Xianzhong [Thu, 20 Jun 2019 17:21:05 +0000 (01:21 +0800)]
MGS-4975 [#imx-1542] fix incomplete dump for gpu command buffer
gpu dump address is aligned with 4096, but dump size is up to 1024,
that causes command dump incomplete around gpu address.
fix gpu dump size with 4096 upper limit.
Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
ming_qian [Wed, 19 Jun 2019 02:39:56 +0000 (10:39 +0800)]
MLK-22037: VPU Decoder: the resolution of g_fmt may be incorrect
sometime firmware may report wrong sequence info, invalid resolution.
getting and setting seqinfo may in the same time, it's a risk
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Peng Fan [Wed, 19 Jun 2019 10:18:51 +0000 (18:18 +0800)]
MLK-22049 imx8mm: fix audio pll setting
The AUDIO PLL max support 650M, so the original clk settings violate
spec. In order not to impact audio functionality, let's div the
clk by 2.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Tested-by: Viorel Suman <viorel.suman@nxp.com>
Reviewed-by: Viorel Suman <viorel.suman@nxp.com>
Jacky Bai [Tue, 18 Jun 2019 10:51:32 +0000 (18:51 +0800)]
MLK-22040 clk: imx: enable qspi clock if m4 enabled on imx8mm
Keep the qspi clock always on if M4 is enabled.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Shijie Qin [Mon, 17 Jun 2019 12:10:36 +0000 (20:10 +0800)]
MLK-22031 VPU Decoder: sync firmware with cleaned up the RPC API
sync firmware with cleaned up the RPC API
Note: just changed name, not changed struct size
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Reviewed-by: ming_qian <ming.qian@nxp.com>
Thinh Nguyen [Wed, 6 Feb 2019 18:49:26 +0000 (10:49 -0800)]
usb: dwc3: haps: Workaround matching VID PID
i.MX6QP and i.MX7D platform use a PCIe controller with the same VID and
PID as this USB controller. The system may incorrectly match this driver
to that PCIe controller. To workaround this, specifically use class type
USB with PCI device ID to prevent incorrect driver matching.
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
(cherry picked from commit
9c7ebc99b382a158fd2594adbfc204e526b16e08)
ming_qian [Mon, 17 Jun 2019 07:07:49 +0000 (15:07 +0800)]
VPU Decoder: synchronize certain commands and events
Firmware requires that certain commands and events not be interrupted,
but firmware didn't implement it,
so it needs driver to ensure that they are sequenced correctly for the FW.
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Sandor Yu [Mon, 10 Jun 2019 07:14:21 +0000 (15:14 +0800)]
MLK-21981: DRM: iMX DP: Avoid unnecessary phy_init function call
Call DP phy initialize function only when dp link rate is changed.
Signed-off-by: Sandor Yu <Sandor.yu@nxp.com>
Reviewed-by: Robby Cai <robby.cai@nxp.com>
Sandor Yu [Thu, 13 Jun 2019 06:10:07 +0000 (14:10 +0800)]
MLK-21982: drm: hdp: Fix DP/HDMI hotplug failed issue
DP need setup link for every cable plugin.
SCDC should setup again for HDMI2.0 4Kp60 video mode when cable plugin.
Call mode_set function to reconfiguration hdmi/dp controller and PHY.
Because mode_set function will called in drm_fb_helper_hotplug_event function
when drm fbdev enable, the patch only active when drm fbdev disable.
Signed-off-by: Sandor Yu <Sandor.yu@nxp.com>
Reviewed-by: Robby Cai <robby.cai@nxp.com>
Sandor Yu [Thu, 13 Jun 2019 05:28:06 +0000 (13:28 +0800)]
MLK-22007: DRM: imx: use the drm_fb_helper_output_poll_changed function
imx drm driver use drm_fb_cma_fbdev_init() and drm_fb_cma_fbdev_fini().
The pointer to the drm_fbdev_cma structure should remove from
struct imx_drm_device
and use the drm_fb_helper_output_poll_changed function.
relative patch:
commit
b2ad94721047c0f047f36c172d68fed93a8f3fcd
Author: Noralf Trønnes <noralf@tronnes.org>
Date: Wed Nov 15 15:19:46 2017 +0100
drm/imx: Use drm_fb_cma_fbdev_init/fini()
Signed-off-by: Sandor Yu <Sandor.yu@nxp.com>
Reviewed-by: Liu Ying <victor.liu@nxp.com>
Reviewed-by: Laurentiu Palcu <laurentiu.palcu@nxp.com>
ming_qian [Thu, 13 Jun 2019 05:27:51 +0000 (13:27 +0800)]
MLK-22006: VPU Decoder: support to find certain pattern from input data
Sometimes for debug, we want to check the
input data whether include certain pattern,
for example, the sequence header.
Driver implement this feature for debug
Fix some typo.
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Mirela Rabulea [Fri, 7 Jun 2019 12:33:34 +0000 (15:33 +0300)]
MLK-21978: mxc-jpeg: Jpeg encoder produces RGB images with distorted colours
Use APP14 marker instead of APP0 marker, for RGB encoding,
jpeg viewers manage to display the colours this way.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Mirela Rabulea [Fri, 7 Jun 2019 12:04:06 +0000 (15:04 +0300)]
MLK-21977: mxc-jpeg: Add sw workaround for jpeg decoder compID limitation
The hardware only supports component IDs between 0-3 or 1-4.
