From: Matthew Brost <matthew.brost@intel.com>
To: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Cc: intel-xe@lists.freedesktop.org,
"Somalapuram Amaranath" <Amaranath.Somalapuram@amd.com>,
"Christian König" <christian.koenig@amd.com>,
"Paulo Zanoni" <paulo.r.zanoni@intel.com>,
"Simona Vetter" <simona.vetter@ffwll.ch>,
dri-devel@lists.freedesktop.org
Subject: Re: [PATCH v11 5/8] drm/ttm: Add a macro to perform LRU iteration
Date: Sun, 20 Oct 2024 18:57:09 +0000 [thread overview]
Message-ID: <ZxVShUFo9PnyWzae@DUT025-TGLU.fm.intel.com> (raw)
In-Reply-To: <20241016085559.11840-6-thomas.hellstrom@linux.intel.com>
On Wed, Oct 16, 2024 at 10:55:56AM +0200, Thomas Hellström wrote:
> Following the design direction communicated here:
>
> https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2d67@amd.com/T/#ma918844aa8a6efe8768fdcda0c6590d5c93850c9
>
> Export a LRU walker for driver shrinker use. The walker
> initially supports only trylocking, since that's the
> method used by shrinkes. The walker makes use of
> scoped_guard() to allow exiting from the LRU walk loop
> without performing any explicit unlocking or
> cleanup.
>
> v8:
> - Split out from another patch.
> - Use a struct for bool arguments to increase readability (Matt Brost).
> - Unmap user-space cpu-mappings before shrinking pages.
> - Explain non-fatal error codes (Matt Brost)
>
> v10:
> - Instead of using the existing helper, Wrap the interface inside out and
> provide a loop to de-midlayer things the LRU iteration (Christian König).
> - Removing the R-B by Matt Brost since the patch was significantly changed.
>
> v11:
> - Split the patch up to include just the LRU walk helper.
>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
> drivers/gpu/drm/ttm/ttm_bo_util.c | 140 +++++++++++++++++++++++++++++-
> include/drm/ttm/ttm_bo.h | 71 +++++++++++++++
> 2 files changed, 207 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
> index 917096bd5f68..0cac02a9764c 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo_util.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
> @@ -769,12 +769,10 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo)
> return ret;
> }
>
> -static bool ttm_lru_walk_trylock(struct ttm_lru_walk *walk,
> +static bool ttm_lru_walk_trylock(struct ttm_operation_ctx *ctx,
> struct ttm_buffer_object *bo,
> bool *needs_unlock)
> {
> - struct ttm_operation_ctx *ctx = walk->ctx;
> -
> *needs_unlock = false;
>
> if (dma_resv_trylock(bo->base.resv)) {
> @@ -877,7 +875,7 @@ s64 ttm_lru_walk_for_evict(struct ttm_lru_walk *walk, struct ttm_device *bdev,
> * since if we do it the other way around, and the trylock fails,
> * we need to drop the lru lock to put the bo.
> */
> - if (ttm_lru_walk_trylock(walk, bo, &bo_needs_unlock))
> + if (ttm_lru_walk_trylock(walk->ctx, bo, &bo_needs_unlock))
> bo_locked = true;
> else if (!walk->ticket || walk->ctx->no_wait_gpu ||
> walk->trylock_only)
> @@ -920,3 +918,137 @@ s64 ttm_lru_walk_for_evict(struct ttm_lru_walk *walk, struct ttm_device *bdev,
>
> return progress;
> }
> +EXPORT_SYMBOL(ttm_lru_walk_for_evict);
> +
> +static void ttm_bo_lru_cursor_cleanup_bo(struct ttm_bo_lru_cursor *curs)
> +{
> + struct ttm_buffer_object *bo = curs->bo;
> +
> + if (bo) {
> + if (curs->needs_unlock)
> + dma_resv_unlock(bo->base.resv);
> + ttm_bo_put(bo);
> + curs->bo = NULL;
> + }
> +}
> +
> +/**
> + * ttm_bo_lru_cursor_fini() - Stop using a struct ttm_bo_lru_cursor
> + * and clean up any iteration it was used for.
