On Thu, 20 Mar 2025 11:17:38 +0000
Karunika Choo <karunika.c...@arm.com> wrote:
> Mali-G715 introduces a new GPU_FEATURES register that provides
> information about GPU-wide supported features. The register value will
> be passed on to userspace via gpu_info. It also adds the following
> registers that are specific to the kernel driver only:
> - ASN_HASH_0~2
> - DOORBELL_FEATURES
> - PRFCNT_FEATURES
> - SYSC_ALLOC0~7
> - SYSC_PBHA_OVERRIDE0~3
>
> Additionally, Mali-G715 presents an 'Immortalis' naming variant
> depending on the shader core count and presence of Ray Intersection
> feature support.
>
> This patch adds:
> - support for correctly identifying the model names for the Mali-G715
> family of GPUs.
> - arch 11.8 FW binary support
> - reading and handling of GPU_FEATURES register
>
> Signed-off-by: Karunika Choo <karunika.c...@arm.com>
> ---
> drivers/gpu/drm/panthor/panthor_fw.c | 1 +
> drivers/gpu/drm/panthor/panthor_hw.c | 26 ++++++++++++++++++++++++++
> drivers/gpu/drm/panthor/panthor_regs.h | 12 ++++++++++++
> include/uapi/drm/panthor_drm.h | 3 +++
> 4 files changed, 42 insertions(+)
>
> diff --git a/drivers/gpu/drm/panthor/panthor_fw.c
> b/drivers/gpu/drm/panthor/panthor_fw.c
> index ecfbe0456f89..0b3fab95f26b 100644
> --- a/drivers/gpu/drm/panthor/panthor_fw.c
> +++ b/drivers/gpu/drm/panthor/panthor_fw.c
> @@ -1398,3 +1398,4 @@ int panthor_fw_init(struct panthor_device *ptdev)
> }
>
> MODULE_FIRMWARE("arm/mali/arch10.8/mali_csffw.bin");
> +MODULE_FIRMWARE("arm/mali/arch11.8/mali_csffw.bin");
> diff --git a/drivers/gpu/drm/panthor/panthor_hw.c
> b/drivers/gpu/drm/panthor/panthor_hw.c
> index 12183c04cd21..d04c8723ac98 100644
> --- a/drivers/gpu/drm/panthor/panthor_hw.c
> +++ b/drivers/gpu/drm/panthor/panthor_hw.c
> @@ -32,15 +32,34 @@ static void arch_10_8_gpu_info_init(struct panthor_device
> *ptdev)
> ptdev->gpu_info.l2_present = gpu_read64(ptdev, GPU_L2_PRESENT_LO);
> }
>
> +static void arch_11_8_gpu_info_init(struct panthor_device *ptdev)
> +{
> + arch_10_8_gpu_info_init(ptdev);
> +
> + ptdev->gpu_info.gpu_features = gpu_read64(ptdev, GPU_FEATURES_LO);
That's typically the sort of specialization I would have done directly
in some csf_gpu_info_init() with a:
if (ARCH_MAJOR(gpu_id) > 11) {
ptdev->gpu_info.gpu_features =
gpu_read64(ptdev, GPU_FEATURES_LO);
}
I do see a benefit in abstracting things away when the layout is
completely different or when registers are conflicting, but that's not
the case AFAICT. I think for this kind of tweaks, we'd rather stick to a
single function with a few conditionals.
Actually, if the registers were readable and returning 0 on v10, you
don't even need the if (ARCH_MAJOR(gpu_id) > 11).
