Fix UDMA throughput bug:  tCYC averages t2CYCTYP/2, but the code
previously assumed it was the same as t2CYCTYP.  (That is, it was
using just one clock edge, not both.)  Move the table's type
declaration so it's adjacent to the table, making it more clear
what those numbers mean.

On one system this change increased throughput by almost 4x: UDMA/66
sometimes topped 23 MB/sec (on a drive known to do much better).  On
another system it was around a 10% win (UDMA/66 up to 7+ MB/sec).

The difference might be caused by the ratio between memory and IDE
clocks.  In the system with large speedup, this was exactly 2 (as a
workaround for a rev 1.1 silicon bug).  The other system used a more
standard ratio of 1.63 (and rev 2.1 silicon) ... clock domain synch
might have some issues, they're not unheard-of.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

The driver somehow got merged with the initializer for the dma_sff_read_status()
method missing which caused kernel panic on bootup.

This should fix the kernel.org bug #13026...
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Reported-by: Arnd Hannemann <hannemann@nets.rwth-aachen.de>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Freeing non-slab objects is bad and results in an oops.  Fix it.
Reported-and-tested-by: Andrew Price <andy@andrewprice.me.uk>
Cc: Theodore Tso <tytso@mit.edu>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Remove uneeded void casts
Signed-off-by: Jack Stone <jwjstone@fastmail.fm>
Cc: jeff@garzik.org
Cc: kernel-janitors@vger.kernel.org
Cc: Jack Stone <jwjstone@fastmail.fm>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use ATA_DMA_* constants instead of the bare numbers for the BMIDE register bits.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

The big driver change in 2.4.19-rc1 introduced a regression for many HPT370[A]
chips -- DMA stopped to work completely, only causing endless timeouts...

The culprit has been identified (at last!): it turned to be the code resetting
the DMA state machine before each transfer. Stop doing it now as this counter-
measure has clearly caused more harm than good.

This should fix the kernel.org bug #7703.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  [libata] fix build error on drivers/ata/pata_legacy.c
  pata_via: Cache and rewrite the device bit
  sata_mv: workaround for multi_count errata sata24
  sata_mv: tidy up qc->tf usage in qc_prep() functions

fix those errors:

drivers/ata/pata_legacy.c: In function ‘pdc_data_xfer_vlb’:
drivers/ata/pata_legacy.c:289: error: ‘ap’ undeclared (first use in this function)
drivers/ata/pata_legacy.c:289: error: (Each undeclared identifier is reported only once
drivers/ata/pata_legacy.c:289: error: for each function it appears in.)
drivers/ata/pata_legacy.c: At top level:
drivers/ata/pata_legacy.c:869: error: ‘ATA_PFLAG_PIO32_CHANGE’ undeclared here (not in a
+function)
make[2]: *** [drivers/ata/pata_legacy.o] Error 1
make[1]: *** [drivers/ata] Error 2
Signed-off-by: Zhenwen Xu <helight.xu@gmail.com>
Acked-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Some VIA chipsets will reset the DEV bit after IEN changes on ctl. Our
optimised write path avoids doing this but we need to remove the
optimisation on these devices.

[Identified and some original patches proposed by Josehn Chan @ VIA but
discussion then all ground to a halt so given a test case I dug it back out]

Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk
Tested-by: Christoph Bisping (bug #13086)
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Workaround for errata SATA#24 in sata_mv.
This errata affects WRITE_MULTI* commands when
the device multi_count produces a DRQ block size >= 4Kbytes.

We work around it here by converting such operations
into ordinary PIO_WRITEs instead.

Note that this might result in a PIO FUA write unavoidably being converted
into a non-FUA write.  In practice, any system using FUA is also going to be
using DMA rather than PIO, so this shouldn't affect anyone in the real world.
Signed-off-by: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Tidy up qc->tf accesses in the mv_qc_prep() functions.
Signed-off-by: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (43 commits)
  staging: slicoss: update README
  otus/zdusb.c: additional USB idnetifier
  Staging: go7007: fix build issues
  Staging: sxg: Fix leaks and checksum errors in transmit code path
  Staging: sxg: Fix sleep in atomic context warning while loading driver
  Staging: sxg: Use correct queue_id for transmitting non-TCP packets
  Staging: sxg: Fire watchdog timer at end of open routine to change the link
  Staging: Pohmelfs: Add load balancing between network states with the same priority.
  Staging: Pohmelfs: Added IO permissions and priorities.
  Staging: Pohmelfs: Added ->show_stats() callback.
  Staging: Pohmelfs: Drop ftrans debugging code.
  Staging: Pohmelfs: Use wait_on_page_timeout when waiting for remote directory sync instead of hardcoded 25 seconds.
  Staging: Pohmelfs: Reduce debugging noise about non-existing objects.
  Staging: Pohmelfs: Sync fs before killing it, since dentry cache is shrunk before writeback is invoked via generic_shutdown_super()
  Staging: Pohmelfs: Extend remount option.
  Staging: Pohmelfs: Set NETFS_INODE_REMOTE_SYNCED and clear NETFS_INODE_OWNED bits in the root inode.
  Staging: Pohmelfs: Added 'need_lock' variable into debug print.
  Staging: Pohmelfs: Disable read lock in pohmelfs_getattr().
  Staging: Pohmelfs: Move parent lock to the place where we really have to send a lookup request to the server.
  Staging: pohmelfs: Populate dentry cache when receiving the new readdir entry.
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6:
  UIO: fix specific device driver missing statement for depmod
  Driver core: remove pr_fmt() from dynamic_dev_dbg() printk
  driver core: prevent device_for_each_child from oopsing
  dynamic debug: resurrect old pr_debug() semantics as pr_devel()
  Driver Core: early platform driver
  proc: mounts_poll() make consistent to mdstat_poll
  sysfs: sysfs poll keep the poll rule of regular file.
  driver core: allow non-root users to listen to uevents
  driver core: fix driver_match_device
  sysfs: don't use global workqueue in sysfs_schedule_callback()