> + * @curs: The cursor.
> + */
> +void ttm_bo_lru_cursor_fini(struct ttm_bo_lru_cursor *curs)
> +{
> + spinlock_t *lru_lock = &curs->res_curs.man->bdev->lru_lock;
> +
> + ttm_bo_lru_cursor_cleanup_bo(curs);
> + spin_lock(lru_lock);
> + ttm_resource_cursor_fini(&curs->res_curs);
> + spin_unlock(lru_lock);
> +}
> +EXPORT_SYMBOL(ttm_bo_lru_cursor_fini);
> +
> +/**
> + * ttm_bo_lru_cursor_init() - Initialize a struct ttm_bo_lru_cursor
> + * @curs: The ttm_bo_lru_cursor to initialize.
> + * @man: The ttm resource_manager whose LRU lists to iterate over.
> + * @ctx: The ttm_operation_ctx to govern the locking.
> + *
> + * Initialize a struct ttm_bo_lru_cursor. Currently only trylocking
> + * or prelocked buffer objects are available as detailed by
> + * @ctx::resv and @ctx::allow_res_evict. Ticketlocking is not
> + * supported.
> + *
> + * Return: Pointer to @curs. The function does not fail.
> + */
> +struct ttm_bo_lru_cursor *
> +ttm_bo_lru_cursor_init(struct ttm_bo_lru_cursor *curs,
> + struct ttm_resource_manager *man,
> + struct ttm_operation_ctx *ctx)
> +{
> + memset(curs, 0, sizeof(*curs));
> + ttm_resource_cursor_init(&curs->res_curs, man);
> + curs->ctx = ctx;
> +
> + return curs;
> +}
> +EXPORT_SYMBOL(ttm_bo_lru_cursor_init);
> +
> +static struct ttm_buffer_object *
> +ttm_bo_from_res_reserved(struct ttm_resource *res, struct ttm_bo_lru_cursor *curs)
> +{
> + struct ttm_buffer_object *bo = res->bo;
> +
> + if (!ttm_lru_walk_trylock(curs->ctx, bo, &curs->needs_unlock))
> + return NULL;
> +
> + if (!ttm_bo_get_unless_zero(bo)) {
> + if (curs->needs_unlock)
> + dma_resv_unlock(bo->base.resv);
> + return NULL;
> + }
> +
> + curs->bo = bo;
> + return bo;
> +}
> +
> +/**
> + * ttm_bo_lru_cursor_next() - Continue iterating a manager's LRU lists
> + * to find and lock buffer object.
> + * @curs: The cursor initialized using ttm_bo_lru_cursor_init() and
> + * ttm_bo_lru_cursor_first().
> + *
> + * Return: A pointer to a locked and reference-counted buffer object,
> + * or NULL if none could be found and looping should be terminated.
> + */
> +struct ttm_buffer_object *ttm_bo_lru_cursor_next(struct ttm_bo_lru_cursor *curs)
> +{
> + spinlock_t *lru_lock = &curs->res_curs.man->bdev->lru_lock;
> + struct ttm_resource *res = NULL;
> + struct ttm_buffer_object *bo;
> +
> + ttm_bo_lru_cursor_cleanup_bo(curs);
> +
> + spin_lock(lru_lock);
> + for (;;) {
> + res = ttm_resource_manager_next(&curs->res_curs);
> + if (!res)
> + break;
> +
> + bo = ttm_bo_from_res_reserved(res, curs);
> + if (bo)
> + break;
> + }
> +
> + spin_unlock(lru_lock);
> + return res ? bo : NULL;
> +}
> +EXPORT_SYMBOL(ttm_bo_lru_cursor_next);
> +
> +/**
> + * ttm_bo_lru_cursor_first() - Start iterating a manager's LRU lists
> + * to find and lock buffer object.
> + * @curs: The cursor initialized using ttm_bo_lru_cursor_init().
> + *
> + * Return: A pointer to a locked and reference-counted buffer object,
> + * or NULL if none could be found and looping should be terminated.
