- 20 Jun, 2022 40 commits
-
-
kyrie wu authored
Add mediatek,mt8186-jpgenc compatible to binding document. MT8186 iommu support 0~16GB iova. We separate it to four banks: 0~4G; 4G~8G; 8G~12G; 12G~16G. The "dma-ranges" could be used to adjust the bank we locate. If we don't set this property. The default range always is 0~4G. Signed-off-by:
kyrie wu <kyrie.wu@mediatek.com> Reviewed-by:
Rob Herring <robh@kernel.org> Signed-off-by:
Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org>
-
keliu authored
Use ida_alloc() and ida_free() instead of the deprecated ida_simple_get() and ida_simple_remove(). Signed-off-by:
keliu <liuke94@huawei.com> Signed-off-by:
Sean Young <sean@mess.org> Signed-off-by:
-
Oliver Neukum authored
While the completion handler is running, usb_unlink_urb() on yourself is a NOP. Remove it. Signed-off-by:
Oliver Neukum <oneukum@suse.com> Signed-off-by:
-
Oliver Neukum authored
During resume we must assume tat devices are not ready for block IO. Use GFP_NOIO. Signed-off-by:
-
Oliver Neukum authored
It cannot fail. Make it void. Signed-off-by:
-
Oliver Neukum authored
During resume() GFP_NOIO is enough. Signed-off-by:
-
Oliver Neukum authored
During resume() GFP_NOIO will do. No need for GFP_ATOMIC Signed-off-by:
-
Oliver Neukum authored
No buffer can be embedded inside a descriptor, not even a simple be64. Use a separate kmalloc() Signed-off-by:
-
Oliver Neukum authored
Unlinking yourself while the completion handler is running is a NOP. Drop it. Signed-off-by:
-
Oliver Neukum authored
Calling that on yourself while the completion handler is running is a NOP. Remove it. Signed-off-by:
-
Oliver Neukum authored
The driver uses a timer, that may submit the URB and the URB may start the timer. No simple order of killing can break te cycle. Poison the URB before killing the timer. Signed-off-by:
-
Oliver Neukum authored
The timer uses the URB. Free it only after the timer has been stopped. Signed-off-by:
-
Oliver Neukum authored
The coherency rules mean that you cannot embed a buffer inside a descriptor. kmalloc() separately. Signed-off-by:
-
Yunfei Dong authored
In order to reduce decoder latency, enable H264 inner racing mode. Send lat trans buffer information to core when trigger lat to work, need not to wait until lat decode done. Signed-off-by:
Yunfei Dong <yunfei.dong@mediatek.com> Signed-off-by:
-
Yunfei Dong authored
When SCP timeout during playing video, kernel crashes with following message. It's caused by accessing NULL pointer in vpu_dec_ipi_handler. This patch doesn't solve the root cause of NULL pointer, but merely prevent kernel crashed when encounter the NULL pointer. After applied this patch, kernel keeps alive, only the video player turns to green screen. [67242.065474] pc : vpu_dec_ipi_handler+0xa0/0xb20 [mtk_vcodec_dec] [67242.065485] [MTK_V4L2] level=0 fops_vcodec_open(),334: 18000000.vcodec_dec decoder [135] [67242.065523] lr : scp_ipi_handler+0x11c/0x244 [mtk_scp] [67242.065540] sp : ffffffbb4207fb10 [67242.065557] x29: ffffffbb4207fb30 x28: ffffffd00a1d5000 [67242.065592] x27: 1ffffffa0143aa24 x26: 0000000000000000 [67242.065625] x25: dfffffd000000000 x24: ffffffd0168bfdb0 [67242.065659] x23: 1ffffff76840ff74 x22: ffffffbb41fa8a88 [67242.065692] x21: ffffffbb4207fb9c x20: ffffffbb4207fba0 [67242.065725] x19: ffffffbb4207fb98 x18: 0000000000000000 [67242.065758] x17: 0000000000000000 x16: ffffffd042022094 [67242.065791] x15: 1ffffff77ed4b71a x14: 1ffffff77ed4b719 [67242.065824] x13: 0000000000000000 x12: 0000000000000000 [67242.065857] x11: 0000000000000000 x10: dfffffd000000001 [67242.065890] x9 : 0000000000000000 x8 : 0000000000000002 [67242.065923] x7 : 0000000000000000 