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (22 commits)
  WUSB: correct format of wusb_chid sysfs file
  WUSB: fix oops when completing URBs for disconnected devices
  WUSB: disconnect all devices when stopping a WUSB HCD
  USB: whci-hcd: check return value of usb_hcd_link_urb_to_ep()
  USB: whci-hcd: provide a endpoint_reset method
  USB: add reset endpoint operations
  USB device codes for Motorola phone.
  usb-storage: fix mistake in Makefile
  USB: usb-serial ch341: support for DTR/RTS/CTS
  Revert USB: usb-serial ch341: support for DTR/RTS/CTS
  USB: musb: fix possible panic while resuming
  USB: musb: fix isochronous TXDMA (take 2)
  USB: musb: sanitize clearing TXCSR DMA bits (take 2)
  USB: musb: bugfixes for multi-packet TXDMA support
  USB: musb_host, fix ep0 fifo flushing
  USB: usb-storage: augment unusual_devs entry for Simple Tech/Datafab
  USB: musb_host, minor enqueue locking fix (v2)
  USB: fix oops in cdc-wdm in case of malformed descriptors
  USB: qcserial: Add extra device IDs
  USB: option: Add ids for D-Link DWM-652 3.5G modem
  ...

* master.kernel.org:/home/rmk/linux-2.6-arm: (48 commits)
  [ARM] S3C24XX: ADC: Check pending queue before freeing adc client
  [ARM] S3C: Fix ADC driver sparse warning
  [ARM] Osiris: Fix double initialisation in machine block
  [ARM] Anubis: Fix sparse warnings for items that should be static
  [ARM] JIVE: Fix sparse warnings about items which should be static
  [ARM] S3C: Fix sparse warning from missing 's3c_device_hwmon'
  [ARM] S3C24XX: Fix sparse error in gpiolib.c
  [ARM] 5455/1: Fix IRQ noise from VIC code
  [ARM] 5454/1: ep93xx_eth: fix sparse warnings
  [ARM] remove .gitignore from include/asm-arm
  Update MAINTAINERS
  mxc defconfig updates
  mx31ads: Mark as having full regulatoion constraints with 1133-EV1 board
  mx31ads: Depend on all the WM8350 core dependencies for WM1133-EV1 board
  Fix ifdef in plat-mxc/irc.c
  MX1ADS: remove I2C ifdefs
  qong: remove AIPS[12] mappings from machine-specific iotable
  mx31ads: imoux pins should be passed in as unsigned int
  MXC: remove orphan imx_init_uart() definition
  mx31: pin definition for csi
  ...

I looked, I gagged, I left
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Provide support for WN111v2 USB 802.11n adapter.
Signed-off-by: Daniele Napolitano <dnax88@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Now that TUNER_SET_TYPE_ADDR is gone from the tree, the older code kicks
in and tries to use TUNER_SET_TYPE, which went away a long time ago.

This patch removes all of this logic, as it should not be needed anymore
now, and by doing so, fixes the build.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Fix the transmit function for the following:
* Free XmtCmd in the error code path. This use to leak memory in
  error conditions.
* Do pci mapping after the checksum operations are over. They can
  reallocate the skb at a different location.
* Fix UDP checksum errors which were seen in wireshark
Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Leave a spinlock before calling request_irq(). request_irq() calls kmalloc
which can sleep. This was generating a warning dump while driver is loaded.
Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Use correct queue_id while transmitting non-TCP packets.
They should always use queue 0.
Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>

The watchdog timer which updates the link status was not fired at the
end of sxg_entry_open(). Add that.
Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: Pohmelfs: Use wait_on_page_timeout when waiting for remote directory sync instead of hardcoded 25 seconds.
Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: Pohmelfs: Sync fs before killing it, since dentry cache is shrunk before writeback is invoked via generic_shutdown_super()
Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: Pohmelfs: Set NETFS_INODE_REMOTE_SYNCED and clear NETFS_INODE_OWNED bits in the root inode.
Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: Pohmelfs: Move parent lock to the place where we really have to send a lookup request to the server.
Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Revert commit 2bb34736

This commit has been reported to cause problems:
  Mar 24 11:50:31 linuxdev kernel: BUG: sleeping function called from invalid context at mm/slub.c:1599 Mar 24 11:50:31 linuxdev kernel: in_atomic(): 1, irqs_disabled(): 0, pid: 3251, name: avahi-daemon

Cc: Lior Dotan <liodot@gmail.com>
Cc: Christopher Harrer <charrer@alacritech.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

m68k allmodconfig:
| drivers/staging/stlc45xx/stlc45xx.c: In function 'stlc45xx_probe':
| drivers/staging/stlc45xx/stlc45xx.c:2456: error: implicit declaration of function 'set_irq_type'
| make[6]: *** [drivers/staging/stlc45xx/stlc45xx.o] Error 1
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

If a transaction that contains a file descriptor fails on a later object,
the new file descriptor needs to be closed. If this is a binder file
descriptor we would deadlock in flush. If there were no other references to
the file at this point release would also be called.
Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>