> +}
> +
> static char *get_gpu_model_name(struct panthor_device *ptdev)
> {
> const u32 gpu_id = ptdev->gpu_info.gpu_id;
> const u32 product_id = GPU_PROD_ID_MAKE(GPU_ARCH_MAJOR(gpu_id),
> GPU_PROD_MAJOR(gpu_id));
> + const bool ray_intersection = !!(ptdev->gpu_info.gpu_features &
> + GPU_FEATURES_RAY_INTERSECTION);
> + const u8 shader_core_count = hweight64(ptdev->gpu_info.shader_present);
>
> switch (product_id) {
> case GPU_PROD_ID_MAKE(10, 7):
> return "Mali-G610";
> + case GPU_PROD_ID_MAKE(11, 2):
> + if (shader_core_count > 10 && ray_intersection)
> + return "Mali-G715-Immortalis";
> + else if (shader_core_count >= 7)
> + return "Mali-G715";
> +
> + fallthrough;
> + case GPU_PROD_ID_MAKE(11, 3):
> + return "Mali-G615";
> }
>
> return "(Unknown Mali GPU)";
> @@ -84,6 +103,13 @@ static struct panthor_hw panthor_hw_devices[] = {
> .gpu_info_init = arch_10_8_gpu_info_init,
> },
> },
> + {
> + .arch_id = GPU_ARCH_ID_MAKE(11, 8, 0),
> + .arch_mask = GPU_ARCH_ID_MAKE(0xFF, 0xFF, 0),
> + .ops = {
> + .gpu_info_init = arch_11_8_gpu_info_init,
> + }
> + },
> };
>
> static int init_gpu_id(struct panthor_device *ptdev)
> diff --git a/drivers/gpu/drm/panthor/panthor_regs.h
> b/drivers/gpu/drm/panthor/panthor_regs.h
> index d9e0769d6f1a..7bc2d838e704 100644
> --- a/drivers/gpu/drm/panthor/panthor_regs.h
> +++ b/drivers/gpu/drm/panthor/panthor_regs.h
> @@ -74,6 +74,11 @@
> #define GPU_PWR_OVERRIDE0 0x54
> #define GPU_PWR_OVERRIDE1 0x58
>
> +#define GPU_FEATURES_LO 0x60
> +#define GPU_FEATURES_HI 0x64
> +#define GPU_FEATURES_RAY_INTERSECTION BIT(2)
> +#define GPU_PRFCNT_FEATURES 0x68
> +
> #define GPU_TIMESTAMP_OFFSET_LO 0x88
> #define GPU_TIMESTAMP_OFFSET_HI 0x8C
> #define GPU_CYCLE_COUNT_LO 0x90
> @@ -88,6 +93,8 @@
>
> #define GPU_TEXTURE_FEATURES(n) (0xB0 + ((n) *
> 4))
>
> +#define GPU_DOORBELL_FEATURES 0xC0
> +
> #define GPU_SHADER_PRESENT_LO 0x100
> #define GPU_SHADER_PRESENT_HI 0x104
> #define GPU_TILER_PRESENT_LO 0x110
> @@ -132,6 +139,8 @@
>
> #define GPU_REVID 0x280
>
> +#define GPU_ASN_HASH(n) (0x2C0 + ((n) *
> 4))
> +
> #define GPU_COHERENCY_FEATURES 0x300
> #define GPU_COHERENCY_PROT_BIT(name) BIT(GPU_COHERENCY_ ##
> name)
>
> @@ -140,6 +149,9 @@
> #define GPU_COHERENCY_ACE_LITE 1
> #define GPU_COHERENCY_NONE 31
>
> +#define GPU_SYSC_PBHA_OVERRIDE(n) (0x320 + ((n) * 4))
> +#define GPU_SYSC_ALLOC(n) (0x340 + ((n) * 4))
> +
> #define MCU_CONTROL 0x700
> #define MCU_CONTROL_ENABLE 1
> #define MCU_CONTROL_AUTO 2
> diff --git a/include/uapi/drm/panthor_drm.h b/include/uapi/drm/panthor_drm.h
> index 97e2c4510e69..4aba8146af3b 100644
> --- a/include/uapi/drm/panthor_drm.h
> +++ b/include/uapi/drm/panthor_drm.h
> @@ -307,6 +307,9 @@ struct drm_panthor_gpu_info {
>
> /** @pad: MBZ. */
> __u32 pad;
> +
> + /** @gpu_features: Bitmask describing supported GPU-wide features */
> + __u64 gpu_features;
> };
>
> /**