> + */
> +struct ttm_buffer_object *ttm_bo_lru_cursor_first(struct ttm_bo_lru_cursor *curs)
> +{
> + spinlock_t *lru_lock = &curs->res_curs.man->bdev->lru_lock;
> + struct ttm_buffer_object *bo;
> + struct ttm_resource *res;
> +
> + spin_lock(lru_lock);
> + res = ttm_resource_manager_first(&curs->res_curs);
> + if (!res) {
> + spin_unlock(lru_lock);
> + return NULL;
> + }
> +
> + bo = ttm_bo_from_res_reserved(res, curs);
> + spin_unlock(lru_lock);
> +
> + return bo ? bo : ttm_bo_lru_cursor_next(curs);
> +}
> +EXPORT_SYMBOL(ttm_bo_lru_cursor_first);
> diff --git a/include/drm/ttm/ttm_bo.h b/include/drm/ttm/ttm_bo.h
> index 5804408815be..91ecbb64f6c8 100644
> --- a/include/drm/ttm/ttm_bo.h
> +++ b/include/drm/ttm/ttm_bo.h
> @@ -465,4 +465,75 @@ void ttm_bo_tt_destroy(struct ttm_buffer_object *bo);
> int ttm_bo_populate(struct ttm_buffer_object *bo,
> struct ttm_operation_ctx *ctx);
>
> +/* Driver LRU walk helpers initially targeted for shrinking. */
> +
> +/**
> + * struct ttm_bo_lru_cursor - Iterator cursor for TTM LRU list looping
> + */
> +struct ttm_bo_lru_cursor {
> + /** @res_curs: Embedded struct ttm_resource_cursor. */
> + struct ttm_resource_cursor res_curs;
> + /**
> + * @ctx: The struct ttm_operation_ctx used while looping.
> + * governs the locking mode.
> + */
> + struct ttm_operation_ctx *ctx;
> + /**
> + * @bo: Buffer object pointer if a buffer object is refcounted,
> + * NULL otherwise.
> + */
> + struct ttm_buffer_object *bo;
> + /**
> + * @needs_unlock: Valid iff @bo != NULL. The bo resv needs
> + * unlock before the next iteration or after loop exit.
> + */
> + bool needs_unlock;
> +};
> +
> +void ttm_bo_lru_cursor_fini(struct ttm_bo_lru_cursor *curs);
> +
> +struct ttm_bo_lru_cursor *
> +ttm_bo_lru_cursor_init(struct ttm_bo_lru_cursor *curs,
> + struct ttm_resource_manager *man,
> + struct ttm_operation_ctx *ctx);
> +
> +struct ttm_buffer_object *ttm_bo_lru_cursor_first(struct ttm_bo_lru_cursor *curs);
> +
> +struct ttm_buffer_object *ttm_bo_lru_cursor_next(struct ttm_bo_lru_cursor *curs);
> +
> +/*
> + * Defines needed to use autocleanup (linux/cleanup.h) with struct ttm_bo_lru_cursor.
> + */
> +DEFINE_CLASS(ttm_bo_lru_cursor, struct ttm_bo_lru_cursor *,
> + if (_T) {ttm_bo_lru_cursor_fini(_T); },
> + ttm_bo_lru_cursor_init(curs, man, ctx),
> + struct ttm_bo_lru_cursor *curs, struct ttm_resource_manager *man,
> + struct ttm_operation_ctx *ctx);
> +static inline void *
> +class_ttm_bo_lru_cursor_lock_ptr(class_ttm_bo_lru_cursor_t *_T)
> +{ return *_T; }
> +
> +/**
> + * ttm_bo_lru_for_each_reserved_guarded() - Iterate over buffer objects owning
> + * resources on LRU lists.
> + * @_cursor: struct ttm_bo_lru_cursor to use for the iteration.
> + * @_man: The resource manager whose LRU lists to iterate over.
> + * @_ctx: The struct ttm_operation_context to govern the @_bo locking.
> + * @_bo: The struct ttm_buffer_object pointer pointing to the buffer object
> + * for the current iteration.