Patch the unsupported component IDs in the input jpeg stream,
before giving it to the CAST-IP jpeg decoder.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Shijie Qin [Tue, 11 Jun 2019 11:30:56 +0000 (19:30 +0800)]
MA-14966 VPU Decoder: check remaining buffer size before copy
check remaining buffer size before copy input buffer to
ring buffer, avoid to copy length longer than remaining
buffer size then lead to memory crossover
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Reviewed-by: ming_qian <ming.qian@nxp.com>
Shijie Qin [Mon, 10 Jun 2019 06:11:21 +0000 (14:11 +0800)]
MLK-21895 VPU Decoder: add buffer threshold for NON_FRAME_LVL mode
use 'stream_buffer_threshold' for NON_FRAME_LVL mode--raw data
stream_buffer_threshold default size is 64K
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Reviewed-by: ming_qian <ming.qian@nxp.com>
Shijie Qin [Wed, 12 Jun 2019 07:35:32 +0000 (15:35 +0800)]
MLK-22002-2 VPU Decoder: cancel return POLLERR at the beginning
The check about if src/dst streaming and empty is superfluous,
henece remove it
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Reviewed-by: ming_qian <ming.qian@nxp.com>
ming_qian [Tue, 11 Jun 2019 02:39:26 +0000 (10:39 +0800)]
MLK-21989: VPU Decoder: avoid crash after vpu decoder hang
If vpu is hang, driver won't really release the instance when v4l2_release,
but driver will release queue and buffer,
the sysfs device instance*_buffer isn't removed.
when we try to access the instance*_buffer in such case,
kernel crash will occur.
Driver release queue and buffer when we really release the instance,
and the crash can be avoided
Signed-off-by: ming_qian <ming.qian@nxp.com>
Shijie Qin [Wed, 12 Jun 2019 02:37:07 +0000 (10:37 +0800)]
MLK-22002 VPU Decoder: cancel return POLLERR at the beginning
Cancel return POLLERR before recevice 'VID_API_EVENT_SEQ_HDR_FOUND'
at the beginning
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Shengjiu Wang [Mon, 3 Jun 2019 02:30:44 +0000 (10:30 +0800)]
MLK-21940-1: ASoC: fsl_asrc: Update mxc_asrc uapi
In order to support the new ASRC in i.MX815, we update the
user api file mxc_asrc.h.
The reason is that the new ASRC support more sample width, and
support endianness, sign, float format, iec958 format setting,
All these type can be expressed by snd_pcm_format_t type.
So we use the in(out)put_format to instead the in(out)put_word_width.
Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Viorel Suman <viorel.suman@nxp.com>
(cherry picked from commit
b95c32c4d4b51e3e55e1e28269a0e8bce0a04f98)
Haibo Chen [Tue, 4 Jun 2019 12:03:32 +0000 (20:03 +0800)]
MLK-21653-2 arm: dts: add 'auto-cmd23-broken' for eMMC on imx6qpdl/imx6sx/imx7d
Broken the auto cmd23 function for eMMC on imx6qpdl/imx6sx/imx7d, otherwise
the eMMC RPMB write access will return general fail.
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Haibo Chen [Tue, 4 Jun 2019 11:42:29 +0000 (19:42 +0800)]
MLK-21653-1 mmc: sdhci-esdhc-imx: handle 'auto-cmd23-broken' from devicetree
Since L4.15, community involve the commit
105819c8a545 ("mmc: core: use mrq->sbc
when sending CMD23 for RPMB"), let the usdhc to decide whether to use ACMD23 for
RPMB. This CMD23 for RPMB need to set the bit 31 to its argument, if not, the
RPMB write operation will return general fail.
According to the sdhci logic, SDMA mode will disable the ACMD23, and only in
ADMA mode, it will chose to use ACMD23 if the host support. But according to
debug, and confirm with IC, the imx6qpdl/imx6sx/imx6sl/imx7d do not support
the ACMD23 feature completely. These SoCs only use the 16 bit block count of
the register 0x4 (BLOCK_ATT) as the CMD23's argument in ACMD23 mode, which
means it will ignore the upper 16 bit of the CMD23's argument. This will block
the reliable write operation in RPMB, because RPMB reliable write need to set
the bit31 of the CMD23's argument. This is the hardware limitation. Due to
imx6sl use SDMA, so for imx6qpdl/imx6sx/imx7d, it need to broke the ACMD23 for
eMMC, SD card do not has this limitation, because SD card do not support reliable
write.
For imx6ul/imx6ull/imx6sll/imx7ulp/imx8, it support the ACMD23 completely, it
change to use the 0x0 register (DS_ADDR) to put the CMD23's argument in ADMA mode.
This patch handle 'auto-cmd23-broken' from devicetree.
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Arulpandiyan Vadivel [Fri, 7 Jun 2019 08:20:28 +0000 (13:50 +0530)]
MLK-21962: media: v4l: capture: ov5640_mipi_v2: Add device tree support
Add device tree support to ov5640_mipi sensor driver. This change helps to
load driver module on booting when it is compiled as loadable module.
Signed-off-by: Arulpandiyan Vadivel <arulpandiyan_vadivel@mentor.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
Mirela Rabulea [Thu, 6 Jun 2019 11:31:13 +0000 (14:31 +0300)]
MLK-21912: mxc-jpeg: Fix for gst decoder broken, after 1080p fix.
Fix broken gst-launch & decoder plugin testcase, because adjusted
width & height were 0, because capture queue settings were improperly set.
Fixes:
776f969c9fe4 ("MLK-21912: mxc-jpeg: Fix for 1080p NV12 broken")
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Arulpandiyan Vadivel [Thu, 6 Jun 2019 08:54:13 +0000 (14:24 +0530)]
MLK-21949: Revert "MLK-14292: 4.9 rebase: LCDIF: 6SX SDB does not boot with lcdif dts"
This chage blocks the suspend/resume operations.
This reverts commit
f07f7a87c3277a614997f104935b70b99f130a9d.
Signed-off-by: Arulpandiyan Vadivel <arulpandiyan_vadivel@mentor.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
Iuliana Prodan [Tue, 28 May 2019 09:52:11 +0000 (12:52 +0300)]
crypto: caam - strip input without changing crypto request
For rsa and pkcs1pad, CAAM expects an input of modulus size.
For this we strip the leading zeros in case the size is more than modulus.
