From: Liviu Dudau <liviu.dudau@arm.com>
To: Cunyuan Liu <cunyuan.liu@cixtech.com>
Cc: boris.brezillon@collabora.com, steven.price@arm.com,
robh@kernel.org, krzk+dt@kernel.org, conor+dt@kernel.org,
gary.yang@cixtech.com, fugang.duan@cixtech.com,
daniel.almeida@collabora.com, aliceryhl@google.com,
p.zabel@pengutronix.de, dri-devel@lists.freedesktop.org,
devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, rust-for-linux@vger.kernel.org,
cix-kernel-upstream@cixtech.com, robin.wang@cixtech.com
Subject: Re: [PATCH v2 2/3] drm/panthor: Add initial support for CIX Sky1
Date: Tue, 10 Nov 2026 10:03:46 +0000 [thread overview]
Message-ID: <avLsAlL048MtIX4x@e142607> (raw)
In-Reply-To: <20261109120902.125242-3-cunyuan.liu@cixtech.com>
Hi Cunyuan,
On Mon, Nov 09, 2026 at 08:09:00PM +0800, Cunyuan Liu wrote:
> Add Sky1 SoC data for system-level cache settings and to skip devfreq.
>
> The GPU on CIX Sky1 needs an external reset before the driver first
> accesses the hardware. Without this reset, the GPU hardware cannot be
> accessed correctly. Acquire an optional exclusive reset for all platforms
> and apply it after enabling the clocks during initial probe. Platforms
> without a reset keep working through the reset framework NULL handling.
Adding support for an external reset should be separate from the enablement
of Sky1 GPU.
>
> Configure SYSC_ALLOC before L2 power-on. Without these settings, GPU
> rendering results are displayed incorrectly.
Adding support for SYSC_ALLOC should be a separate patch. It can be used
by other SoCs as well, not just Sky1.
>
> Keep the clock settings provided by the PM firmware. This first version
> does not read OPPs or enable devfreq. It does not register a GPU cooling
> device.
This looks like a stop gap measure. What is your plan here? When do you
think you might be able to provide OPP and devfreq support and what is
the hold up?
>
> Tested on Orion O6: offscreen pixel readback, kmscube, glmark2 at 1080p,
> runtime suspend and resume with a retained EGL context, software-triggered
> reset-worker tests, and module unload and reload. Display tests used
> local Komeda support.
>
> Signed-off-by: Cunyuan Liu <cunyuan.liu@cixtech.com>
> ---
> drivers/gpu/drm/panthor/panthor_devfreq.c | 11 ++++++++++-
> drivers/gpu/drm/panthor/panthor_device.c | 13 +++++++++++++
> drivers/gpu/drm/panthor/panthor_device.h | 19 +++++++++++++++++++
> drivers/gpu/drm/panthor/panthor_drv.c | 14 ++++++++++++++
> drivers/gpu/drm/panthor/panthor_gpu.c | 14 ++++++++++++++
> drivers/gpu/drm/panthor/panthor_gpu_regs.h | 2 ++
> 6 files changed, 72 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/panthor/panthor_devfreq.c b/drivers/gpu/drm/panthor/panthor_devfreq.c
> index 2249b41ca4af..1914549f7b1f 100644
> --- a/drivers/gpu/drm/panthor/panthor_devfreq.c
> +++ b/drivers/gpu/drm/panthor/panthor_devfreq.c
> @@ -154,6 +154,12 @@ int panthor_devfreq_init(struct panthor_device *ptdev)
>
> ptdev->devfreq = pdevfreq;
>
> + /* Keep the firmware clock settings until Sky1 devfreq is supported. */
> + if (ptdev->soc_data && ptdev->soc_data->skip_devfreq) {
> + ptdev->fast_rate = clk_get_rate(ptdev->clks.core);
> + return 0;
> + }
This should not be in the driver. If you want to have a fixed frequency just use a fixed
clock provider, or make the clock provider query the firmware for the value.
