Eric W. Biederman says:
====================
Using dev_kfree/consume_skb_any for functions called in multiple contexts
These changes are a result of walking through the network drivers
supporting netpoll and verifying the code paths that netpoll can cause
to be called in hard irq context use an appropriate flavor of
kfree_skb. Either dev_kfree_skb_any or dev_consume_skb_any.
Since my last pass at this I have become aware of the small differences
between dev_kfree_skb_any and dev_consume_skb_any.
net/core/drop_monitor.c reports the dev_kfree_skb_any as a drop and
while being quite about the second. With the weird twist that
dev_kfree_skb is unintuitively consume_skb.
As netpoll now calls the napi poll function with budget == 0, pieces of
a drivers the napi poll function that don't run when budget == 0 have
been ignored.
The most interesting change is to the atl1c which tried unsuccesfully to
tell one of it's functions which context it is called in so that it
could call dev_kfree_skb_irq or dev_kfree_skb as appropriate. I have
just removed the extra parameter and called dev_consume_skb_any.
At 54 separate changes I will post each change as a separate patch (so
they can be reviewed) but for general sanity sake I have gathered them
all into a git branch for easy acces.
====================
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull nfsd fix frm Bruce Fields:
"J R Okajima sent this early and I was just slow to pass it along,
apologies. Fortunately it's a simple fix"
* 'nfsd-next' of git://linux-nfs.org/~bfields/linux:
nfsd: fix lost nfserrno() call in nfsd_setattr()
Pull vfs fixes from Al Viro:
"These four commits are obvious fixes (a couple of fdget_pos()-related
ones from Eric Biggers, prepend_name() fix, missing checks for false
negatives from __lookup_mnt() in fs/namei.c)"
For now I'm pulling just the four obvious fixes, there's another four
pending in Al's 'for-linus' branch wrt the mnt_hash list that were more
involved.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
rcuwalk: recheck mount_lock after mountpoint crossing attempts
make prepend_name() work correctly when called with negative *buflen
vfs: Don't let __fdget_pos() get FMODE_PATH files
vfs: atomic f_pos access in llseek()
When parsing options, make sure we have found a proper token before
doing a numeric conversion.
Without this check, the current code will end up following random
pointers that just happened to be on the stack when this function was
called, because match_token() will not touch the 'args' list unless a
valid token is found.
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
There is no point in allocating these structs separately.
Changing this makes the code a little simpler and saves a few bytes of
memory.
Reported-by: Herve Vico
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
This request state is mostly useless, and properly implementing it
for RDMA would require an extra lock to be taken in handle_recv()
and in rdma_cancel() to avoid this race:
handle_recv() rdma_cancel()
. .
. if req->state == SENT
req->state = RCVD .
. req->state = FLSH
So just get rid of it.
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Take into account posted recv buffers that will never receive their
reply.
The RDMA code posts a recv buffer for each request that it sends.
When a request is flushed, it is possible that this request will
never receive a reply, and that one recv buffer will stay unused on
the recv queue.
It is then possible, if this scenario happens several times, to have the
recv queue full, and have the 9pnet_rmda module unable to send new requests.
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
And move transport-specific code out of net/9p/client.c
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Mark function as static in net/9p/client.c because it is not used
outside this file.
This eliminates the following warning in net/9p/client.c:
net/9p/client.c:207:18: warning: no previous prototype for ‘p9_fcall_alloc’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
We need barriers to guarantee this pattern works as intended:
[w] req->rc, 1 [r] req->status, 1
wmb rmb
[w] req->status, 1 [r] req->rc
Where the wmb ensures that rc gets written before status,
and the rmb ensures that if you observe status == 1, rc is the new value.
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
ASoC: A few more fixes for rcar
Since Linus hasn't done the release yet and Morimoto-san here are some
additional bug fixes for rcar which Morimoto-san sent overnight.
# gpg: Signature made Mon 24 Mar 2014 11:21:18 GMT using RSA key ID 7EA229BD
# gpg: Good signature from "Mark Brown <broonie@sirena.org.uk>"
# gpg: aka "Mark Brown <broonie@debian.org>"
# gpg: aka "Mark Brown <broonie@kernel.org>"
# gpg: aka "Mark Brown <broonie@tardis.ed.ac.uk>"
# gpg: aka "Mark Brown <broonie@linaro.org>"
# gpg: aka "Mark Brown <Mark.Brown@linaro.org>"
ASoC: Updates for v3.15
A few more updates for the merge window:
- Fixes for the simple-card DAI format DT mess.
- A new driver for Cirrus cs42xx8 devices.
- DT support for a couple more devices.
- A revert of a previous buggy fix for soc-pcm, plus a few more fixes
and cleanups.