This commit avoids modifying the crypto request while stripping zeros from
input, to comply with the crypto API requirement. This is done by adding
a fixup input pointer and length.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Horia Geanta <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
(cherry picked from commit
3b2614cb864d4114250f043ce6ff94e152198ae5)
Reviewed-by: Silvano Di Ninno <silvano.dininno@nxp.com>
Mirela Rabulea [Wed, 5 Jun 2019 07:33:18 +0000 (10:33 +0300)]
MLK-21929: mxc-jpeg: Set colorspace in jpeg try_fmt ioctl
Respond with colorspace REC709 for gst negotiations, to be in sync
with the camera, otherwise, the default would be REC601.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Suggested-by: Bing Song <bing.song@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Mirela Rabulea [Fri, 31 May 2019 11:02:01 +0000 (14:02 +0300)]
MLK-21912: mxc-jpeg: Fix for 1080p NV12 broken, after GRAY fix.
Up/down alignment of w x h needs to happen before the
configuration phase for encoder, so do it once in s_fmt.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Mirela Rabulea [Thu, 30 May 2019 17:03:11 +0000 (20:03 +0300)]
MLK-21912: mxc-jpeg: Add encoder support for GRAY format
Break the encoder configuration stream (hactbl) into marker segments,
because the SOF & SOS segments need to have variable length, depending
on the number of colour components. This will allow ARGB (4 components)
and GRAY (1 component) to work.
So far, the configuration stream was hardcoded for 3 components.
Add new functions mxc_jpeg_fixup_sof & mxc_jpeg_fixup_sos, both will start
form a maximal SOF/SOS segment and fix it to match the current format.
Add new function mxc_jpeg_setup_cfg_stream, to copy all the default marker
segments into the configuration stream, and patch it to match the current
format.
Move the setup/fixup of encoder configuration stream from configuration
phase(mxc_jpeg_config_enc_desc) to mxc_jpeg_s_fmt (when format changes).
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Xianzhong [Wed, 5 Jun 2019 15:26:37 +0000 (23:26 +0800)]
MGS-4763 [#imx-320] fix global state error in GPU kernel
broadcast power off will not acquire global semaphore,
gpu suspend will pull up power to avoid command stuck,
then global semphore can be acquired in global power on.
remove global state error and add error check in place.
Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
ming_qian [Wed, 5 Jun 2019 07:53:32 +0000 (15:53 +0800)]
MLK-21951: VPU Decoder: fix stop cmd may be missed
If user open and close in a short time,
call close before driver receive the start done event,
driver won't send stop cmd to firmware, and then release the context.
next time user open device, the state of firmware is mismatch with driver,
it will cause some unexpected error
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Zhou Peng <eagle.zhou@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Vipul Kumar [Tue, 4 Jun 2019 06:21:44 +0000 (11:51 +0530)]
MLK-21875 ARM: imx: Remove cpu check before ethernet mac initialization
fsl,imx6ul-ocotp is used for for imx6ul, imx6ull and imx6ulz boards. So,
remove cpu check before calling imx6_enet_mac_init() function.
This patch fix the following error message at boot up:
[ 0.237021] failed to find ocotp node
[ 0.237389] failed to find ocotp node
Signed-off-by: Vipul Kumar <vipul_kumar@mentor.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
Fancy Fang [Mon, 3 Jun 2019 02:45:59 +0000 (10:45 +0800)]
MLK-21911 drm/bridge: sec-dsim: free 'pll_pms' to avoid memory leak
The 'pll_pms' data is allocated dynamically and it should be
freed after its usage is done explicitly, otherwise there is
a serious memory leak problem, since this data allocation
frequency is really high during video playback and graphic
cases running.
Signed-off-by: Fancy Fang <chen.fang@nxp.com>
Signed-off-by: Jason Liu <jason.hui.liu@nxp.com>
(cherry picked from commit
f28801c464ac6d43893584af85bc348e66149e9d)
ming_qian [Tue, 4 Jun 2019 02:54:14 +0000 (10:54 +0800)]
MLK-21931: VPU Decoder: avoid abort during source change
when vpu parse the first seq header,
driver will send source change event to user.
when vpu send res_change event,
driver will send source change event to user.
So when driver send source change event to user,
user may execute streamoff, and then streamon,
or only streamon if the streamon isn't executed before.
driver should avoid abort cmd in streamoff in this case.
In normal case, user shouldn't streamon capture stream
before it receive the first source change event.
Signed-off-by: ming_qian <ming.qian@nxp.com>
Yong Gan [Thu, 15 Nov 2018 14:53:45 +0000 (22:53 +0800)]
MGS-4381 [#ccc] Refined the licence for the file viv-metadata.h
Change the licence to GPL.
Signed-off-by: Yong Gan <yong.gan@nxp.com>
(cherry picked from commit
c339c74ea52406f966507b89aa076ae0cdfd0ab6)
Ranjani Vaidyanathan [Wed, 29 May 2019 16:10:14 +0000 (11:10 -0500)]
MLK-21896 arm64: dts: imx8qm/imx8qxp Fix PWM clocks and power domain
Add references to the power-domains in the PWM dts nodes.
Fix the ipg clock in PWM dts nodes.
Add PWM nodes to imx8qxp dts file
Signed-off-by: Ranjani Vaidyanathan <ranjani.vaidyanathan@nxp.com>
Acked-by: User <leonard.Crestez.nxp.com>.
Tested-by: User <Dragos.Mincu.nxp.com>.
Robby Cai [Mon, 27 May 2019 07:47:15 +0000 (15:47 +0800)]
MLK-21865 Revert "regmap: use debugfs even when no device"
This reverts commit
9b947a13e7f6017f18470f665992dbae267852b3.
The commit
9b947a13e7f6017f18470f665992dbae267852b3 does not consider power and
clock when expose the regmap into debugsfs. This may bring some problem for some
case. E.g., For reset registers (not real GPR, but csi bridge register) for
MIPI DSI, MIPI CSI and LCDIF on i.mx8mm, they only can be accessed after the
disp_root clock and dispmix power has been turned on, otherwise the system may
be stuck.
on i.mx8mm evk, the command to reproduce (with default dtb, no display/capture
running):
cat /sys/kernel/debug/regmap/dummy-display-gpr@
32e28000/registers
This patch reverted previous commit to avoid the exposure.
