This patch allows transferring of some requests in PIO and some in DMA
mode and defaults to using DMA only for transfers longer than 8 bytes.
This is especially useful with SDIO, which can have lots of 2- and 4-byte
transfers, creating unnecessary high overhead, when executed in DMA.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
On some SDHI tmio implementations the order of DMA and command completion
interrupts swaps, which leads to malfunction. This patch postpones
DMA activation until the MMC command completion IRQ time.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
This patch is based on the commit "af51715079":
* Bugfix to that mmc_send_cxd_data() code: dma-to-stack is
unsafe/nonportable, so kmalloc a bounce buffer instead.
The driver may invalidate the mmc_card->csd when host driver uses dma.
So this subroutine also needs a kmalloc buffer.
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Fix the NO INT in the Multi-BLK IO in SD/MMC, and Multi-BLK read in
SDIO on imx53.
The CMDTYPE of the CMD register (offset 0xE) should be set to "11"
when the STOP CMD12 is issued on imx53 to abort one open ended
multi-blk IO. Otherwise the TC INT wouldn't be generated.
In exact block transfer, the controller doesn't complete the
operations automatically as required at the end of the transfer
and remains on hold if the abort command is not sent on imx53.
As a result, the TC flag is not asserted and SW receives timeout
exception. Set bit1 of Vendor Spec register to fix it.
Signed-off-by: Richard Zhu <Hong-Xing.Zhu@freescale.com>
Signed-off-by: Richard Zhao <richard.zhao@freescale.com>
Reviewed-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
Add one flag to indicate the GPIO CD/WP is enabled or not
on imx platforms, and reuse the writel/readl as the general
APIs for imx SOCs.
Signed-off-by: Richard Zhu <Hong-Xing.Zhu@freescale.com>
Reviewed-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
Add the abort CMDTYPE bits definition of command register (offset 0xE)
Signed-off-by: Richard Zhu <Hong-Xing.Zhu@freescale.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
sdhci-esdhc-imx does not need SDHCI_QUIRK_NO_CARD_NO_RESET.
Make it OF-specific.
Signed-off-by: Richard Zhu <Hong-Xing.Zhu@freescale.com>
Tested-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
Since these are the headers that provide irq_of_parse_and_map()
and NO_IRQ.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Chris Ball <cjb@laptop.org>
New IO FPGA implementation for Versatile Express boards contain
MMCI (PL180) cell with FIFO extended to 128 words (512 bytes).
Matt Waddel reports that this patch improves MMC performance on
his vexpress system, and also fixes "mmcblk0: error -5 transferring
data" errors.
Signed-off-by: Pawel Moll <pawel.moll@arm.com>
Tested-by: Matt Waddel <matt.waddel@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
This patch uses the fixed burst bit when using an internal DMA controller.
I found increased performance with IDMAC when this bit is set.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
The SDHCI driver uses PCI_CLASS_REVISION instead of PCI_REVISION_ID, so
it was not converted by commit 44c10138fd
(PCI: Change all drivers to use pci_device->revision). The newer VIA
driver has similar code too. This patch converts both drivers to use
the 'revision' field of 'struct pci_dev'.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Acked-by: Harald Welte <HaraldWelte@viatech.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Fixes:
drivers/mmc/card/mmc_test.c: In function ‘mmc_test_seq_perf’:
drivers/mmc/card/mmc_test.c:1878:28: warning: variable ‘ts’ set but not
used [-Wunused-but-set-variable]
There's no reason to be calling timespec_sub() here, because
mmc_test_print_avg_rate() is going to do that itself.
Signed-off-by: Chris Ball <cjb@laptop.org>
Cc: Adrian Hunter <adrian.hunter@nokia.com>
This gives much better performance at fast 3x3 rates (up to ~160 Mbps).
The scheduler will still make most decisions about batch size based on
available packets and RX parameters.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Throughout the code we use rate_n_flags & 0xff to extract the lower byte
of the rate_n_flags u32 that contains the information about the rate.
Add a #define and remove the use of the magic number.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since the driver split, there's no need for no_agg_framecnt_info since
all devices have this set to false. Secondly, the compressed block ack
handling code was broken. Fix this.
(1) A shift less than zero simply implies that the buffer wrapped, this
is expected. Remove the incorrect comment.
(2) The (agg->frame_count > (64-sh)) condition can happen if the last
frame is dropped. E.g., if I send 7 frames and the 6th is received but
the 7th is lost, the other side may only shift the window 6, not 7
frames since the last bit is a 0. This is perfectly fine behavior and
doesn't invalidate the feedback.
(3) Store the feedback from a Compressed BA in the first newly received
frame, rather than the start of the window. This way it will get
processed by the rate selection code. Feedback stored in a non-received
frame is likely to get overwritten by the retransmission.