> +
> /*
> * The power domain associated with the GPU may have already added an
> * OPP table, complete with OPPs, as part of the platform bus
> @@ -323,8 +329,11 @@ unsigned long panthor_devfreq_get_freq(struct panthor_device *ptdev)
> unsigned long freq = 0;
> int ret;
>
> - if (!pdevfreq->devfreq)
> + if (!pdevfreq->devfreq) {
> + if (ptdev->soc_data && ptdev->soc_data->skip_devfreq)
> + return clk_get_rate(ptdev->clks.core);
> return 0;
> + }
>
> ret = pdevfreq->devfreq->profile->get_cur_freq(ptdev->base.dev, &freq);
> if (ret)
> diff --git a/drivers/gpu/drm/panthor/panthor_device.c b/drivers/gpu/drm/panthor/panthor_device.c
> index 9687c59de350..a3a32abec5f7 100644
> --- a/drivers/gpu/drm/panthor/panthor_device.c
> +++ b/drivers/gpu/drm/panthor/panthor_device.c
> @@ -229,6 +229,11 @@ int panthor_device_init(struct panthor_device *ptdev)
> return ret;
> }
>
> + ptdev->reset_control = devm_reset_control_get_optional_exclusive(ptdev->base.dev, NULL);
> + if (IS_ERR(ptdev->reset_control))
> + return dev_err_probe(ptdev->base.dev, PTR_ERR(ptdev->reset_control),
> + "Failed to get GPU reset\n");
> +
> ret = panthor_devfreq_init(ptdev);
> if (ret)
> return ret;
> @@ -511,6 +516,12 @@ int panthor_device_resume(struct device *dev)
> if (ret)
> goto err_disable_stacks_clk;
>
> + if (!panthor_device_is_initialized(ptdev)) {
> + ret = reset_control_reset(ptdev->reset_control);
> + if (ret)
> + goto err_disable_coregroup_clk;
> + }
> +
> panthor_devfreq_resume(ptdev);
>
> if (panthor_device_is_initialized(ptdev) &&
> @@ -553,6 +564,8 @@ int panthor_device_resume(struct device *dev)
>
> err_suspend_devfreq:
> panthor_devfreq_suspend(ptdev);
> +
> +err_disable_coregroup_clk:
> clk_disable_unprepare(ptdev->clks.coregroup);
>
> err_disable_stacks_clk:
> diff --git a/drivers/gpu/drm/panthor/panthor_device.h b/drivers/gpu/drm/panthor/panthor_device.h
> index b55a3f9edd41..afb4167bd364 100644
> --- a/drivers/gpu/drm/panthor/panthor_device.h
> +++ b/drivers/gpu/drm/panthor/panthor_device.h
> @@ -19,6 +19,8 @@
> #include <drm/gpu_scheduler.h>
> #include <drm/panthor_drm.h>
>
> +struct reset_control;
> +
> struct panthor_csf;
> struct panthor_csf_ctx;
> struct panthor_device;
> @@ -34,10 +36,24 @@ struct panthor_pwr;
> struct panthor_vm;
> struct panthor_vm_pool;
>
> +/**
> + * struct panthor_sysc_config - GPU system-cache configuration
> + */
> +struct panthor_sysc_config {
> + /** @alloc: System-cache allocation policies for GPU access types. */
> + u32 alloc[8];
> +};
> +
> /**
> * struct panthor_soc_data - Panthor SoC Data
> */
> struct panthor_soc_data {
> + /** @skip_devfreq: Keep the firmware clock settings without devfreq. */
> + bool skip_devfreq;
> +
> + /** @sysc_config: System-cache configuration, or NULL to keep defaults. */
> + const struct panthor_sysc_config *sysc_config;
> +
> /** @asn_hash_enable: True if GPU_L2_CONFIG_ASN_HASH_ENABLE must be set. */
> bool asn_hash_enable;
>
> @@ -138,6 +154,9 @@ struct panthor_device {
> /** @soc_data: Optional SoC data. */
> const struct panthor_soc_data *soc_data;
>
> + /** @reset_control: Optional external GPU reset. */
> + struct reset_control *reset_control;
> +
> /** @phys_addr: Physical address of the iomem region. */
> phys_addr_t phys_addr;
>
> diff --git a/drivers/gpu/drm/panthor/panthor_drv.c b/drivers/gpu/drm/panthor/panthor_drv.c
> index dc8d0191aef4..4d7b1230cd7b 100644
> --- a/drivers/gpu/drm/panthor/panthor_drv.c
> +++ b/drivers/gpu/drm/panthor/panthor_drv.c
> @@ -1872,12 +1872,26 @@ static struct attribute *panthor_attrs[] = {
>
> ATTRIBUTE_GROUPS(panthor);
>
> +static const struct panthor_sysc_config sysc_config_cix_sky1 = {
> + /* Sky1 needs these SYSC_ALLOC policies for correct display of GPU output. */
> + .alloc = {
> + 0x00230000, 0x00000023, 0x00000000, 0x00000000,
> + 0x00523222, 0x00523200, 0x00000022, 0x00000032,
> + },
> +};
These should not be in the driver but in the device tree bindings. If you extract
the system cache code into a separate patch we can have a proper discussion there.