Signed-off-by: Robby Cai <robby.cai@nxp.com>
Reviewed-by: Fancy Fang <chen.fang@nxp.com>
Abel Vesa [Mon, 13 May 2019 16:22:24 +0000 (19:22 +0300)]
MLK-21399 irqchip: gic-v3: Rework the ERR11171 workaround
Instead of just raising irq0 for all the cores, we mask the irq0 for all the
non-target cores, this way waking up only the core we want. All of this
is done now in TF-A.
Also, since this new workaround doesn't need the IOMUX_GPR1 register here
in kernel, the IOMUX_GPR reg entry inside the gic dts node can be removed.
In order for this to work, the following commit is needed in TF-A:
0e91ff59720d0756 ("MLK-21399 plat: imx8mq: gpc: Workaround for ERR11171")
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
ming_qian [Fri, 31 May 2019 05:23:28 +0000 (13:23 +0800)]
MMFMWK-8542: VPU Decoder: correct the consumed length if rptr cross over
after abort, if there is no IDR frame, the decoder can't decode any
frame.
If the gop is large, and the stream buffer is full, it will cause hang.
Decoder can skip the undecoded P/B frames and update the rptr.
If the rptr cross over the pre end address,
the consumed length should add stream buffer length.
clear req_frame_count before update uActiveSeqTag
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Zhou Peng <eagle.zhou@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Shijie Qin [Fri, 31 May 2019 05:48:48 +0000 (13:48 +0800)]
MLK-21901: VPU Decoder: add eos padding data for AVS format
the AVS eos padding data is same as ASP
last = 0xb1010000
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Viorel Suman [Thu, 30 May 2019 08:17:53 +0000 (11:17 +0300)]
MLK-18795-2: Revert "MLK-17034-4: ASoC: wm8960: add pm runtime suspend and resume"
Clock management is already implemented in wm8960_set_bias_level_out3,
there is no need to do it in runtime suspend/resume. Another aspect is
that the component_driver.idle_bias_on = 1, because of this codec will
be in runtime_resume all the time and runtime_suspend will never be called,
with current code the implication of this is that the clock will be
enabled all the time even codec not in use.
This reverts commit
fa77799df0a76a8ed1f07fca73fb19964a3a14c5.
Signed-off-by: Viorel Suman <viorel.suman@nxp.com>
Viorel Suman [Thu, 30 May 2019 08:17:48 +0000 (11:17 +0300)]
MLK-18795-1: Revert "MLK-17089-3: ASoC: wm8960: support suspend & resume for imx8"
Clock management is already implemented in wm8960_set_bias_level_out3,
there is no need to do it in runtime suspend/resume. Another aspect is
that the component_driver.idle_bias_on = 1, because of this codec will
be in runtime_resume all the time and runtime_suspend will never be called,
with current code the implication of this is that the clock will be
enabled all the time even codec not in use.
This reverts commit
d845971fdd5531a1f3c37bbc45560aedd3da0773.
Signed-off-by: Viorel Suman <viorel.suman@nxp.com>
ming_qian [Thu, 30 May 2019 06:29:25 +0000 (14:29 +0800)]
MLK-21888: VPU Decoder: release frame when resolution change
driver may alloc frame after firmware send res_change event,
but before driver receive the event.
the frame buffer status maybe incorrect in this case.
So release all the frames when driver receive the res_change event.
remove some unused code
show more debug info
Signed-off-by: ming_qian <ming.qian@nxp.com>
Reviewed-by: Shijie Qin <shijie.qin@nxp.com>
Reviewed-by: Zhou Peng <eagle.zhou@nxp.com>
Xianzhong [Thu, 23 May 2019 11:43:56 +0000 (19:43 +0800)]
MGS-4524 [#imx-1353] wait the signal before destroy context buffer
glslsanbox player test cause GPU hang with repeat timeout command,
the problem is related with glslsanbox context command buffer destroy,
it will unmap GPU MMU pages immediately, but still being used by GPU.
so that GPU will run into corruption with wrong MMU pages suddenly,
need wait for the signal before destroy context command buffer.
Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
Raul E Rangel [Thu, 2 May 2019 19:07:14 +0000 (13:07 -0600)]
mmc: core: Fix tag set memory leak
The tag set is allocated in mmc_init_queue but never freed. This results
in a memory leak. This change makes sure we free the tag set when the
queue is also freed.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Fixes:
81196976ed94 ("mmc: block: Add blk-mq support")
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
(cherry picked from commit
43d8dabb4074cf7f3b1404bfbaeba5aa6f3e5cfc)
Shijie Qin [Tue, 28 May 2019 11:36:52 +0000 (19:36 +0800)]
MLK-21866 VPU Decoder: add __GFP_NOWARN flag for allor dma buffer
Add __GFP_NOWARN flag for allor dma buffer, avoid kernel panic when
alloc failed as no available space.
MLK-21248-3 VPU Decoder: support adaptive playback in driver
Shall cleanup seqinfo after reset setting.
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Mirela Rabulea [Tue, 28 May 2019 10:40:18 +0000 (13:40 +0300)]
MLK-21882: mxc-jpeg: Allow DMABUF for mxc-jpeg encoder/decoder
Proposed & requested by Bing Song, for performance measurement.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Laurentiu Palcu <laurentiu.palcu@nxp.com>
Silvano di Ninno [Thu, 23 May 2019 07:53:14 +0000 (09:53 +0200)]
MLK-21698: tee:optee: fix shared memory page attribute checks
When allocating pages for share memory with OP-TEE,
the driver checks the page attribute (pte).
The current checks only allow writealloc pages.
i.MX 6SLL sets the page attribute to writeback.
Relax this check to allow writealloc, writeback and writethrough.
Signed-off-by: Silvano di Ninno <silvano.dininno@nxp.com>
Reviewed-by: Franck Lenormand <franck.lenormand@nxp.com>
Reviewed-by: Horia Geanta <horia.geanta@nxp.com>
Acked-by: Leonard Crestez <leonard.crestez@nxp.com>
Li Jun [Mon, 27 May 2019 08:35:50 +0000 (16:35 +0800)]
MLK-21692 usb: chipidea: udc: protect usb interrupt enable
We hit the problem with below sequence:
- ci_udc_vbus_session() update vbus_active flag and ci->driver
is valid,
- before calling the ci_hdrc_gadget_connect(),
usb_gadget_udc_stop() is called by application remove gadget
driver,
- ci_udc_vbus_session() will contine do ci_hdrc_gadget_connect() as
gadget_ready is 1, so udc interrupt is enabled, but ci->driver is
NULL.