# gpg: Signature made Sun 23 Mar 2014 16:56:11 GMT using RSA key ID 7EA229BD
# gpg: Good signature from "Mark Brown <broonie@sirena.org.uk>"
# gpg: aka "Mark Brown <broonie@debian.org>"
# gpg: aka "Mark Brown <broonie@kernel.org>"
# gpg: aka "Mark Brown <broonie@tardis.ed.ac.uk>"
# gpg: aka "Mark Brown <broonie@linaro.org>"
# gpg: aka "Mark Brown <Mark.Brown@linaro.org>"
ASoC: Updates for v3.15
This is mostly a few additional fixes from Lars-Peter, a new driver and
cleaning up a git failure with merging the Intel branch (combined with
an xargs failure to pay attention to error codes). The history lists a
bunch of additional commits for the branch but the content of those
commits is actually present already but not recorded in history due to
git failing. Unfortunately xargs is used in the merge script and it
doesn't do a good job of noticing errors from the commands it invokes.
# gpg: Signature made Thu 13 Mar 2014 14:25:44 GMT using RSA key ID 7EA229BD
# gpg: Good signature from "Mark Brown <broonie@sirena.org.uk>"
# gpg: aka "Mark Brown <broonie@debian.org>"
# gpg: aka "Mark Brown <broonie@kernel.org>"
# gpg: aka "Mark Brown <broonie@tardis.ed.ac.uk>"
# gpg: aka "Mark Brown <broonie@linaro.org>"
# gpg: aka "Mark Brown <Mark.Brown@linaro.org>"
ASoC: Updates for v3.15
Quite a busy release for ASoC this time, more on janitorial work than
exciting new features but welcome nontheless:
- Lots of cleanups from Takashi for enumerations; the original API for
these was error prone so he's refactored lots of code to use more
modern APIs which avoid issues.
- Elimination of the ASoC level wrappers for I2C and SPI moving us
closer to converting to regmap completely and avoiding some
randconfig hassle.
- Provide both manually and transparently locked DAPM APIs rather than
a mix of the two fixing some concurrency issues.
- Start converting CODEC drivers to use separate bus interface drivers
rather than having them all in one file helping avoid dependency
issues.
- DPCM support for Intel Haswell and Bay Trail platforms.
- Lots of work on improvements for simple-card, DaVinci and the Renesas
rcar drivers.
- New drivers for Analog Devices ADAU1977, TI PCM512x and parts of the
CSR SiRF SoC.
# gpg: Signature made Wed 12 Mar 2014 23:05:45 GMT using RSA key ID 7EA229BD
# gpg: Good signature from "Mark Brown <broonie@sirena.org.uk>"
# gpg: aka "Mark Brown <broonie@debian.org>"
# gpg: aka "Mark Brown <broonie@kernel.org>"
# gpg: aka "Mark Brown <broonie@tardis.ed.ac.uk>"
# gpg: aka "Mark Brown <broonie@linaro.org>"
# gpg: aka "Mark Brown <Mark.Brown@linaro.org>"
When the host->tuning_count is zero it means that the retuning is
disabled. This is checked on the first run of sdhci_execute_tuning()
by the if statement below:
if (!(host->flags & SDHCI_NEEDS_RETUNING) && host->tuning_count &&
(host->tuning_mode == SDHCI_TUNING_MODE_1)) {
So only when tuning_count is non-zero it will set the host flag
SDHCI_USING_RETUNING_TIMER. The else statement is only for re-programming
the timer, which means that flag must be set. Because that is not checked
the else statement is executed in the first run when tuning_count is zero.
This was seen on a host controller which indicated SDHCI_TUNING_MODE_1 (0)
and tuning_count being zero. Suspect that (one of) these registers is not
properly set.
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Aaron Lu <aaron.lu@intel.com>
Signed-off-by: Chris Ball <chris@printf.net>
When ir_dev_scope_init() is called via a rootfs initcall it
will check for irq_remapping_enabled before it calls
(indirectly) into dmar_acpi_dev_scope_init() which uses the
dmar_tbl pointer without any checks.
The AMD IOMMU driver also sets the irq_remapping_enabled
flag which causes the dmar_acpi_dev_scope_init() function to
be called on systems with AMD IOMMU hardware too, causing a
boot-time kernel crash.
Signed-off-by: Joerg Roedel <joro@8bytes.org>
There is no real reason why we require transfers to have a completion and
the only user of the completion now checks to see if one has been provided
before using it so stop enforcing this. This makes it more convenient for
drivers to chain multiple asynchronous transfers together.
Signed-off-by: Mark Brown <broonie@linaro.org>
Update SATA AHCI support to make use of the new ahci_da850 host
driver (instead of the generic ahci_platform one) and remove
deprecated ahci_platform_data code.
Cc: Sekhar Nori <nsekhar@ti.com>
Cc: Kevin Hilman <khilman@deeprootsystems.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Add the new ahci_da850 host driver.
Platform changes needed to make DaVinci DA850 SATA AHCI support
fully functional are in the separate "ARM: davinci: da850: update
SATA AHCI support" commit.