x6 : 000000000000003f [67242.065956] x5 : 0000000000000040 x4 : ffffffffffffffe0 [67242.065989] x3 : ffffffd043b841b8 x2 : 0000000000000000 [67242.066021] x1 : 0000000000000010 x0 : 0000000000000010 [67242.066055] Call trace: [67242.066092] vpu_dec_ipi_handler+0xa0/0xb20 [mtk_vcodec_dec 12220d230d83a7426fc38c56b3e7bc6066955bae] [67242.066119] scp_ipi_handler+0x11c/0x244 [mtk_scp 8fb69c2ef141dd3192518b952b65aba35627b8bf] [67242.066145] mt8192_scp_irq_handler+0x70/0x128 [mtk_scp 8fb69c2ef141dd3192518b952b65aba35627b8bf] [67242.066172] scp_irq_handler+0xa0/0x114 [mtk_scp 8fb69c2ef141dd3192518b952b65aba35627b8bf] [67242.066200] irq_thread_fn+0x84/0xf8 [67242.066220] irq_thread+0x170/0x1ec [67242.066242] kthread+0x2f8/0x3b8 [67242.066264] ret_from_fork+0x10/0x30 [67242.066292] Code: 38f96908 35003628 91004340 d343fc08 (38f96908) Signed-off-by:
Tinghan Shen <tinghan.shen@mediatek.com> Signed-off-by:
Macpaul Lin <macpaul.lin@mediatek.com> Signed-off-by:
-
Yunfei Dong authored
Vp8 need to use MM21, but vp9 and h264 need to use HyFbc mode for mt8195. Vp8/vp9/h264 use the same MM21 format for mt8192. Signed-off-by:
-
Yunfei Dong authored
Adds compatible for mt8195 platform. Signed-off-by:
-
Yunfei Dong authored
Adds decoder dt-bindings for mt8195. Signed-off-by:
-
Yunfei Dong authored
Add lat soc compatible and to support lat soc power/clk helper. Signed-off-by:
-
Yunfei Dong authored
Adds decoder dt-bindings for compatible "mediatek,mtk-vcodec-lat-soc". Signed-off-by:
-
Yunfei Dong authored
Add vp9 single core decode driver to support mt8186. [hverkuil: fix 3 small checkpatch warnings] Signed-off-by:
Mingjia Zhang <mingjia.zhang@mediatek.com> Signed-off-by:
Xiaoyong Lu <xiaoyong.lu@mediatek.com> Signed-off-by:
-
Hans Verkuil authored
Document this new vivid test control. Signed-off-by:
-
Benjamin Gaignard authored
P010 is a YUV format with 10-bits per component with interleaved UV. Signed-off-by:
Benjamin Gaignard <benjamin.gaignard@collabora.com> Acked-by:
Nicolas Dufresne <nicolas.dufresne@collabora.com> Signed-off-by:
-
Ming Qian authored
Implement V4L2_CID_JPEG_COMPRESSION_QUALITY to set jpeg quality Signed-off-by:
Ming Qian <ming.qian@nxp.com> Signed-off-by:
-
Guo Zhengkui authored
Fix the following coccicheck warning: drivers/media/platform/samsung/s5p-jpeg/jpeg-core.c:1712:24-25: WARNING opportunity for max() max() macro is defined in include/linux/minmax.h. It avoids multiple evaluations of the arguments when non-constant and performs strict type-checking. Signed-off-by:
Guo Zhengkui <guozhengkui@vivo.com> Signed-off-by:
-
Benjamin Marty authored
This bit is marked as reserved in the ADV Hardware Reference Manual. Resetting this bit seems to cause increased video noise. Setting this bit according to the Hardware Reference Manual reduces the video noise immediately. Signed-off-by:
Benjamin Marty <info@benjaminmarty.ch> Reviewed-by:
Dave Stevenson <dave.stevenson@raspberrypi.com> Reviewed-by:
Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Tested-by:
-
Hans Verkuil authored
This is a very old driver for very old hardware and it is one of the very few remaining that does not use the vb2 framework (or even the older videobuf framework), so deprecate this driver and move it to staging with the intent to removing it altogether by the end of 2022. If someone wants to keep this driver, then it has to be converted to use vb2. Reviewed-by:
Ricardo Ribalda <ribalda@chromium.org> Signed-off-by:
-
Colin Ian King authored
The variables f_chk_len and f_chk_addr are being initialized to zero and then being re-assigned in the next statement. The initializations are redundant and can be removed. Also initialize s_chk_addr and s_chk_len at the declaration statement. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Signed-off-by:
-
Ming Qian authored
As there is no interface to power off vpu core. So it need to boot from cold on first load. but on the second load, driver need to restore the status instead of booting. Signed-off-by:
-
Yang Li authored
Eliminate the following coccicheck warning: ./drivers/media/platform/atmel/atmel-isc-base.c:1469:2-3: Unneeded semicolon Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Signed-off-by:
Yang Li <yang.lee@linux.alibaba.com> Signed-off-by:
-
Wan Jiabing authored
clk_prepare_enable() already checks NULL by using !clk. clk_disable_unprepare() already checks NULL by using IS_ERR_OR_NULL. Remove unneeded NULL checks before them. Signed-off-by:
Wan Jiabing <wanjiabing@vivo.com> Signed-off-by:
-
Jorge Maidana authored
Add PAL-Nc support to usbtv, the corresponding register value "0x00fe" comes from the Windows driver "Active Development Co., Ltd. v2.1.1.2". Signed-off-by:
Jorge Maidana <jorgem.linux@gmail.com> Signed-off-by:
-
Jorge Maidana authored
* Fix PAL-Nc horizontal parameters according to DScaler: https://github.com/JohnAdders/DScaler/blob/f7d92b76678e/DScaler/TVFormats.cpp#L88-L94 * Add PAL-Nc TV audio support. * Tested with CVBS and RF on a Pinnacle PCTV HD 800i. Signed-off-by:
-
Hans Verkuil authored
Add a vivid control to insert the HDMI Video Guard Band in the image. Signed-off-by:
-
Hans Verkuil authored
This inserts 4 pixels of the RGB color 0xab55ab at the left hand side of the image. This is only done for 3 or 4 byte RGB pixel formats. The HDMI TMDS encoding of this pixel value equals the Video Guard Band value as defined by HDMI (see section 5.2.2.1 in the HDMI 1.3 Specification) that preceeds the first actual pixel of a video line. If an HDMI receiver doesn't handle this correctly, then it might keep skipping these Video Guard Band patterns and end up with a shorter video line. So this is a nice pattern to test with. Signed-off-by:
-
Hans Verkuil authored
If there is a hardware problem such as someone pulling the CEC line low continuously, then the POLL message will fail with an error other than OK, NACK, ABORTED or TIMEOUT. Log the tx_status value in that case to help debug this. Signed-off-by:
-
Hans Verkuil authored
When logging the infoframes the adv7604 driver stopped reading infoframes at the first missing infoframe type. Instead it should try all supported infoframe types. Signed-off-by:
-
Vladimir Zapolskiy authored
To simplify the driver's maintenance it makes sense to escape from hardcoded numbers of power domain resources per platform and statical allocation of the resources. For instance on a QCOM SM8450 platform the number of CAMSS power domains shall be bumped up to 6, also notably CAMSS on MSM8916 has only one power domain, however it expects to get 2, and thus it should result in a runtime error on driver probe. The change fixes an issue mentioned above and gives more flexibility to support more platforms in future. Signed-off-by:
Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org> Reviewed-by:
Robert Foss <robert.foss@linaro.org> Signed-off-by:
-
Yang Yingliang authored
of_graph_get_next_endpoint() returns an 'endpoint' node pointer with refcount incremented. The refcount should be decremented before returning from vpif_probe(). Reported-by:
Hulk Robot <hulkci@huawei.com> Signed-off-by:
Yang Yingliang <yangyingliang@huawei.com> Signed-off-by:
-
Zheyu Ma authored
The driver should disable the regulator when failing to probe. Signed-off-by:
Zheyu Ma <zheyuma97@gmail.com> Acked-by:
Tim Harvey <tharvey@gateworks.com> Signed-off-by:
-