- USB connection irq generated but ci->driver is NULL.
As udc irq only should be enabled when gadget driver is binded, so
add spinlock to protect the usb irq enable&disable for vbus session
handling.
Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
(cherry picked from commit
fe39b8fd45898a3efa054238219b03ff3b91cd42)
Mirela Rabulea [Fri, 24 May 2019 16:49:34 +0000 (19:49 +0300)]
MLK-21438: mxc-jpeg: Fix mxc-jpeg enc/dec for NV12 1080p
The CAST IP requires for NV12 that the resolution is aligned to 16, and
1080 is not a multiple of 16. On the other hand, gstreamer does not allow
to adjust the buffer size down, only up.
For decoding, just fool the CAST-IP the resolution is a bit bigger, but
keep the buffer size the same. Increasing the buffer size would result
in a few extra garbage lines, decreasing is not allowed by gst.
For encoding, just fool the CAST-IP to encode a little bit less,
if we would encode some extra lines, it would not be possible to
discard them from the resulting jpeg.
Add alignment restrictions to the format descriptions in mxc_formats[].
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Clark Wang [Mon, 27 May 2019 07:48:52 +0000 (15:48 +0800)]
MLK-21870 spi: lpspi: remove initialization of cs-gpios in slave mode
Common code will do spi_set_cs() if cs-gpios has been initialized
whether controller is in master or slave mode. So remove allocating
memory for cs-gpios operation to avoid cs control in slave sending
sequence.
Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Acked-by: Fugang Duan <fugang.duan@nxp.com>
Shijie Qin [Mon, 27 May 2019 04:03:42 +0000 (12:03 +0800)]
MLK-21859-2 VPU Decoder: modify sequence header for
SPK format
Inser frame header for first frame
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Shijie Qin [Fri, 24 May 2019 11:57:09 +0000 (19:57 +0800)]
MLK-21859 VPU Decoder: VPU Decoder: modify sequence
header for SPK format
the payloadSize of SPK sequence header shall
be 0
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Richard Zhu [Fri, 24 May 2019 07:58:51 +0000 (15:58 +0800)]
MLK-21824 PCI: imx: msi enable bit of rc should be set in resume
The MSI Enable bit controls delivery of MSI interrupts from components
below the Root Port.
This bit would be lost during the suspend, should be re-configured
during resume.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
ming_qian [Fri, 24 May 2019 08:05:58 +0000 (16:05 +0800)]
MLK-21811-2: VPU Decoder: fix typo and add some log
Signed-off-by: ming_qian <ming.qian@nxp.com>
Liu Ying [Thu, 23 May 2019 05:11:33 +0000 (13:11 +0800)]
MLK-21799 drm/imx: dpu: kms: Suppress build warning for variable 'm'
This patch initializes variable 'm' in function
dpu_atomic_assign_plane_source_per_crtc() to
suppress the below build warning:
drivers/gpu/drm/imx/dpu/dpu-kms.c: In function ‘dpu_drm_atomic_check’:
drivers/gpu/drm/imx/dpu/dpu-kms.c:217:18: warning: ‘m’ may be used uninitialized in this function [-Wmaybe-uninitialized]
int i, j, k, l, m;
^
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Richard Zhu [Thu, 23 May 2019 07:43:56 +0000 (15:43 +0800)]
MLK-21812 PCI: imx: remove gpc related codes
iMX8M PCIe doesn't need to touch GPC, since they are contained in the
power domain operations already.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Acked-by: Fugang Duan <fugang.duan@nxp.com>
Shijie Qin [Fri, 24 May 2019 02:58:36 +0000 (10:58 +0800)]
MLK-21817 VPU Decoder: use getTSManagerPreBufferCnt
as a reference count of delay decode frames
Use getTSManagerPreBufferCnt from TSM to delay
frames, insead of the difference between input
buffer times and output.
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Richard Zhu [Wed, 22 May 2019 08:32:21 +0000 (16:32 +0800)]
MLK-21807-3 use the bus dma mask refer to the 32 bit dma limitation
Refer to the 32bits DMA limitation on some iMX8 SOCs, use the
bus_dma_mask to specify the limitation.
Get the dma-ranges from the node at firstly, then the next parent.
Otherwise, the dma-ranges defined in the PCIe node, wouldn't be found.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Wed, 22 May 2019 08:38:53 +0000 (16:38 +0800)]
MLK-21807-2 arm64: dts: imx8: specify the imx pcie inbound memory region
Refer to the 32bits DMA limitation, specify the dma ranges memory region
of iMX8 PCIe in the 32bits address space.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Mon, 20 May 2019 08:22:09 +0000 (16:22 +0800)]
MLK-21807-1 Revert "MLK-15064-2 ARM64: DMA: limit the dma mask to be 32bit"
This reverts commit
1953b8289b51b1ebc42e8511a6b7b12de7142c1e.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Shijie Qin [Thu, 23 May 2019 11:40:53 +0000 (19:40 +0800)]
MLK-21816 VPU Decoder: modify insert_header to fix mosaic for VC1 format
extend 12 byte for load size about VC1 insert_header
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Arulpandiyan Vadivel [Wed, 22 May 2019 13:15:12 +0000 (18:45 +0530)]
MLK-21621 clk: imx6sll / imx6ul: correct gpio flags
Adding CLK_IS_CRITICAL flags to all GPIO banks of imx6sll and imx6ul
boards, As gpio pins might be accessed during the suspend mode.