Please note that this driver doesn't have the superfluous clock
control code as clock is already handled by the generic AHCI
platform library code.
Cc: Sekhar Nori <nsekhar@ti.com>
Cc: Kevin Hilman <khilman@deeprootsystems.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Move AHCI platform library code from ahci_platform.c to
libahci_platform.c and fix dependencies for ahci_st,
ahci_imx and ahci_sunxi drivers.
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Looking at ST SPEAr1340 AHCI code (the only user of the deprecated
pdata->suspend and pdata->resume) it is obvious the we should return
after calling pdata->suspend() only if the function have returned
non-zero return value. The code has been broken since commit 1e70c2
("ata/ahci_platform: Add clock framework support"). Fix it.
Cc: Viresh Kumar <viresh.linux@gmail.com>
Cc: Shiraz Hashim <shiraz.hashim@st.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
If during a SPI transfer with len larger than PCH_MAX_FIFO_DEPTH and the
IRQ handler happens to be called when the transmit FIFO is already empty,
and SPSR_FI_BIT is set consequently, the message
"spi_master spi32766: pch_spi_handler_sub : Transfer is not completed"
is spammed to the systemlog, because tx_index has already increased
further due to the next bytes to be written. This case is uncritical as
new bytes have already been written.
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
On Tegra the convention is to have a single machine driver
that's controlling the whole audio subsystem.
Move the clock handling to the machine driver, to be in line
with the other Tegra drivers and give the machine driver
full control over the single Tegra audio PLL.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Some simple audio cards may have many DAI links.
This patch extends the simple-card driver for handling such cards.
Signed-off-by: Jean-Francois Moine <moinejf@free.fr>
Signed-off-by: Mark Brown <broonie@linaro.org>
Many couples of CPU/CODEC DAI links may be described in the DT
thanks to 'simple-audio-card,dai-link' containers.
Signed-off-by: Jean-Francois Moine <moinejf@free.fr>
Signed-off-by: Mark Brown <broonie@linaro.org>
The DAI link array and the properties (fmt, sysclk slots) are
hard-coded for a single CPU / CODEC link.
This patch dynamically allocates the DAI link array and the
properties with the aim of supporting many DAI links.
Signed-off-by: Jean-Francois Moine <moinejf@free.fr>
Signed-off-by: Mark Brown <broonie@linaro.org>
spi core will use spi->max_speed_hz as transfer speed if the transfer speed was
not set. So we don't need to test t->speed_hz in mcfqspi_transfer_one().
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Wolfram Sang pointed out that "efm32,$device" is non-standard. So use the
common scheme and prefix device with "efm32-". The old compatible string
is left in place until arch/arm/boot/dts/efm32* is fixed.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Mark Brown <broonie@linaro.org>
Cc: stable@vger.kernel.org
3 fixes
- memory leak on certain SIGP conditions
- wrong size for idle bitmap (always too big)
- clear local interrupts on initial CPU reset
1 performance improvement
- improve performance with many guests on certain workloads
This patch add .xlate_tdm_slot_mask support for IMX SSI, and this
will generate the TDM slot TX and RX masks.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This patch add .xlate_tdm_slot_mask support for ESAI, and this will
generate the TDM slot TX and RX masks.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This patch add fsl_asoc_xlate_tdm_slot_mask() support for utils.
For the some spcified DAI driver, this will be used to generate the
TDM slot TX/RX mask. And the TX/RX mask will use a 0 bit for an active
slot as default, and the default active bits are at the LSB of
the masks.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
The 'of_' is not appropriate here for there hasn't any DT parsing.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
This patch removes <mach/hardware.h> dependency. This is performed
by replace hard coded used memory regions and interrupt to getting
these values from resources passed to the driver. For the system-wide
registers we now able to use SYSCON driver.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Mark Brown <broonie@linaro.org>
Before we had card level support for DAPM and controls machine drivers would
register their controls and DAPM elements with the CODEC. This required us to
temporarily unset the name_prefix of a CODEC during the rtd init callback to
avoid the machine level controls getting the CODEC's prefix. Now that all
machine drivers properly register their machine level controls and DAPM elements
with the card rather than with the CODEC we can drop the hack that sets the
CODEC's name_prefix to NULL while calling the DAI link or AUX dev init callback.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@linaro.org>
Empty list of local interrupts when vcpu goes through initial reset
to provide a clean state
Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
When kvm_get_vcpu() returned NULL for the destination CPU in
__sigp_emergency() or __sigp_external_call(), the memory for the
"inti" structure was not released anymore. This patch fixes this
issue by moving the check for !dst_vcpu before the kzalloc() call.
Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
We need BITS_TO_LONGS, not sizeof(long) to calculate
the correct size.
idle_mask is a bitmask, each bit representing the state
of a cpu. The desired outcome is an array of unsigned long
fields that can fit KVM_MAX_VCPUS bits. We should not use
sizeof(long) which returnes the size in bytes, but BITS_TO_LONGS
Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
