From: "Ceraolo Spurio, Daniele" <daniele.ceraolospurio@intel.com>
To: Matt Roper <matthew.d.roper@intel.com>,
<intel-gfx@lists.freedesktop.org>
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [PATCH v3.1 10/14] drm/i915/uncore: Add GSI offset to uncore
Date: Thu, 8 Sep 2022 15:53:11 -0700 [thread overview]
Message-ID: <0c41e953-b90e-e4ac-24da-4573fe2c175e@intel.com> (raw)
In-Reply-To: <20220908224550.821257-1-matthew.d.roper@intel.com>
On 9/8/2022 3:45 PM, Matt Roper wrote:
> GT non-engine registers (referred to as "GSI" registers by the spec)
> have the same relative offsets on standalone media as they do on the
> primary GT, just with an additional "GSI offset" added to their MMIO
> address. If we store this GSI offset in the standalone media's
> intel_uncore structure, it can be automatically applied to all GSI reg
> reads/writes that happen on that GT, allowing us to re-use our existing
> GT code with minimal changes.
>
> Forcewake and shadowed register tables for the media GT (which will be
> added in a future patch) are listed as final addresses that already
> include the GSI offset, so we also need to add the GSI offset before
> doing lookups of registers in one of those tables.
>
> v2:
> - Add comment on raw_reg_*() macros explaining why we don't bother with
> GSI offsets in them. (Daniele)
>
> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Daniele
> ---
> drivers/gpu/drm/i915/gt/intel_gt_types.h | 1 +
> drivers/gpu/drm/i915/intel_uncore.c | 10 +++++--
> drivers/gpu/drm/i915/intel_uncore.h | 34 ++++++++++++++++++++++--
> 3 files changed, 41 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_gt_types.h b/drivers/gpu/drm/i915/gt/intel_gt_types.h
> index 0e139f7d75ed..82dc28643572 100644
> --- a/drivers/gpu/drm/i915/gt/intel_gt_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_gt_types.h
> @@ -274,6 +274,7 @@ struct intel_gt_definition {
> enum intel_gt_type type;
> char *name;
> u32 mapping_base;
> + u32 gsi_offset;
> intel_engine_mask_t engine_mask;
> };
>
> diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
> index 452b3a31e965..5cd423c7b646 100644
> --- a/drivers/gpu/drm/i915/intel_uncore.c
> +++ b/drivers/gpu/drm/i915/intel_uncore.c
> @@ -928,6 +928,9 @@ find_fw_domain(struct intel_uncore *uncore, u32 offset)
> {
> const struct intel_forcewake_range *entry;
>
> + if (IS_GSI_REG(offset))
> + offset += uncore->gsi_offset;
> +
> entry = BSEARCH(offset,
> uncore->fw_domains_table,
> uncore->fw_domains_table_entries,
> @@ -1143,6 +1146,9 @@ static bool is_shadowed(struct intel_uncore *uncore, u32 offset)
> if (drm_WARN_ON(&uncore->i915->drm, !uncore->shadowed_reg_table))
> return false;
>
> + if (IS_GSI_REG(offset))
> + offset += uncore->gsi_offset;
> +
> return BSEARCH(offset,
> uncore->shadowed_reg_table,
> uncore->shadowed_reg_table_entries,
> @@ -1995,8 +2001,8 @@ static int __fw_domain_init(struct intel_uncore *uncore,
>
> d->uncore = uncore;
> d->wake_count = 0;
> - d->reg_set = uncore->regs + i915_mmio_reg_offset(reg_set);
> - d->reg_ack = uncore->regs + i915_mmio_reg_offset(reg_ack);
> + d->reg_set = uncore->regs + i915_mmio_reg_offset(reg_set) + uncore->gsi_offset;
> + d->reg_ack = uncore->regs + i915_mmio_reg_offset(reg_ack) + uncore->gsi_offset;
>
> d->id = domain_id;
>
> diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h
> index 4acb78a03233..5022bac80b67 100644
> --- a/drivers/gpu/drm/i915/intel_uncore.h
> +++ b/drivers/gpu/drm/i915/intel_uncore.h
> @@ -136,6 +136,16 @@ struct intel_uncore {
>
> spinlock_t lock; /** lock is also taken in irq contexts. */
>
> + /*
> + * Do we need to apply an additional offset to reach the beginning
> + * of the basic non-engine GT registers (referred to as "GSI" on
> + * newer platforms, or "GT block" on older platforms)? If so, we'll
> + * track that here and apply it transparently to registers in the
> + * appropriate range to maintain compatibility with our existing
> + * register definitions and GT code.