Signed-off-by: Arulpandiyan Vadivel <arulpandiyan_vadivel@mentor.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
Mirela Rabulea [Wed, 22 May 2019 13:50:41 +0000 (16:50 +0300)]
MLK-21808: mxc-jpeg: Take resolution from jpeg when the user provides 0x0
The jpegparse gst plugin sets the resolution to 0x0 for progressive and
extended sequential jpegs, so use the resolution from the jpeg file
instead of returning error. This will allow extended sequential jpegs
to be decoded. Progressive jpegs are not supported.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
Mirela Rabulea [Wed, 22 May 2019 08:58:28 +0000 (11:58 +0300)]
MLK-21804: mxc-jpeg: Report error for progressive jpegs
The mxc-jpeg decoder only supports baseline and extended
sequential jpegs. Progressive jpegs are not supported,
this is confirmed with Alma Technologies.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>
ming_qian [Thu, 23 May 2019 09:48:35 +0000 (17:48 +0800)]
MA-14498: respond eos immediately if firmware is not started
[Android_8QM/8QXP_MEK]RTSP: Play a RTSP streaming and turn off/on
screen, display of streaming will become black.(100%)
Signed-off-by: ming_qian <ming.qian@nxp.com>
Anson Huang [Thu, 23 May 2019 06:37:28 +0000 (14:37 +0800)]
MLK-21596 soc: imx: make sure MU irq can wake up system from freeze mode
Commit
b24e5c5fca92 ("MLK-21078-3 soc: imx: enable RX interrupt
for IPC response") adds IPC RX IRQ support and need to add
IRQF_NO_SUSPEND flag for MU IRQ to make IPC work during system
suspend phase, but with this flag set, IRQD_WAKEUP_ARMED flag will
NOT be set during suspend_device_irq() phase, then when MU IRQ
arrives, it will NOT wake up system from s2idle.
To fix this issue, pm_system_wakeup() is called in general MU IRQ
handler to make sure system can be waked up when MU IRQ arrives.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Acked-by: Robin Gong<yibin.gong@nxp.com>
ming_qian [Thu, 23 May 2019 05:49:41 +0000 (13:49 +0800)]
MLK-21811: VPU Decoder: allocate mbi buffer when requested
1. allocate mbi buffer when requested
2. don't block when receive request frame event
Signed-off-by: ming_qian <ming.qian@nxp.com>
Leonard Crestez [Tue, 21 May 2019 12:55:49 +0000 (15:55 +0300)]
MLK-21768-3 clk: imx6sll: Drop unused clks_init_on
All the clks in this array are already marked as critical so this unused
array is not required.
Fixes:
3b78a5b6263c ("MLK-13303-5 ARM: imx: add clock driver for i.mx6sll")
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Leonard Crestez [Tue, 21 May 2019 12:54:42 +0000 (15:54 +0300)]
MLK-21768-2 clk: imx6sll: Port extra clk flags from 4.14
These flags were not included in upstream and were missed during
porting.
Fixes:
3b78a5b6263c ("MLK-13303-5 ARM: imx: add clock driver for i.mx6sll")
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Leonard Crestez [Tue, 21 May 2019 19:30:16 +0000 (22:30 +0300)]
MLK-21768-1 arm: dts: imx6sll: Fix sram nodes
In upstream a single large area was pushed but in internal tree the sram
is partitioned into multiple blocks used for low power idle and busfreq.
Use definition from imx_4.14.y and prevent the generic mmio-sram from
overwriting low power code.
Fixes:
248bf3ba3b70 ("MLK-13344-03 ARM: dts: imx: add busfreq node on imx6sll")
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Leonard Crestez [Tue, 21 May 2019 11:32:41 +0000 (14:32 +0300)]
MLK-21800 arm: dts: imx6sll: Add extra cpufreq clocks
These extra clocks are used by NXP imx6q cpufreq driver.
They were not included when upstreaming imx6sll dtsi and were missed
during porting.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Shijie Qin [Wed, 22 May 2019 06:17:32 +0000 (14:17 +0800)]
MLK-21248-2 VPU Decoder: support adaptive playback in driver
add mutex lock when send abort/stop cmd to firmware,
avoid twice abort cmd send at the same time in streamoff
and reset.
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Shijie Qin [Tue, 21 May 2019 05:42:10 +0000 (13:42 +0800)]
MLK-21794 VPU Decoder: add new v4l2 dec event for report HW decode error
1. add new v4l2 dec event for report HW decode error
#define V4L2_EVENT_DECODE_ERROR (V4L2_EVENT_PRIVATE_START + 1)
2. refine define of V4L2_EVENT_SKIP and IMX_V4L2_DEC_CMD_RESET
#define V4L2_EVENT_SKIP (V4L2_EVENT_PRIVATE_START + 2)
#define IMX_V4L2_DEC_CMD_START (0x09000000)
#define IMX_V4L2_DEC_CMD_RESET (IMX_V4L2_DEC_CMD_START +1)
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Peng Fan [Tue, 21 May 2019 04:40:55 +0000 (12:40 +0800)]
MLK-21795 clk: imx: imx8mm: correct audio_pll2_clk to audio_pll2_out
There is no audio_pll2_clk registered, it should be audio_pll2_out.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Richard Zhu [Tue, 21 May 2019 09:00:11 +0000 (17:00 +0800)]
MLK-21765-3 clk: imx: remove the redundant codes
Remove the M4 enabled check codes, since the uart clocks is not related
to M4 at all after the commit "
3a2c36e6" MLK-21432 clk: imx: only enable
clocks for earlycon/earlyprintk port.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Tue, 21 May 2019 08:06:51 +0000 (16:06 +0800)]
MLK-21765-2 clk: imx7d: enable uart2 clock when m4 is enabled
iMX7D has the similar potential issue, that the UART clock used by M4
maybe turned off by Linux side, after the initialization of the clocks.
Enable the UART2 clock when M4 is enabled to fix it.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Tue, 21 May 2019 07:57:46 +0000 (15:57 +0800)]
MLK-21765-1 arm64: dts: imx8m: define the init-on-array clocks
The UART clock used by M4 maybe turned off by Linux side, after the
initialization of the clocks.
Re-define the init-on-array clocks in the dts, Add the UART clock used
by M4 into the init_on clock array to fix this issue.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Tue, 14 May 2019 09:43:45 +0000 (17:43 +0800)]
MLK-21695-3: PCI: imx: set up the hsio regmap itself
Setup PCI its own HSIO regmap to fix the kernel dump, when the HSIO
regmap is set as system syscon.