> + *
> + * Iterate over all resources of @_man and for each resource, attempt to
> + * reference and lock (using the locking mode detailed in @_ctx) the buffer
> + * object it points to. If successful, assign @_bo to the address of the
> + * buffer object and update @_cursor. The iteration is guarded in the
> + * sense that @_cursor will be initialized before looping start and cleaned
> + * up at looping termination, even if terminated prematurely by, for
> + * example a return or break statement. Exiting the loop will also unlock
> + * (if needed) and unreference @_bo.
> + */
> +#define ttm_bo_lru_for_each_reserved_guarded(_cursor, _man, _ctx, _bo) \
> + scoped_guard(ttm_bo_lru_cursor, _cursor, _man, _ctx) \
> + for ((_bo) = ttm_bo_lru_cursor_first(_cursor); (_bo); \
Actually 1 nit.
scoped_guard()
for()
To show the scope of the guard. RB holds though.
Matt
> + (_bo) = ttm_bo_lru_cursor_next(_cursor))
> +
> #endif
> --
> 2.46.0
>
next prev parent reply other threads:[~2024-10-20 18:58 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-16 8:55 [PATCH v11 0/8] TTM shrinker helpers and xe buffer object shrinker Thomas Hellström
2024-10-16 8:55 ` [PATCH v11 1/8] drm/ttm: Balance ttm_resource_cursor_init() and ttm_resource_cursor_fini() Thomas Hellström
2024-10-16 8:55 ` [PATCH v11 2/8] drm/ttm: Add a virtual base class for graphics memory backup Thomas Hellström
2024-11-08 14:32 ` Christian König
2024-11-11 14:38 ` Thomas Hellström
2024-11-11 16:13 ` Christian König
2024-10-16 8:55 ` [PATCH v11 3/8] drm/ttm/pool: Provide a helper to shrink pages Thomas Hellström
2024-10-20 18:38 ` Matthew Brost
2024-10-16 8:55 ` [PATCH v11 4/8] drm/ttm: Use fault-injection to test error paths Thomas Hellström
2024-10-16 8:55 ` [PATCH v11 5/8] drm/ttm: Add a macro to perform LRU iteration Thomas Hellström
2024-10-20 18:54 ` Matthew Brost
2024-10-20 18:57 ` Matthew Brost [this message]
2024-10-16 8:55 ` [PATCH v11 6/8] drm/ttm: Add helpers for shrinking Thomas Hellström
2024-10-21 21:37 ` Matthew Brost
2024-10-16 8:55 ` [PATCH v11 7/8] drm/xe: Add a shrinker for xe bos Thomas Hellström
2024-10-24 17:00 ` Matthew Brost
2024-10-16 8:55 ` [PATCH v11 8/8] drm/xe: Increase the XE_PL_TT watermark Thomas Hellström
2024-10-16 10:56 ` ✓ CI.Patch_applied: success for TTM shrinker helpers and xe buffer object shrinker (rev10) Patchwork
2024-10-16 10:57 ` ✗ CI.checkpatch: warning " Patchwork
2024-10-16 10:58 ` ✓ CI.KUnit: success " Patchwork
2024-10-16 11:09 ` ✓ CI.Build: " Patchwork
2024-10-16 11:11 ` ✓ CI.Hooks: " Patchwork
2024-10-16 11:13 ` ✗ CI.checksparse: warning " Patchwork
2024-10-16 11:59 ` ✗ CI.BAT: failure " Patchwork
2024-10-20 19:01 ` Matthew Brost
2024-10-21 10:50 ` Thomas Hellström
2024-10-17 1:10 ` ✗ CI.FULL: " Patchwork
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=ZxVShUFo9PnyWzae@DUT025-TGLU.fm.intel.com \
--to=matthew.brost@intel.com \
--cc=Amaranath.Somalapuram@amd.com \
--cc=christian.koenig@amd.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=paulo.r.zanoni@intel.com \
--cc=simona.vetter@ffwll.ch \
--cc=thomas.hellstrom@linux.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