s390 with

drivers/char/hvc_iucv.c:1331: error: 'param_ops_vmidfilter' undeclared here (not in a function)
make[2]: *** [drivers/char/hvc_iucv.o] Error 1

Most probably caused by commit 684a6d340b8a5767db4670031b0f39455346018a
(param:param_ops) which introduced a param_ops structure.

The following compile tested patch adds a param_ops structure for hvc_iucv.
Signed-off-by: NSachin Sant <sachinp@in.ibm.com>
Acked-by: NHeiko Carstens <heiko.carstens@de.ibm.com>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

After merging the rr tree, today's linux-next build (powerpc
ppc64_defconfig) failed like this:

net/sunrpc/auth.c:74: error: 'param_ops_hashtbl_sz' undeclared here (not in a function)

Caused by commit 0685652df0929cec7d78efa85127f6eb34962132
("param:param_ops") interacting with commit
f8f853ab19fcc415b6eadd273373edc424916212 ("SUNRPC: Make the credential
cache hashtable size configurable") from the nfs tree.
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

Fixes these build errors:
security/apparmor/lsm.c:701: error: 'param_ops_aabool' undeclared here (not in a function)
security/apparmor/lsm.c:721: error: 'param_ops_aalockpolicy' undeclared here (not in a function)
security/apparmor/lsm.c:729: error: 'param_ops_aauint' undeclared here (not in a function)
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NJohn Johansen <john.johansen@canonical.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

This is more kernel-ish, saves some space, and also allows us to
expand the ops without breaking all the callers who are happy for the
new members to be NULL.

The few places which defined their own param types are changed to the
new scheme (more which crept in recently fixed in following patches).

Since we're touching them anyway, we change get() and set() to take a
const struct kernel_param (which they really are).  This causes some
harmless warnings until we fix them (in following patches).

To reduce churn, module_param_call creates the ops struct so the callers
don't have to change (and casts the functions to reduce warnings).
The modern version which takes an ops struct is called module_param_cb.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Ville Syrjala <syrjala@sci.fi>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: Alessandro Rubini <rubini@ipvvis.unipv.it>
Cc: Michal Januszewski <spock@gentoo.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Neil Brown <neilb@suse.de>
Cc: linux-kernel@vger.kernel.org
Cc: linux-input@vger.kernel.org
Cc: linux-fbdev-devel@lists.sourceforge.net
Cc: linux-nfs@vger.kernel.org
Cc: netdev@vger.kernel.org

This is modern style, and good to do before we start changing things.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Reviewed-by: NTakashi Iwai <tiwai@suse.de>
Tested-by: NPhil Carmody <ext-phil.2.carmody@nokia.com>

An audit by Dongdong Deng revealed that most driver-author-written param
calls don't handle val == NULL (which happens when parameters are specified
with no =, eg "foo" instead of "foo=1").

The only real case to use this is boolean, so handle it specially for that
case and remove a source of bugs for everyone else.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Cc: Dongdong Deng <dongdong.deng@windriver.com>
Cc: Américo Wang <xiyou.wangcong@gmail.com>

Hey, at least it has both l's.
Reported-by: NMarin Mitov <mitov@issp.bas.bg>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

* 'for-linus' of git://neil.brown.name/md: (24 commits)
  md: clean up do_md_stop
  md: fix another deadlock with removing sysfs attributes.
  md: move revalidate_disk() back outside open_mutex
  md/raid10: fix deadlock with unaligned read during resync
  md/bitmap:  separate out loading a bitmap from initialising the structures.
  md/bitmap: prepare for storing write-intent-bitmap via dm-dirty-log.
  md/bitmap: optimise scanning of empty bitmaps.
  md/bitmap: clean up plugging calls.
  md/bitmap: reduce dependence on sysfs.
  md/bitmap: white space clean up and similar.
  md/raid5: export raid5 unplugging interface.
  md/plug: optionally use plugger to unplug an array during resync/recovery.
  md/raid5: add simple plugging infrastructure.
  md/raid5: export is_congested test
  raid5: Don't set read-ahead when there is no queue
  md: add support for raising dm events.
  md: export various start/stop interfaces
  md: split out md_rdev_init
  md: be more careful setting MD_CHANGE_CLEAN
  md/raid5: ensure we create a unique name for kmem_cache when mddev has no gendisk
  ...