/sys/kernel/debug/regmap# cat dummy-hsio@
5f080000/register
NOTE: devm_ioremap is used to get the virtual address, because that the
devm_ioremap_resource would return -EBUSY when there is a resource
overlap between different HSIO consumers.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Tue, 14 May 2019 09:42:30 +0000 (17:42 +0800)]
MLK-21695-2: ata: imx: set up the hsio regmap itself
Setup SATA its own HSIO regmap to fix the kernel dump,
when the HSIO regmap is set as system syscon.
/sys/kernel/debug/regmap# cat dummy-hsio@
5f080000/register
NOTE: devm_ioremap is used to get the virtual address, because that the
devm_ioremap_resource would return -EBUSY when there is a resource
overlap between different HSIO consumers.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Richard Zhu [Thu, 9 May 2019 10:50:18 +0000 (18:50 +0800)]
MLK-21695-1 arm64: dts: add the peripheral clocks for hsio
In order to resolve the HSIO regmap dump issue.
- Add the requrired HSIO peripheral clocks for different consumers.
- Remove the global syscon HSIO. And setup the separated regmap
by each device itself.
- Remove the HSIO node to fix the dtb build errors.
- Correct the HSIO power domain name on iMX8QXP, otherwise, the
IMX8QXP_HSIO_MISC_PER_CLK clock clock wouldn't be enabled.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Signed-off-by: Ranjani Vaidyanathan <ranjani.vaidyanathan@nxp.com>
ming_qian [Tue, 21 May 2019 08:55:26 +0000 (16:55 +0800)]
MLK-21798: VPU Decoder: dump firmware log using debugfs instead of sysfs
The size of sysfs is limited to PAGE_SIZE.
But there is no limit if using debugfs.
Signed-off-by: ming_qian <ming.qian@nxp.com>
Leonard Crestez [Mon, 20 May 2019 13:20:13 +0000 (16:20 +0300)]
MLK-21782 cpufreq: imx6q: Fetch additional clocks not used in upstream
Unlike upstream we use additional clks and we need to fetch them all.
Those additional clocks are present on all platforms, not just 6ul.
This additional manipulation is relied-upon by busfreq.
Fixes:
eac2948ba4c8 ("MLK-11343-02 cpufreq: imx: add more clk used by cpufreq")
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
Leonard Crestez [Mon, 20 May 2019 15:36:40 +0000 (18:36 +0300)]
MLK-21781 cpufreq: imx6q: Assume no speed grading if nothing found
The imx6q-cpufreq driver currently fails to probe on imx6sl because no
speed grading info is found.
For imx6sl there is no reference of speed grading in data sheet or
reference manual, apparently all parts go up to 1ghz. Fix this by using
the entire OPP table if no speed grading information is available.
This is similar to behavior in older releases.
Fixes:
b7655355f521 ("cpufreq: imx6sx/imx6q: read OCOTP through nvmem for imx6sx/imx6q")
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
Haibo Chen [Fri, 17 May 2019 07:27:44 +0000 (15:27 +0800)]
MLK-20420 ARM: dts: imx7ulp-evk: add delay cell for DDR50/DDR52 mode
We find some imx7ulp evk board, SD card work in DDR50 mode will meet
data CRC error. Only some board has this issue. And eMMC DDR50 mode
also has this issue on these boards. For DDR50, do tuning can fix
this issue, but eMMC DDR52 do not support tuning. So this patch
manually add the delay cell on the fixed clock (FBCLK_SEL = 0).
Currently, add 15 delay cell, which can make DDR50/DDR52 works stable
on all imx7ulp evk board.
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Reviewed-by: Dong Aisheng <aisheng.dong@nxp.com>
(cherry picked from commit
ef369313de747251ff11c108e7fd5bf2b92df603)
ming_qian [Mon, 20 May 2019 09:30:33 +0000 (17:30 +0800)]
MLK-21778: VPU Decoder: check vb2 buffer is changed
1. check vb2 buffer is changed
2. avoid duplicate qbuf with same index
3. add some debug log
Signed-off-by: ming_qian <ming.qian@nxp.com>
Richard Zhu [Mon, 13 May 2019 08:11:56 +0000 (16:11 +0800)]
MLK-21694 ata: imx: fix can't find ahb clk issue and enable imx8qm sata
- The ahb clock is not mandatory required by iMX8QM SATA, fetch it only
when there is "ahb" clock.
- Specify the 32-bit dma limitation and the softreset for iMX8QM SATA.
- Use the standard ahci_error_handler on iMX8QM SATA.
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
Mirela Rabulea [Fri, 17 May 2019 15:35:54 +0000 (18:35 +0300)]
MLK-21770: mxc-jpeg: Add packed YUV444 24bpp pixel format for mxc-jpeg
Previously, the mxc-jpeg decoder was using V4L2_PIX_FMT_YUV32(AYUV,
32 bpp), but this format was not exactly suited for what the CAST-IP
supports (1 Plannar in YUVYUV sequence, see Chapter 14.4.4.5.1.13,
STM_CTRL description for JPEG Decoder Wrapper in iMX8DQXP_RM), so decided
to add a new format: V4L2_PIX_FMT_YUV24.
This will also be used by the gst plugins(MMFMWK-8504) and ISI.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewd-by: Laurentiu Palcu <laurentiu.palcu@nxp.com>
Mirela Rabulea [Fri, 17 May 2019 15:27:34 +0000 (18:27 +0300)]
MLK-21770: media: v4l: Add packed YUV444 24bpp pixel format
The added format is V4L2_PIX_FMT_YUV24, this is a packed
YUV 4:4:4 format, with 8 bits for each component, 24 bits
per sample.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Laurentiu Palcu <laurentiu.palcu@nxp.com>
Acked-by: Leonard Crestez <leonard.crestez@nxp.com>
Leonard Crestez [Sun, 19 May 2019 10:28:26 +0000 (13:28 +0300)]
MLK-21750 mfd: bd718x7: Remove hardcoded config for button press duration
The reset button on imx8mm-evk is tied to PWRON_B pin of
bd71847 and the
long press duration is set to zero from OTP. The linux driver overrides
those values and breaks reset from button.
