[Intel-gfx] [PATCH v3 02/11] drm/i915: Read graphics/media/display arch version from hw
Jani Nikula
jani.nikula at linux.intel.com
Thu Sep 1 07:58:20 UTC 2022
On Wed, 31 Aug 2022, Radhakrishna Sripada <radhakrishna.sripada at intel.com> wrote:
> From: Matt Roper <matthew.d.roper at intel.com>
>
> Going forward, the hardware teams no longer consider new platforms to
> have a "generation" in the way we've defined it for past platforms.
> Instead, each IP block (graphics, media, display) will have their own
> architecture major.minor versions and stepping ID's which should be read
> directly from a register in the MMIO space. New hardware programming
> styles, features, and workarounds should be conditional solely on the
> architecture version, and should no longer be derived from the PCI
> device ID, revision ID, or platform-specific feature flags.
>
> Bspec: 63361, 64111
>
> v2:
> - Move the IP version readout to intel_device_info.c
> - Convert the macro into a function
>
> Signed-off-by: Matt Roper <matthew.d.roper at intel.com>
> Signed-off-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
> Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada at intel.com>
> ---
> drivers/gpu/drm/i915/gt/intel_gt_regs.h | 2 +
> drivers/gpu/drm/i915/i915_driver.c | 2 +
> drivers/gpu/drm/i915/i915_drv.h | 2 +
> drivers/gpu/drm/i915/i915_pci.c | 1 +
> drivers/gpu/drm/i915/i915_reg.h | 6 ++
> drivers/gpu/drm/i915/intel_device_info.c | 73 ++++++++++++++++++++++++
> drivers/gpu/drm/i915/intel_device_info.h | 3 +
> 7 files changed, 89 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_gt_regs.h b/drivers/gpu/drm/i915/gt/intel_gt_regs.h
> index d414785003cc..579da62158c4 100644
> --- a/drivers/gpu/drm/i915/gt/intel_gt_regs.h
> +++ b/drivers/gpu/drm/i915/gt/intel_gt_regs.h
> @@ -39,6 +39,8 @@
> #define FORCEWAKE_ACK_RENDER_GEN9 _MMIO(0xd84)
> #define FORCEWAKE_ACK_MEDIA_GEN9 _MMIO(0xd88)
>
> +#define GMD_ID_GRAPHICS _MMIO(0xd8c)
> +
> #define MCFG_MCR_SELECTOR _MMIO(0xfd0)
> #define SF_MCR_SELECTOR _MMIO(0xfd8)
> #define GEN8_MCR_SELECTOR _MMIO(0xfdc)
> diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
> index 3aedc33ded57..5826c70d6fa5 100644
> --- a/drivers/gpu/drm/i915/i915_driver.c
> +++ b/drivers/gpu/drm/i915/i915_driver.c
> @@ -323,6 +323,8 @@ static int i915_driver_early_probe(struct drm_i915_private *dev_priv)
> if (i915_inject_probe_failure(dev_priv))
> return -ENODEV;
>
> + intel_device_info_runtime_init_early(dev_priv);
> +
> intel_device_info_subplatform_init(dev_priv);
Hmm, so why not move the subplatform init call to runtime init early?