As pointed out by Jiri Slaby: when I resolved the the 32-bit x85 system
call entry tables for prlimit (due to the conflict with fanotify), I
forgot to add the numbering in comments that we do for every fifth entry.
Reported-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix kernel-doc warnings in linux/i2c.h:

  Warning(include/linux/i2c.h:176): No description found for parameter 'alert'
  Warning(include/linux/i2c.h:259): No description found for parameter 'of_node'
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix kernel-doc warning, add @timer description:

  Warning(kernel/timer.c:335): No description found for parameter 'timer'
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

'make htmldocs' produces errors due to missing a supporting media
file, so add 'xmldoclinks' to the htmldocs dependencies so that the
needed supporting file will be present.

  Documentation/DocBook/media.xml:4: warning: failed to load external entity "Documentation/DocBook/media-entities.tmpl"
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Acked-by: NMauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

'make htmldocs' has a fatal error when processing libata.xml, as seen
below.  The string "Example patterns:" (or any string with "example.*:"
in it AFAIK) causes some part of the doc generation tool chain to try to
produce an <informalexample> block without a beginning <para>, but
there is an ending </para> generated, which throws things out of kilter.

I don't even know where (what program) this is happening in.
I searched in docproc and xmlto and in some XML stylesheets without
finding anything.  If anyone can give me pointers about this, please do.
Until this is fixed, let's just spell "Example" as "Sample"
and match up the double quotation marks while there.

  Documentation/DocBook/libata.xml:6575: parser error : Opening and ending tag mismatch: programlisting line 6573 and para
     </para><para>
            ^
  Documentation/DocBook/libata.xml:6580: parser error : Opening and ending tag mismatch: para line 6575 and programlisting
  </programlisting></informalexample>
                   ^
  unable to parse Documentation/DocBook/libata.xml
  make[2]: *** [Documentation/DocBook/libata.html] Error 1
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It's currently possible to bypass xattr namespace access rules by
prefixing valid xattr names with "os2.", since the os2 namespace stores
extended attributes in a legacy format with no prefix.

This patch adds checking to deny access to any valid namespace prefix
following "os2.".
Signed-off-by: NDave Kleikamp <shaggy@linux.vnet.ibm.com>
Reported-by: NSergey Vlasov <vsu@altlinux.ru>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

* 'for-2.6.36' of git://git.kernel.dk/linux-2.6-block: (149 commits)
  block: make sure that REQ_* types are seen even with CONFIG_BLOCK=n
  xen-blkfront: fix missing out label
  blkdev: fix blkdev_issue_zeroout return value
  block: update request stacking methods to support discards
  block: fix missing export of blk_types.h
  writeback: fix bad _bh spinlock nesting
  drbd: revert "delay probes", feature is being re-implemented differently
  drbd: Initialize all members of sync_conf to their defaults [Bugz 315]
  drbd: Disable delay probes for the upcomming release
  writeback: cleanup bdi_register
  writeback: add new tracepoints
  writeback: remove unnecessary init_timer call
  writeback: optimize periodic bdi thread wakeups
  writeback: prevent unnecessary bdi threads wakeups
  writeback: move bdi threads exiting logic to the forker thread
  writeback: restructure bdi forker loop a little
  writeback: move last_active to bdi
  writeback: do not remove bdi from bdi_list
  writeback: simplify bdi code a little
  writeback: do not lose wake-ups in bdi threads
  ...

Fixed up pretty trivial conflicts in drivers/block/virtio_blk.c and
drivers/scsi/scsi_error.c as per Jens.

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (94 commits)
  V4L/DVB: tvp7002: fix write to H-PLL Feedback Divider LSB register
  V4L/DVB: dvb: siano: free spinlock before schedule()
  V4L/DVB: media: video: pvrusb2: remove custom hex_to_bin()
  V4L/DVB: drivers: usbvideo: remove custom implementation of hex_to_bin()
  V4L/DVB: Report supported QAM modes on bt8xx
  V4L/DVB: media: ir-keytable: null dereference in debug code
  V4L/DVB: ivtv: convert to the new control framework
  V4L/DVB: ivtv: convert gpio subdev to new control framework
  V4L/DVB: wm8739: convert to the new control framework
  V4L/DVB: cs53l32a: convert to new control framework
  V4L/DVB: wm8775: convert to the new control framework
  V4L/DVB: cx2341x: convert to the control framework
  V4L/DVB: cx25840: convert to the new control framework
  V4L/DVB: cx25840/ivtv: replace ugly priv control with s_config
  V4L/DVB: saa717x: convert to the new control framework
  V4L/DVB: msp3400: convert to the new control framework
  V4L/DVB: saa7115: convert to the new control framework
  V4L/DVB: v4l2: hook up the new control framework into the core framework
  V4L/DVB: Documentation: add v4l2-controls.txt documenting the new controls API
  V4L/DVB: v4l2-ctrls: Whitespace cleanups
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  watchdog: hpwdt: formatting of pointers in printk()
  watchdog: Adding support for ARM Primecell SP805 Watchdog
  watchdog: f71808e_wdt: new watchdog driver for Fintek F71808E and F71882FG
  watchdog: sch311x_wdt.c: set parent before registeriing the misc device in probe() function
  watchdog: wdt_pci.c: move ids to pci_ids.h
  watchdog: s3c2410_wdt - Fix removing of platform device