Overwriting OTP or bootloader configuration with some hardcoded defaults
is not desirable, keep already programmed values instead.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Acked-By: Matti Vaittinen <matti.vaittinen@fi.rohmeurope.com>
Fugang Duan [Fri, 17 May 2019 09:18:04 +0000 (17:18 +0800)]
MLK-21756 brcmfmac: bcmsdh: attach device only func1 compatible string match
brcmfamc driver attach mmc func devices only when func1 device's
node exist, and whose compatible string match "brcm,bcm4329-fmac".
Reviewed-by: Richard Zhu <hongxing.zhu@nxp.com>
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Leonard Crestez [Sun, 19 May 2019 13:39:27 +0000 (16:39 +0300)]
MLK-21650 clk: imx6sx: Fix reparenting perclk
Remove the CLK_IS_CRITICAL flag because this breaks the reparenting of
perclk_sel to osc. Replace this with an explicit clk_prepare_enable
instead.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
Peng Fan [Mon, 20 May 2019 06:53:16 +0000 (14:53 +0800)]
MLK-21776 iommu: arm: pgtable: fix pagetable allocation
ZONE_DMA is removed by commit
ad67f5a6545("arm64: replace ZONE_DMA with ZONE_DMA32").
So need use __GFP_DMA32, otherwise meet
"
[ 2.560837] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 2.569456] iommu: Failed to add device
5f020000.sata to group 0: -12
[ 2.623471] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 2.632022] iommu: Failed to add device
5b040000.ethernet to group 0: -12
[ 2.703749] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 2.712294] iommu: Failed to add device
5b050000.ethernet to group 0: -12
[ 3.294966] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 3.303514] iommu: Failed to add device
5b010000.usdhc to group 0: -12
[ 3.358298] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 3.366858] iommu: Failed to add device
5b020000.usdhc to group 0: -12
[ 4.184750] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 4.193311] iommu: Failed to add device
5b050000.ethernet to group 0: -12
[ 4.205606] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 4.214175] iommu: Failed to add device
5b050000.ethernet to group 0: -12
[ 4.224358] arm-smmu
51400000.iommu: Cannot accommodate DMA translation for IOMMU page tables
[ 4.232938] iommu: Failed to add device
5b050000.ethernet to group 0: -12
"
Fixes:
90d512cf0ab ("MLK-15007-1 iommu: arm: pgtable: alloc pagetable in DMA area")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Richard Zhu <hongxing.zhu@nxp.com>
Shijie Qin [Mon, 20 May 2019 04:30:23 +0000 (12:30 +0800)]
MMFMWK-8516 VPU decoder: memory leak issue when playing
RV format video
Memory leak issue when playing RV format video. Hence, will
consume all available cmd size and cause kernel panic if repeatedly
playing.
Free arv_frame memory before return.
Signed-off-by: Shijie Qin <shijie.qin@nxp.com>
Xianzhong [Thu, 16 May 2019 17:03:55 +0000 (01:03 +0800)]
MGS-4716-2 [#imx-1412] set contiguous 4GB flag for CMA limit
CMA limit shall not allocate non-contiguous 4GB memory
Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
Xianzhong [Fri, 17 May 2019 15:23:33 +0000 (23:23 +0800)]
MGS-4861 [#imx-1482] fix AXI bus error for g2d compositor
dpu g2d does not support non-contiguous system memory,
fix video memory block to avoid wrong MMU mapping.
Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
ming_qian [Mon, 20 May 2019 02:16:51 +0000 (10:16 +0800)]
MLK-21774: VPU Decoder: avoid request frame buffer after abort
firmware may restart request frame buffer after
send event STR_BUF_RST to driver.
So frame buffer may be requested before driver clear queue.
It's not safe.
Signed-off-by: ming_qian <ming.qian@nxp.com>
Guoniu.Zhou [Wed, 15 May 2019 06:01:16 +0000 (14:01 +0800)]
MLK-21743: dma: pxp_v3: the crop height need to be even when input format is YUV420/YVU420
YVU420/YUV420 has three components which are separated into three
sub-images or planes and the Cr plane is half the width and half
the height of the Y plane (and of the image). So the crop width
and height need to be even in order to meet the format requirements.
Signed-off-by: Guoniu.Zhou <guoniu.zhou@nxp.com>
Signed-off-by: Richard Liu <xuegang.liu@nxp.com>
Tested-by: Richard Liu <xuegang.liu@nxp.com>
(cherry picked from commit
f2edfadf34c31cfa8b53fa60a88ccc90d4d6667d)
Leonard Crestez [Thu, 16 May 2019 12:16:16 +0000 (15:16 +0300)]
MLK-21554-4 clk: imx: Relax CLK_SET_RATE_GATE for imx6/7 busfreq
This fixes busfreq work on various 6s/6u/7d chips and realigns us with
upstream.
The clk framework in new versions rejects clk_set_parent if any child
has CLK_SET_RATE_GATE set. However doing glitchless muxing upstream
shouldn't cause problems for dividers lower down.
All the CLK_SET_RATE_GATE flag ever did is reject certain invalid
clk_set_rate calls however at this point our imx6/7 driver are
reasonably stable anyway. Perhaps this flag could be maintained in some
form but that should be done directly upstream.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Acked-by: Nitin Garg <nitin.garg@nxp.com>
Anson Huang [Wed, 17 Apr 2019 05:39:28 +0000 (13:39 +0800)]
MLK-21554-3 clk: imx6sx: keep OCRAM_S always ON
OCRAM_S is used as iram tlb table for low power modes, clock
needs to be always ON.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Leonard Crestez <leonard.crestez@nxp.com>
Mirela Rabulea [Wed, 15 May 2019 11:04:52 +0000 (14:04 +0300)]
MLK-21742: mxc-jpeg: Fix extended sequential jpeg decoding
For extended sequential jpegs, the frame definition is
in the SOF1 marker, so use that, also keep using SOF0 for
baseline sequential and SOF1 for progressive jpegs.
Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
Reviewed-by: Robert Chiras <robert.chiras@nxp.com>