> intel_step_init(dev_priv);
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index bf60593a4ce5..935ff3486fef 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -936,6 +936,8 @@ IS_SUBPLATFORM(const struct drm_i915_private *i915,
>
> #define HAS_GMCH(dev_priv) (INTEL_INFO(dev_priv)->display.has_gmch)
>
> +#define HAS_GMD_ID(i915) INTEL_INFO(i915)->has_gmd_id
> +
> #define HAS_LSPCON(dev_priv) (IS_DISPLAY_VER(dev_priv, 9, 10))
>
> #define HAS_L3_CCS_READ(i915) (INTEL_INFO(i915)->has_l3_ccs_read)
> diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
> index 72577e327c71..9772c315185d 100644
> --- a/drivers/gpu/drm/i915/i915_pci.c
> +++ b/drivers/gpu/drm/i915/i915_pci.c
> @@ -1129,6 +1129,7 @@ static const struct intel_device_info mtl_info = {
> PLATFORM(INTEL_METEORLAKE),
> .display.has_modular_fia = 1,
> .has_flat_ccs = 0,
> + .has_gmd_id = 1,
> .has_snoop = 1,
> .__runtime.memory_regions = REGION_SMEM | REGION_STOLEN_LMEM,
> .__runtime.platform_engine_mask = BIT(RCS0) | BIT(BCS0) | BIT(CCS0),
> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> index 5e6239864c35..f52ed6d00030 100644
> --- a/drivers/gpu/drm/i915/i915_reg.h
> +++ b/drivers/gpu/drm/i915/i915_reg.h
> @@ -5798,6 +5798,11 @@
> #define ICL_DSSM_CDCLK_PLL_REFCLK_19_2MHz (1 << 29)
> #define ICL_DSSM_CDCLK_PLL_REFCLK_38_4MHz (2 << 29)
>
> +#define GMD_ID_DISPLAY _MMIO(0x510a0)
> +#define GMD_ID_ARCH_MASK REG_GENMASK(31, 22)
> +#define GMD_ID_RELEASE_MASK REG_GENMASK(21, 14)
> +#define GMD_ID_STEP REG_GENMASK(5, 0)
> +
> /*GEN11 chicken */
> #define _PIPEA_CHICKEN 0x70038
> #define _PIPEB_CHICKEN 0x71038
> @@ -8298,4 +8303,5 @@ enum skl_power_gate {
> #define MTL_LATENCY_LEVEL_EVEN_MASK REG_GENMASK(12, 0)
> #define MTL_LATENCY_LEVEL_ODD_MASK REG_GENMASK(28, 16)
>
> +#define MTL_MEDIA_GSI_BASE 0x380000
Blank line here.
> #endif /* _I915_REG_H_ */
> diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
> index 8ff66b4e11c1..e36ba0520be7 100644
> --- a/drivers/gpu/drm/i915/intel_device_info.c
> +++ b/drivers/gpu/drm/i915/intel_device_info.c
> @@ -29,6 +29,7 @@
>
> #include "display/intel_cdclk.h"
> #include "display/intel_de.h"
> +#include "gt/intel_gt_regs.h"
> #include "intel_device_info.h"
> #include "i915_drv.h"
> #include "i915_utils.h"
> @@ -133,6 +134,78 @@ void intel_device_info_print(const struct intel_device_info *info,
> drm_printf(p, "rawclk rate: %u kHz\n", runtime->rawclk_freq);
> }
>
> +static void ip_ver_read(struct drm_i915_private *i915, u32 offset, struct ip_version *ip,
> + const struct ip_version *dr_ip)
> +{
> + struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
> + void __iomem *addr;
> + u32 ver;
> +
> + addr = pci_iomap_range(pdev, 0, offset, sizeof(u32));
> + if (drm_WARN_ON(&i915->drm, !addr))
> + return;
> +
> + ver = ioread32(addr);
> + pci_iounmap(pdev, addr);
> +
> + ip->ver = REG_FIELD_GET(GMD_ID_ARCH_MASK, ver);
> + ip->rel = REG_FIELD_GET(GMD_ID_RELEASE_MASK, ver);
> + ip->step = REG_FIELD_GET(GMD_ID_STEP, ver);
> +
> + /* Sanity check against expected versions from device info */
> + if (ip->ver != dr_ip->ver || ip->rel > dr_ip->rel)
> + drm_dbg(&i915->drm,
> + "Hardware reports GMD IP version %u.%u but minimum expected is %u.%u\n",
> + ip->ver, ip->rel, dr_ip->ver, dr_ip->rel);
So this doesn't work because you pass in the same pointer in both ip and
dr_ip, and you overwrite it above. These will always match. I think just
drop the 2nd pointer from parameters, and make a local copy. Or read the
values to a local copy, check them, and assign.