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/cxgb4: Obtain RDMA QID ranges from LLD/FW

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: xpad - add USB-ID for PL-3601 Xbox 360 pad
  Input: cy8ctmg100_ts - signedness bug
  Input: elantech - report position also with 3 fingers
  Input: elantech - discard the first 2 positions on some firmwares
  Input: adxl34x - do not mark device as disabled on startup
  Input: gpio_keys - add hooks to enable/disable device
  Input: evdev - rearrange ioctl handling
  Input: dynamically allocate ABS information
  Input: switch to input_abs_*() access functions
  Input: add static inline accessors for ABS properties

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (148 commits)
  USB: serial: fix stalled writes
  USB: remove fake "address-of" expressions
  USB: fix thread-unsafe anchor utiliy routines
  USB: usbtest: support test device with only one iso-in or iso-out endpoint
  USB: usbtest: avoid to free coherent buffer in atomic context
  USB: xhci: Set DMA mask for host.
  USB: xhci: Don't flush doorbell writes.
  USB: xhci: Reduce reads and writes of interrupter registers.
  USB: xhci: Make xhci_set_hc_event_deq() static.
  USB: xhci: Minimize HW event ring dequeue pointer writes.
  USB: xhci: Make xhci_handle_event() static.
  USB: xhci: Remove unnecessary reads of IRQ_PENDING register.
  USB: xhci: Performance - move xhci_work() into xhci_irq()
  USB: xhci: Performance - move interrupt handlers into xhci-ring.c
  USB: xhci: Performance - move functions that find ep ring.
  USB:: fix linux/usb.h kernel-doc warnings
  USB: add USB serial ssu100 driver
  USB: usb-storage: implement autosuspend
  USB: ehci: fix remove of ehci debugfs dir
  USB: Add USB 2.0 to ssb ohci driver
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6: (68 commits)
  U6715 16550A serial driver support
  Char: nozomi, set tty->driver_data appropriately
  Char: nozomi, fix tty->count counting
  serial: max3107: Fix gpiolib support
  hsu: call PCI pm hooks in suspend/resume function
  hsu: some code cleanup
  hsu: add a periodic timer to check dma rx channel
  hsu: driver for Medfield High Speed UART device
  mxser: remove unnesesary NULL check
  serial: add support for OX16PCI958 card
  serial: 68328serial.c: remove dead (ALMA_ANS | DRAGONIXVZ | M68EZ328ADS)
  timbuart: use __devinit and __devexit macros for probe and remove
  serial: MMIO32 support for 8250_early.c
  serial: mcf: don't take spinlocks in already protected functions
  serial: general fixes in the serial_rs485 structure
  serial: fix missing bit coverage of ASYNC_FLAGS
  serial: "altera_uart: simplify altera_uart_console_putc()" checkpatch fixes
  serial: crisv10: formatting of pointers in printk()
  vt: Fix warning: statement with no effect due to vt_kern.h
  tty_io: remove casts from void*
  ...

As David VomLehn points out, it was possible to receive an interrupt
before clearing the free-urb flag which could lead to the urb being
incorrectly marked as busy.

For the same reason, move tx_bytes accounting so that it will never be
negative.

Note that the free-flags set and clear operations do not need any
additional locking as they are manipulated while USB_SERIAL_WRITE_BUSY
is set.
Reported-by: NDavid VomLehn <dvomlehn@cisco.com>
Tested-by: NDavid VomLehn <dvomlehn@cisco.com>
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Cc: stable <stable@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Fake "address-of" expressions that evaluate to NULL generally confuse
readers and can provoke compiler warnings.  This patch (as1412)
removes three such fake expressions, using "#ifdef"s in their place.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch fixes a race condition in two utility routines
related to the removal/unlinking of urbs from an anchor.

If two threads are concurrently accessing the same anchor,
both could end up with the same urb - thinking they are
the exclusive owner.

Alan Stern pointed out a related issue in
usb_unlink_anchored_urbs:

"The URB isn't removed from the anchor until it completes
 (as a by-product of completion, in fact), which might not
 be for quite some time after the unlink call returns.
 In the meantime, the subroutine will keep trying to unlink
 it, over and over again."