> +
> +static const struct panthor_soc_data soc_data_cix_sky1 = {
> + .skip_devfreq = true,
> + .sysc_config = &sysc_config_cix_sky1,
> +};
> +
> static const struct panthor_soc_data soc_data_mediatek_mt8196 = {
> .asn_hash_enable = true,
> .asn_hash = { 0xb, 0xe, 0x0, },
> };
>
> static const struct of_device_id dt_match[] = {
> + { .compatible = "cix,sky1-mali", .data = &soc_data_cix_sky1 },
> { .compatible = "mediatek,mt8196-mali", .data = &soc_data_mediatek_mt8196, },
> { .compatible = "rockchip,rk3588-mali" },
> { .compatible = "arm,mali-valhall-csf" },
> diff --git a/drivers/gpu/drm/panthor/panthor_gpu.c b/drivers/gpu/drm/panthor/panthor_gpu.c
> index a383b04f101e..9a6963fe62ae 100644
> --- a/drivers/gpu/drm/panthor/panthor_gpu.c
> +++ b/drivers/gpu/drm/panthor/panthor_gpu.c
> @@ -63,6 +63,19 @@ static void panthor_gpu_coherency_set(struct panthor_device *ptdev)
> ptdev->gpu_info.selected_coherency);
> }
>
> +static void panthor_gpu_sysc_config_set(struct panthor_device *ptdev)
> +{
> + const struct panthor_sysc_config *config;
> + unsigned int i;
> +
> + if (!ptdev->soc_data || !ptdev->soc_data->sysc_config)
> + return;
> +
> + config = ptdev->soc_data->sysc_config;
> + for (i = 0; i < ARRAY_SIZE(config->alloc); i++)
> + gpu_write(ptdev->gpu->iomem, GPU_SYSC_ALLOC(i), config->alloc[i]);
This will crash for arch10 GPUs. I know that you're relying here on soc_data not having
the fields filled in, but if we move the sysc_config data to being provided by DT, we
need to check here.
> +}
> +
> static void panthor_gpu_l2_config_set(struct panthor_device *ptdev)
> {
> struct panthor_gpu *gpu = ptdev->gpu;
> @@ -317,6 +330,7 @@ int panthor_gpu_l2_power_on(struct panthor_device *ptdev)
> }
>
> /* Set the desired coherency mode and L2 config before the power up of L2 */
> + panthor_gpu_sysc_config_set(ptdev);
> panthor_gpu_coherency_set(ptdev);
> panthor_gpu_l2_config_set(ptdev);
>
> diff --git a/drivers/gpu/drm/panthor/panthor_gpu_regs.h b/drivers/gpu/drm/panthor/panthor_gpu_regs.h
> index 4c5b953796e4..f9b65a4dae83 100644
> --- a/drivers/gpu/drm/panthor/panthor_gpu_regs.h
> +++ b/drivers/gpu/drm/panthor/panthor_gpu_regs.h
> @@ -116,4 +116,6 @@
> #define GPU_COHERENCY_ACE 1
> #define GPU_COHERENCY_NONE 31
>
> +#define GPU_SYSC_ALLOC(n) (0x340 + (n) * 4)
> +
> #endif /* __PANTHOR_GPU_REGS_H__ */
> --
> 2.53.0
>
Best regards,
Liviu
--
====================
| I would like to |
| fix the world, |
| but they're not |
| giving me the |
\ source code! /
---------------
¯\_(ツ)_/¯
next prev parent reply other threads:[~2026-11-10 10:04 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-09 12:08 [PATCH v2 0/3] Add initial GPU " Cunyuan Liu
2026-11-09 12:08 ` [PATCH v2 1/3] dt-bindings: gpu: Add CIX Sky1 Mali support Cunyuan Liu
2026-11-09 12:09 ` [PATCH v2 2/3] drm/panthor: Add initial support for CIX Sky1 Cunyuan Liu
2026-11-10 10:03 ` Liviu Dudau [this message]
2026-11-09 12:09 ` [PATCH v2 3/3] arm64: dts: cix: Add GPU support for Sky1 Orion O6 Cunyuan Liu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=avLsAlL048MtIX4x@e142607 \
--to=liviu.dudau@arm.com \
--cc=aliceryhl@google.com \
--cc=boris.brezillon@collabora.com \
--cc=cix-kernel-upstream@cixtech.com \
--cc=conor+dt@kernel.org \
--cc=cunyuan.liu@cixtech.com \
--cc=daniel.almeida@collabora.com \
--cc=devicetree@vger.kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=fugang.duan@cixtech.com \
--cc=gary.yang@cixtech.com \
--cc=krzk+dt@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=p.zabel@pengutronix.de \
--cc=robh@kernel.org \
--cc=robin.wang@cixtech.com \
--cc=rust-for-linux@vger.kernel.org \
--cc=steven.price@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®