> + */
> + u32 gsi_offset;
> +
> unsigned int flags;
> #define UNCORE_HAS_FORCEWAKE BIT(0)
> #define UNCORE_HAS_FPGA_DBG_UNCLAIMED BIT(1)
> @@ -294,19 +304,27 @@ intel_wait_for_register_fw(struct intel_uncore *uncore,
> 2, timeout_ms, NULL);
> }
>
> +#define IS_GSI_REG(reg) ((reg) < 0x40000)
> +
> /* register access functions */
> #define __raw_read(x__, s__) \
> static inline u##x__ __raw_uncore_read##x__(const struct intel_uncore *uncore, \
> i915_reg_t reg) \
> { \
> - return read##s__(uncore->regs + i915_mmio_reg_offset(reg)); \
> + u32 offset = i915_mmio_reg_offset(reg); \
> + if (IS_GSI_REG(offset)) \
> + offset += uncore->gsi_offset; \
> + return read##s__(uncore->regs + offset); \
> }
>
> #define __raw_write(x__, s__) \
> static inline void __raw_uncore_write##x__(const struct intel_uncore *uncore, \
> i915_reg_t reg, u##x__ val) \
> { \
> - write##s__(val, uncore->regs + i915_mmio_reg_offset(reg)); \
> + u32 offset = i915_mmio_reg_offset(reg); \
> + if (IS_GSI_REG(offset)) \
> + offset += uncore->gsi_offset; \
> + write##s__(val, uncore->regs + offset); \
> }
> __raw_read(8, b)
> __raw_read(16, w)
> @@ -447,6 +465,18 @@ static inline int intel_uncore_write_and_verify(struct intel_uncore *uncore,
> return (reg_val & mask) != expected_val ? -EINVAL : 0;
> }
>
> +/*
> + * The raw_reg_{read,write} macros are intended as a micro-optimization for
> + * interrupt handlers so that the pointer indirection on uncore->regs can
> + * be computed once (and presumably cached in a register) instead of generating
> + * extra load instructions for each MMIO access.
> + *
> + * Given that these macros are only intended for non-GSI interrupt registers
> + * (and the goal is to avoid extra instructions generated by the compiler),
> + * these macros do not account for uncore->gsi_offset. Any caller that needs
> + * to use these macros on a GSI register is responsible for adding the
> + * appropriate GSI offset to the 'base' parameter.
> + */
> #define raw_reg_read(base, reg) \
> readl(base + i915_mmio_reg_offset(reg))
> #define raw_reg_write(base, reg, value) \
next prev parent reply other threads:[~2022-09-08 22:53 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-06 23:49 [PATCH v3 00/14] i915: Add "standalone media" support for MTL Matt Roper
2022-09-06 23:49 ` [PATCH v3 01/14] drm/i915: Move locking and unclaimed check into mmio_debug_{suspend, resume} Matt Roper
2022-09-06 23:49 ` [PATCH v3 02/14] drm/i915: Only hook up uncore->debug for primary uncore Matt Roper
2022-09-06 23:49 ` [PATCH v3 03/14] drm/i915: Use managed allocations for extra uncore objects Matt Roper
2022-09-06 23:49 ` [PATCH v3 04/14] drm/i915: Drop intel_gt_tile_cleanup() Matt Roper
2022-09-07 0:07 ` Lucas De Marchi
2022-09-07 11:18 ` kernel test robot
2022-09-06 23:49 ` [PATCH v3 05/14] drm/i915: Prepare more multi-GT initialization Matt Roper
2022-09-08 16:19 ` Iddamsetty, Aravind
2022-09-06 23:49 ` [PATCH v3 06/14] drm/i915: Rename and expose common GT early init routine Matt Roper
2022-09-06 23:49 ` [PATCH v3 07/14] drm/i915: Use a DRM-managed action to release the PCI bridge device Matt Roper
2022-09-09 20:57 ` Sripada, Radhakrishna
2022-09-06 23:49 ` [PATCH v3 08/14] drm/i915: Initialize MMIO access for each GT Matt Roper
2022-09-08 20:52 ` Ceraolo Spurio, Daniele
2022-09-06 23:49 ` [PATCH v3 09/14] drm/i915: Handle each GT on init/release and suspend/resume Matt Roper
2022-09-08 20:55 ` Ceraolo Spurio, Daniele
2022-09-06 23:49 ` [PATCH v3 10/14] drm/i915/uncore: Add GSI offset to uncore Matt Roper
2022-09-08 21:16 ` Ceraolo Spurio, Daniele
2022-09-08 22:29 ` Matt Roper
2022-09-08 22:45 ` [PATCH v3.1 " Matt Roper
2022-09-08 22:53 ` Ceraolo Spurio, Daniele [this message]
2022-09-06 23:49 ` [PATCH v3 11/14] drm/i915/mtl: Add gsi_offset when emitting aux table invalidation Matt Roper
2022-09-07 16:16 ` Iddamsetty, Aravind
2022-09-06 23:49 ` [PATCH v3 12/14] drm/i915/xelpmp: Expose media as another GT Matt Roper
2022-09-08 16:22 ` Iddamsetty, Aravind
2022-09-06 23:49 ` [PATCH v3 13/14] drm/i915/mtl: Use primary GT's irq lock for media GT Matt Roper
2022-09-08 21:20 ` Ceraolo Spurio, Daniele
2022-09-06 23:49 ` [PATCH v3 14/14] drm/i915/mtl: Hook up interrupts for standalone media Matt Roper
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=0c41e953-b90e-e4ac-24da-4573fe2c175e@intel.com \
--to=daniele.ceraolospurio@intel.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=matthew.d.roper@intel.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;
as well as URLs for NNTP newsgroup(s).