> +}
> +
> +/**
> + * intel_ipver_early_init - setup IP version values
> + * @dev_priv: device private
We don't need (or want) kernel-doc for static functions. It just
pollutes the generated documentation. The informative part can stay as a
regular comment.
> + *
> + * Setup the graphics version for the current device. This must be done before
> + * any code that performs checks on GRAPHICS_VER or DISPLAY_VER, so this
> + * function should be called very early in the driver initialization sequence.
> + *
> + * Regular MMIO access is not yet setup at the point this function is called so
> + * we peek at the appropriate MMIO offset directly. The GMD_ID register is
> + * part of an 'always on' power well by design, so we don't need to worry about
> + * forcewake while reading it.
> + */
> +static void intel_ipver_early_init(struct drm_i915_private *i915)
> +{
> + struct intel_runtime_info *rtime = RUNTIME_INFO(i915);
> + const struct intel_runtime_info *drinfo = &INTEL_INFO(i915)->__runtime;
"runtime" and "info" have been used elsewhere as the variable
names. Whether they are good is debatable, but better to stick to one
naming.
> +
> + if (!HAS_GMD_ID(i915))
> + return;
> +
> + ip_ver_read(i915, i915_mmio_reg_offset(GMD_ID_GRAPHICS), &rtime->graphics,
> + &drinfo->graphics);
> + ip_ver_read(i915, i915_mmio_reg_offset(GMD_ID_DISPLAY), &rtime->display,
> + &drinfo->display);
> + ip_ver_read(i915, MTL_MEDIA_GSI_BASE + i915_mmio_reg_offset(GMD_ID_GRAPHICS),
> + &rtime->media, &drinfo->media);
> +}
> +
> +/**
> + * intel_device_info_runtime_init_early - initialize early runtime info
> + * @i915: the i915 device
Even if this isn't a static function, I think it does beg the question
does it need to be part of the generated documentation?
> + *
> + * Determine early intel_device_info fields at runtime.
> + *
> + * Use it when:
> + * - Early init of certain runtime info fields are to be initialized
> + *
> + * This function needs to be called:
> + * - before the MMIO has been setup as we are reading registers,
> + * - before the PCH has been detected,
> + * - before the first usage of the fields it can tweak.
> + */
> +void intel_device_info_runtime_init_early(struct drm_i915_private *i915)
> +{
> + intel_ipver_early_init(i915);
> +}
Blank line here.
> #undef INTEL_VGA_DEVICE
> #define INTEL_VGA_DEVICE(id, info) (id)
>
> diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h
> index 6511b25277dc..6031e729714e 100644
> --- a/drivers/gpu/drm/i915/intel_device_info.h
> +++ b/drivers/gpu/drm/i915/intel_device_info.h
> @@ -152,6 +152,7 @@ enum intel_ppgtt_type {
> func(has_4tile); \
> func(has_flat_ccs); \
> func(has_global_mocs); \
> + func(has_gmd_id); \
> func(has_gt_uc); \
> func(has_heci_pxp); \
> func(has_heci_gscfi); \
> @@ -197,6 +198,7 @@ enum intel_ppgtt_type {
> struct ip_version {
> u8 ver;
> u8 rel;
> + u8 step;
> };
>
> struct intel_runtime_info {
> @@ -307,6 +309,7 @@ struct intel_driver_caps {
>
> const char *intel_platform_name(enum intel_platform platform);
>
> +void intel_device_info_runtime_init_early(struct drm_i915_private *dev_priv);
> void intel_device_info_subplatform_init(struct drm_i915_private *dev_priv);
> void intel_device_info_runtime_init(struct drm_i915_private *dev_priv);
--
Jani Nikula, Intel Open Source Graphics Center
More information about the dri-devel
mailing list