Cc: stable <stable@kernel.org>
Cc: Oliver Neukum <oneukum@suse.de>
Cc: Greg Kroah-Hartman <greg@kroah.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NChristian Lamparter <chunkeey@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

It is very common that one altsetting may include only one iso-in or iso-out
single endpoint, especially for high bandwidth endpoint, so support it.
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch fixes the warning below:
[30753.755998] ------------[ cut here ]------------
[30753.755998] WARNING: at /home/tom/git/linux-2.6/linux-2.6-next/arch/x86/include/asm/dma-mapping.h:155 hcd_buffer_free+0xb1/0xd4 [usbcore]()
[30753.755998] Hardware name: 6475EK2
[30753.755998] Modules linked in: uvcvideo ehci_hcd usbtest cdc_ether usbnet vfat fat usb_storage nfsd lockd nfs_acl auth_rpcgss exportfs mii tun videodev v4l1_compat v4l2_compat_ioctl32 fuse bridge stp llc sunrpc ipv6 cpufreq_ondemand acpi_cpufreq freq_table mperf kvm_intel kvm arc4 ecb ath5k usbhid mac80211 snd_hda_codec_conexant ch341 usbserial ath cfg80211 thinkpad_acpi snd_hda_intel pcspkr wmi hwmon yenta_socket iTCO_wdt iTCO_vendor_support i2c_i801 e1000e snd_hda_codec snd_hwdep snd_pcm snd_timer snd soundcore snd_page_alloc pata_acpi uhci_hcd ohci_hcd usbcore i915 drm_kms_helper drm i2c_algo_bit i2c_core video output [last unloaded: uvcvideo]
[30753.755998] Pid: 0, comm: swapper Tainted: G        W   2.6.35-rc6-gkh-wl+ #49
[30753.755998] Call Trace:
[30753.755998]  <IRQ>  [<ffffffff8104478a>] warn_slowpath_common+0x80/0x98
[30753.755998]  [<ffffffff810447b7>] warn_slowpath_null+0x15/0x17
[30753.755998]  [<ffffffffa00ce02d>] hcd_buffer_free+0xb1/0xd4 [usbcore]
[30753.755998]  [<ffffffffa00c1345>] usb_free_coherent+0x1c/0x1e [usbcore]
[30753.755998]  [<ffffffffa00b13e4>] simple_free_urb+0x23/0x2f [usbtest]
[30753.755998]  [<ffffffffa00b210b>] iso_callback+0xbb/0x10f [usbtest]
[30753.755998]  [<ffffffffa00c7390>] usb_hcd_giveback_urb+0x8c/0xc0 [usbcore]
[30753.755998]  [<ffffffffa0449b35>] ehci_urb_done+0x84/0x95 [ehci_hcd]
[30753.755998]  [<ffffffffa044b5a5>] ehci_work+0x41a/0x7dd [ehci_hcd]
[30753.755998]  [<ffffffffa044e298>] ehci_irq+0x33b/0x370 [ehci_hcd]
[30753.755998]  [<ffffffff8100fb05>] ? sched_clock+0x9/0xd
[30753.755998]  [<ffffffff8105e641>] ? sched_clock_local+0x1c/0x82
[30753.755998]  [<ffffffff8105e76a>] ? sched_clock_cpu+0xc3/0xce
[30753.755998]  [<ffffffff81067c7e>] ? trace_hardirqs_off+0xd/0xf
[30753.755998]  [<ffffffff8105e7b8>] ? cpu_clock+0x43/0x5e
[30753.755998]  [<ffffffffa00c6999>] usb_hcd_irq+0x45/0xa1 [usbcore]
[30753.755998]  [<ffffffff81092e02>] handle_IRQ_event+0x20/0xa5
[30753.755998]  [<ffffffff81094cea>] handle_fasteoi_irq+0x92/0xd2
[30753.755998]  [<ffffffff8100c0ed>] handle_irq+0x1f/0x2a
[30753.755998]  [<ffffffff8100b75d>] do_IRQ+0x57/0xbe
[30753.755998]  [<ffffffff8136a693>] ret_from_intr+0x0/0x16
[30753.755998]  <EOI>  [<ffffffff81223baa>] ? acpi_idle_enter_bm+0x231/0x269
[30753.755998]  [<ffffffff81223ba3>] ? acpi_idle_enter_bm+0x22a/0x269
[30753.755998]  [<ffffffff812c4b6b>] cpuidle_idle_call+0x99/0xce
[30753.755998]  [<ffffffff81008dd5>] cpu_idle+0x61/0xaa
[30753.755998]  [<ffffffff8136374b>] start_secondary+0x1c2/0x1c6
[30753.755998] ---[ end trace 904cfaf7ab4cb1a2 ]---
Signed-off-by: NMing Lei <tom.leiming@gmail.com>
Cc: stable <stable@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Tell the USB core that we can do DMA directly (instead of needing it to
memory-map the buffers for PIO).  If the xHCI host supports 64-bit addresses,
set the DMA mask accordingly.  Otherwise indicate the host can handle 32-bit DMA
addresses.