This is based on the approach taken by minstrel_ht.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
When filling out its rate scale table, iwlwifi repeats the first HT rate
IWL_HT_NUMBER_TRY times. The hardware scheduler will stop using
aggregation for any frame that fails LINK_QUAL_AGG_DISABLE_START_DEF
times. Currently, both these constants equal 3.
If iwlwifi probes a faster rate than the link supports, all frames in a
(potentially tens of frames large) batch will fail IWL_HT_NUMBER_TRY
times. Because this happens to be as large as
LINK_QUAL_AGG_DISABLE_START_DEF, all frames will then be sent
individually. This leads to a short, but performance-degrading window
where the legacy stop-and-wait MAC takes over.
Bounding the initial rate by (LINK_QUAL_AGG_DISABLE_START_DEF-1)
attempts makes the third try use a lower rate and hence more be likely
to succeed. This somewhat mitigates the above described behavior.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
iwlagn's hardware scheduler needs to be set up
with the right aggregation frame limit and
buffer sizes. To achieve this, we need to move
the hardware queue setup to when the session
becomes operational.
Tested-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since the driver split, there's no need for
function pointers any more for aggregation
queue setup and teardown as all devices now
share the same code. Simplify this.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
without this patch, we get :
------------[ cut here ]------------
WARNING: at drivers/gpio/gpiolib.c:99 gpio_ensure_requested+0x4c/0x104()
autorequest GPIO-4
.../...
[<c01bcfb0>] (platform_lcd_set_power+0x34/0x40) from [<c033954c>] (platform_lcd_probe+0xb8/0xd8)
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Replace local VPRINTK1 with snd_printdd.
Create local snd_printddd instead of VPRINTK2 for most verbose debug.
In most cases let snd_printk supply default level for messages.
Signed-off-by: Eliot Blennerhassett <eblennerhassett@audioscience.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Make playback silence callback a no-op, card automatically outputs
silence when written data runs out.
Increasing update interval and thus minimum period avoids xrun on startup
or because of timer jitter.
Signed-off-by: Eliot Blennerhassett <eblennerhassett@audioscience.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cards without settable local samplerate and without SRC
still must have a valid samplerate.
This fixed rate is determined by reading the current rate for the card.
Signed-off-by: Eliot Blennerhassett <eblennerhassett@audioscience.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
An update of the futex code had a
WARN_ON(!spin_is_locked(q->lock_ptr))
But on UP, spin_is_locked() is always false, and will
trigger this warning, and even worse, it will exit the function
without doing the necessary work.
Converting this to a WARN_ON_SMP() fixes the problem.
Reported-by: Richard Weinberger <richard@nod.at>
Tested-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
LKML-Reference: <20110317192208.682654502@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Both WARN_ON() and WARN_ON_SMP() should be able to be used in
an if statement.
if (WARN_ON_SMP(foo)) { ... }
Because WARN_ON_SMP() is defined as a do { } while (0) on UP,
it can not be used this way.
Convert it to the same form that WARN_ON() is, even when
CONFIG_SMP is off.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <20110317192208.444147791@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
For codec AD1984A, add a new model to support Dell Precision R5500
or the microphone jack won't work correctly.
BugLink: http://bugs.launchpad.net/bugs/741516
Tested-by: Kent Baxley <kent.baxley@canonical.com>
Cc: stable@kernel.org
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Eric Dumazet reported that hardware PMU events do not work on his
system, due to the BIOS corrupting PMU state:
Performance Events: PEBS fmt0+, Core2 events, Broken BIOS detected, using software events only.
[Firmware Bug]: the BIOS has corrupted hw-PMU resources (MSR 186 is 43003c)
Linus suggested that we continue in the face of such BIOS-induced CPU
state corruption:
http://lkml.org/lkml/2011/3/24/608
Such BIOSes will have to be fixed - Linux developers rely on a working and
fully capable PMU and the BIOS interfering with the CPU's PMU state is simply
not acceptable.
So this patch changes perf to continue when it detects such BIOS
interaction, some hardware events may be unreliable due to the BIOS
writing and re-writing them - there's not much the kernel can do
about that but to detect the corruption and report it.
Reported-and-tested-by: Eric Dumazet <eric.dumazet@gmail.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
It's a big no-no to use pgprot_noncached() when mmap'ing such buffers
into userspace since they are mapped cachable in kernel space.
This can cause all sort of interesting things ranging from to garbled
sound to lockups on various architectures. I've observed that usb-audio
is broken on powerpc 4xx for example because of that.
Also remove the now unused snd_pcm_lib_mmap_noncached(). It's
an arch business to know when to use uncached mappings, there's
already hacks for MIPS inside snd_pcm_default_mmap() and other
archs are supposed to use dma_mmap_coherent().
