[PATCH 9/9] drm/verisilicon: Add starfive hdmi driver

[Keith Zhao]

yes I tested 
modetest -M starfive -D 0 -s 116 at 31:1280x720-59.94 -v
modetest -M starfive -D 0 -s 116 at 31:1920x1080 -v

and the second command will repeat the problem
as you advise at the beginning
I call the "starfive_hdmi_setup"  function in the "starfive_hdmi_encoder_enable"
instead of "starfive_hdmi_encoder_mode_set"
resolve the problem
i will add this modify in my next patch

Thank you Hoegeun

On 2023/6/23 10:38, Hoegeun Kwon wrote:
> Hi Keith,
> 
> There is a problem with stopping when changing modes.
> 
> Below test log
> 
> root:~> modetest -Mstarfive -c
> Connectors:
> id      encoder status          name            size (mm)       modes
> encoders
> 116     115     connected       HDMI-A-1        320x180         51      115
>   modes:
>         index name refresh (Hz) hdisp hss hse htot vdisp vss vse vtot
>   #0 1280x800 59.91 1280 1328 1360 1440 800 803 809 823 71000 flags: phsync,
> pvsync; type: preferred, driver
>   #1 1920x1080 60.00 1920 2008 2052 2200 1080 1084 1089 1125 148500 flags:
> phsync, pvsync; type: driver
> [...]
> 
> root:~> modetest -Mstarfive -s 116:#0 -v
> setting mode 1280x800-59.91Hz on connectors 116, crtc 31
> freq: 60.65Hz
> freq: 59.91Hz
> freq: 59.91Hz
> 
> root:~> modetest -Mstarfive -s 116:#1 -v
> setting mode 1920x1080-60.00Hz on connectors 116, crtc 31
> [   94.535626] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> [   94.560985] rcu:     1-...0: (20 ticks this GP)
> idle=c9bc/1/0x4000000000000000 softirq=3869/3871 fqs=1120
> [   94.589532] rcu:     (detected by 3, t=5264 jiffies, g=4645, q=63
> ncpus=4)
> [   94.615335] Task dump for CPU 1:
> [   94.637723] task:modetest        state:R  running task     stack:0
> pid:407   ppid:397    flags:0x00000008
> [   94.667299] Call Trace:
> [   94.689297] [<ffffffff80d1e8fc>] __schedule+0x2a8/0xa52
> [   94.714221] [<ffffffff80d1f100>] schedule+0x5a/0xdc
> [   94.738626] [<ffffffff80d25a14>] schedule_timeout+0x220/0x2a6
> [   94.763762] [<ffffffff80d2037a>] wait_for_completion+0xfe/0x126
> [   94.789073] [<ffffffff8002ffe4>] kthread_flush_worker+0x82/0xa0
> 
> 
>> -----Original Message-----
>> From: dri-devel <dri-devel-bounces at lists.freedesktop.org> On Behalf Of
>> Keith Zhao
>> Sent: Friday, June 2, 2023 4:41 PM
>> To: dri-devel at lists.freedesktop.org; devicetree at vger.kernel.org; linux-
>> kernel at vger.kernel.org; linux-riscv at lists.infradead.org; linux-
>> media at vger.kernel.org; linaro-mm-sig at lists.linaro.org
>> Cc: Krzysztof Kozlowski <krzysztof.kozlowski+dt at linaro.org>; Sumit Semwal
>> <sumit.semwal at linaro.org>; Emil Renner Berthing <kernel at esmil.dk>;
>> Shengyang Chen <shengyang.chen at starfivetech.com>; Conor Dooley
>> <conor+dt at kernel.org>; Albert Ou <aou at eecs.berkeley.edu>; Thomas
>> Zimmermann <tzimmermann at suse.de>; Jagan Teki <jagan at edgeble.ai>; Rob
>> Herring <robh+dt at kernel.org>; Chris Morgan <macromorgan at hotmail.com>; Paul
>> Walmsley <paul.walmsley at sifive.com>; Keith Zhao
>> <keith.zhao at starfivetech.com>; Bjorn Andersson <andersson at kernel.org>;
>> Changhuang Liang <changhuang.liang at starfivetech.com>; Jack Zhu
>> <jack.zhu at starfivetech.com>; Palmer Dabbelt <palmer at dabbelt.com>; Shawn
>> Guo <shawnguo at kernel.org>; christian.koenig at amd.com
>> Subject: [PATCH 9/9] drm/verisilicon: Add starfive hdmi driver
>> 
>> Add HDMI dirver for StarFive SoC JH7110.
>> 
>> Signed-off-by: Keith Zhao <keith.zhao at starfivetech.com>
>> ---
>>  drivers/gpu/drm/verisilicon/Kconfig         |  11 +
>>  drivers/gpu/drm/verisilicon/Makefile        |   1 +
>>  drivers/gpu/drm/verisilicon/starfive_hdmi.c | 928 ++++++++++++++++++++
>> drivers/gpu/drm/verisilicon/starfive_hdmi.h | 296 +++++++
>>  drivers/gpu/drm/verisilicon/vs_drv.c        |   6 +
>>  drivers/gpu/drm/verisilicon/vs_drv.h        |   4 +
>>  6 files changed, 1246 insertions(+)
>>  create mode 100644 drivers/gpu/drm/verisilicon/starfive_hdmi.c
>>  create mode 100644 drivers/gpu/drm/verisilicon/starfive_hdmi.h
> 
> [...]
> 
>> diff --git a/drivers/gpu/drm/verisilicon/starfive_hdmi.c
>> b/drivers/gpu/drm/verisilicon/starfive_hdmi.c
>> new file mode 100644
>> index 000000000000..128ecca03309
>> --- /dev/null
>> +++ b/drivers/gpu/drm/verisilicon/starfive_hdmi.c
>> @@ -0,0 +1,928 @@
> 
> [...]
> 
>> +static int starfive_hdmi_setup(struct starfive_hdmi *hdmi,
>> +			       struct drm_display_mode *mode) {
> 
> [...]
> 
>> +	return 0;
>> +}
>> +
>> +static void starfive_hdmi_encoder_mode_set(struct drm_encoder *encoder,
>> +					   struct drm_display_mode *mode,
>> +					   struct drm_display_mode
> *adj_mode) {
>> +	struct starfive_hdmi *hdmi = encoder_to_hdmi(encoder);
>> +
>> +	starfive_hdmi_setup(hdmi, adj_mode);
> 
> When starfive_hdmi_setup runs here,
> when changing the mode, a problem occurs because try to write a value to reg
> in a state that is not resumed after suspend.
> 
>> +
>> +	memcpy(&hdmi->previous_mode, adj_mode, sizeof(hdmi-
>> >previous_mode)); }
>> +
>> +static void starfive_hdmi_encoder_enable(struct drm_encoder *encoder) {
>> +	struct starfive_hdmi *hdmi = encoder_to_hdmi(encoder);
>> +
>> +	pm_runtime_get_sync(hdmi->dev);
> 
> So if move the call point of starfive_hdmi_setup here, it works normally.
> 
>> +}
> 
> Best regards,
> Hoegeun
> 
>