Compared with other drivers, the "ret" should be nagative and
returned. But in vhci_hdc, it always return 0;
I dont't use the driver, and I'm not familiar with the code.
Hope the patch is helpful.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This includes fixes for all of the legit checkpatch.pl errors and
warnings. I have also included several of the suggestions from the
linux-kernel mailing list when the USB-IP code was first added.
Signed-off-by: Brian G. Merrell <bgmerrell@novell.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drag mxc_ipuv3_fb with it (remove final efikamx specific stuff)
update some IPU stuff and ioctls for it all to interoperate
notes:
that line_length bug is still there, I really can't find why it's 0...
doesn't register because the platform_device isn't in a/a/p-mxc/devices.?
ENGR00133954 ipuv3: remove FB_SYNC_EXT flag
ENGR00133034 ipuv3: fix black flash issue during high resolution video playback
ENGR00134076 ipuv3: should not set burst size again
ENGR00133345 IPUv3 VDI:Support bottom field first setting
ENGR00131660 ipuv3: left stripe output column should be set to zero
This actually fixes the clock sync problem with the AOC HDMI monitor we have (perfectly
working display but "input not support" floating around). Happily, the boot splash
actually works, too (previously it would not work until the second set of the display
when GDM booted). Hopefully IPU support is nice and stable now.
includes patches:
ENGR00132153 sgtl5000: do not enforce symmetry if rate is unknown
ENGR00132525 sgtl5000: audio clock gating
ENGR00132580 allow prediv by 1 for ext1 clk
(audio clock gating on Efika will be patched in later)
Workaround for imx5x M4IF burst len errata when armv7 neon enabled, use
pgprot_writecombine instead of pgprot_noncached, to avoid burst len > 8.
Signed-off-by: Sammy He <r62914@freescale.com>
HW flow control should be enable/disable by userspace.
It can be set by tty option: options.c_cflag of CRTSCTS
bit.
This bug is found by testing UART BT AR3001 chip,
this chip will wake up the chip by set RTS to high,
it needs frist disable HW flow control, set RTS,
then disable HW flow control.
Signed-off-by: Zhang Jiejing <jiejing.zhang@freescale.com>
ENGR00133429 MX51 DVFS_PER: Set LP voltage to 1.25 V
MX51 DVFS_PER: Set high setpoint and low setpoint of LP voltage
from 1.2 V to 1.25 V.
Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
If VPU doesn't use IRAM, there will be a kernel dump when
we enter VPU driver probe function.
This issue is caused by allocating IRAM with zero buffer
size.
Signed-off-by: Liu Ying <b17645@freescale.com>
When apply rt patch or turn on hardirq threaded in kernel,
play audio by application with realtime schedule policy hangs the system.
It is caused by:
Requesting SDMA channel, a specific channel is used to load script
and context from memory.
This load phase also uses sdma to transfer, then poll
a complete flag within a while(1) loop.
When application with realtime schedule policy request sdma channel,
it will preempt threaded irqs and prevent complete flag to be set.
So a deadlock appears.
Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
disable dp-dc in the display flow end irq handler, otherwise,
ipu may get NFB4EOF error for next fb enabling under big resolution(like 1080p).
Signed-off-by: Jason Chen <b02280@freescale.com>
* sanitize modelist used preferred_mode which was __initdata and shouldn't have been
* init_wdog only runs once and accesses __initdata, so it's been made __init
This patch adds support for Spansion S25FL016K and S25FL064K SPI flash.
It has been tested with physical devices. Note that both parts exhibit
a Winbond manufacturer ID so they might also be added to that section.
Signed-off-by: Gernot Hoyler <Gernot.Hoyler@spansion.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
"plat_id" is always non-NULL here. There is a zero element on the end
of the m25p_ids[] array and if we hit the end of the loop then plat_id
points to that.
This would lead to a NULL pointer dereference later on in the function.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Anton Vorontsov <cbouatmailru@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
spi_write_then_read() may return its own return codes (e.g. -EIO),
so let's propagate the value down to the probe().
Also, remove jedec == 0 check, it isn't needed as nowadays we use
dedicated SPI device IDs for non-JEDEC flashes.
Suggested-by: Barry Song <21cnbao@gmail.com>
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>