(See my separate patch that adds dma_mmap_coherent() to powerpc)
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
CC: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Commit 9cd0314(ACPI / ACPICA: Fix global lock acquisition) was backported
into ACPICA code base, and some divergence was introduced.
This patch fixed it,
- rename acpi_ev_global_lock_pending/acpi_ev_global_lock_pending_lock
to acpi_gbl_global_lock_pending/acpi_gbl_global_lock_pending_lock.
- move the initialization of acpi_gbl_global_lock_pending_lock from
acpi_ut_mutex_initialize to acpi_ev_init_global_lock_handler.
Signed-off-by: Lin Ming <ming.m.lin@intel.com>
Reviewed-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
When we set up the flow informations in ip_route_newports(), we take
the address informations from the the rt_key_src and rt_key_dst fields
of the rtable. They appear to be empty. So take the address
informations from rt_src and rt_dst instead. This issue was introduced
by commit 5e2b61f784 ("ipv4: Remove
flowi from struct rtable.")
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Note that this relies on stuff currently in mfd's next tree, but this
is also a newer driver. I'm not sure which tree it should go through,
as it's a problem that shows up in next.
From: Andres Salomon <dilinger@queued.net>
MFD changes in c738892f cause the mc13xxx_platform_data struct
to change. This changes one more (new) user of it, fixing a build
error.
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
There is a lot of common code that could be shared between the m68k
and m68knommu arch branches. It makes sense to merge the two branches
into a single directory structure so that we can more easily share
that common code.
This is a brute force merge, based on a script from Stephen King
<sfking@fdwdc.com>, which was originally written by Arnd Bergmann
<arnd@arndb.de>.
> The script was inspired by the script Sam Ravnborg used to merge the
> includes from m68knommu. For those files common to both arches but
> differing in content, the m68k version of the file is renamed to
> <file>_mm.<ext> and the m68knommu version of the file is moved into the
> corresponding m68k directory and renamed <file>_no.<ext> and a small
> wrapper file <file>.<ext> is used to select between the two version. Files
> that are common to both but don't differ are removed from the m68knommu
> tree and files and directories that are unique to the m68knommu tree are
> moved to the m68k tree. Finally, the arch/m68knommu tree is removed.
>
> To select between the the versions of the files, the wrapper uses
>
> #ifdef CONFIG_MMU
> #include <file>_mm.<ext>
> #else
> #include <file>_no.<ext>
> #endif
On top of this file merge I have done a simplistic merge of m68k and
m68knommu Kconfig, which primarily attempts to keep existing options and
menus in place. Other than a handful of options being moved it produces
identical .config outputs on m68k and m68knommu targets I tested it on.
With this in place there is now quite a bit of scope for merge cleanups
in future patches.
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
This was noticed by users who performed more than 2^32 lock operations
and hence made this counter overflow (eventually leading to
use-after-free's). Setting rq_client to NULL here means that it won't
later get auth_domain_put() when it should be.
Appears to have been introduced in 2.5.42 by "[PATCH] kNFSd: Move auth
domain lookup into svcauth" which moved most of the rq_client handling
to common svcauth code, but left behind this one line.
Cc: Neil Brown <neilb@suse.de>
Cc: stable@kernel.org
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
fs: simplify iget & friends
fs: pull inode->i_lock up out of writeback_single_inode
fs: rename inode_lock to inode_hash_lock
fs: move i_wb_list out from under inode_lock
fs: move i_sb_list out from under inode_lock
fs: remove inode_lock from iput_final and prune_icache
fs: Lock the inode LRU list separately
fs: factor inode disposal
fs: protect inode->i_state with inode->i_lock
autofs4: Do not potentially dereference NULL pointer returned by fget() in autofs_dev_ioctl_setpipefd()
autofs4 - remove autofs4_lock
autofs4 - fix d_manage() return on rcu-walk
autofs4 - fix autofs4_expire_indirect() traversal
autofs4 - fix dentry leak in autofs4_expire_direct()
autofs4 - reinstate last used update on access
vfs - check non-mountpoint dentry might block in __follow_mount_rcu()
Fix the following compile failure:
drivers/media/rc/ite-cir.c: In function 'ite_decode_bytes':
drivers/media/rc/ite-cir.c:190: error: implicit declaration of function 'generic_find_next_le_bit'
drivers/media/rc/ite-cir.c:199: error: implicit declaration of function 'generic_find_next_zero_le_bit'
Caused by commit 620a32bba4 ("[media] rc: New rc-based ite-cir driver
for several ITE CIRs") interacting with commit c4945b9ed4
("asm-generic: rename generic little-endian bitops functions").
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