This improves performance because the USB core doesn't have to spend time
remapping buffers in high memory into the 32-bit address range.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

To tell the host controller that there are transfers on the endpoint
rings, we need to ring the endpoint doorbell.  This is a PCI MMIO write,
which can be delayed until another register read is queued.

The previous code would flush the doorbell write by reading the doorbell
register after the write.  This may take time, and it's not necessary to
force the host controller to know about the transfers right away.  Don't
flush the doorbell register writes.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The interrupter register set includes a register that says whether interrupts
are pending for each event ring (the IP bit).  Each MSI-X vector will get its
own interrupter set with separate IP bits.  The status register includes an
"Event Interrupt (EINT)" bit that is set when an IP bit is set in any of the
interrupters.

When PCI interrupts are used, the EINT bit exactly mirrors the IP bit in the
single interrupter set, and it is a waste of time to check both registers when
trying to figure out if the xHC interrupted or another device on the shared IRQ
line interrupted.  Only check the IP bit to reduce register reads.

The IP bit is automatically cleared by the xHC when MSI or MSI-X is enabled.  It
doesn't make sense to read that register to check for shared interrupts (since
MSI and MSI-X aren't shared).  It also doesn't make sense to write to that
register to clear the IP bit, since it is cleared by the hardware.

We can tell whether MSI or MSI-X is enabled by looking at the irq number in
hcd->irq.  If it's -1, we know MSI or MSI-X is enabled.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Now that the event handler functions no longer use xhci_set_hc_event_deq()
to update the event ring dequeue pointer, that function is not used by
anything in xhci-ring.c.  Move that function into xhci-mem.c and make it
static.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The xHCI specification suggests that writing the hardware event ring dequeue
pointer register too often can be an expensive operation for the xHCI hardware
to manage.  It suggests minimizing the number of writes to that register.

Originally, the driver wrote the event ring dequeue pointer after each
event was processed.  Depending on how the event ring moderation register
is set up and how fast the transfers are completing, there may be several
events processed for each interrupt.  This patch makes the hardware event
ring dequeue pointer be written only once per interrupt.

Make the transfer event handler and port status event handler only write
the software event ring dequeue pointer.  Move the updating of the
hardware event ring dequeue pointer into the interrupt function.  Move the
contents of xhci_set_hc_event_deq() into the interrupt handler.  The
interrupt handler must clear the event handler busy flag, so it might as
well also write the dequeue pointer to the same register.  This eliminates
two 32-bit PCI reads and two 32-bit PCI writes.
Reported-by: NAndiry Xu <andiry.xu@amd.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

xhci_handle_event() is now only called from within xhci-ring.c, so make it
static.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Remove a duplicate register read of the interrupt pending register from
xhci_irq().  Also, remove waiting on the posted write of that register.
The host will see it eventually.  It will probably read the register
itself before deciding whether to interrupt the system again, forcing the
posted write to complete.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

When we move xhci_work() into xhci_irq(), we don't need to read the operational
register status field twice.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Most of the work for interrupt handling is done in xhci-ring.c, so it makes
sense to move the functions that are first called when an interrupt happens
(xhci_irq() or xhci_msi_irq()) into xhci-ring.c, so that the compiler can better
optimize them.

Shorten some lines to make it pass checkpatch.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

I've been using perf to measure the top symbols while transferring 1GB of data
on a USB 3.0 drive with dd.  This is using the raw disk with /dev/sdb, with a
block size of 1K.

During performance testing, the top symbol was xhci_triad_to_transfer_ring(), a
function that should return immediately if streams are not enabled for an
endpoint.  It turned out that the functions to find the endpoint ring was
defined in xhci-mem.c and used in xhci-ring.c and xhci-hcd.c.  I moved a copy of
xhci_triad_to_transfer_ring() and xhci_urb_to_transfer_ring() into xhci-ring.c
and declared them static.  I also made a static version of
xhci_urb_to_transfer_ring() in xhci.c.

This improved throughput on a 1GB read of the raw disk with dd from
186MB/s to 195MB/s, and perf reported sampling the xhci_triad_to_transfer_ring()
0.06% of the time, rather than 9.26% of the time.
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>