public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
	intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH 1/6] drm/i915/ttm: add ttm_buddy_man
Date: Tue, 8 Jun 2021 09:11:05 +0100	[thread overview]
Message-ID: <125c067c-b94a-c218-5ebc-a57b7d75402d@intel.com> (raw)
In-Reply-To: <eaad0953f7699a906cc590023aab9d11a93df005.camel@linux.intel.com>

On 08/06/2021 08:11, Thomas Hellström wrote:
> On Mon, 2021-06-07 at 19:22 +0100, Matthew Auld wrote:
>> Add back our standalone i915_buddy allocator and integrate it into a
>> ttm_resource_manager. This will plug into our ttm backend for
>> managing
>> device local-memory in the next couple of patches.
>>
>> Signed-off-by: Matthew Auld <matthew.auld@intel.com>
>> Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
>> ---
>>
> 
> Since the buddy + selftests have been part of the driver before, I
> didn't review them separately, but for the TTM interface, some minor
> comments below. With those fixed,
> 
> Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> 
> 
>> diff --git a/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c
>> b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c
>> new file mode 100644
>> index 000000000000..d7bf37be1932
>> --- /dev/null
>> +++ b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c
>> @@ -0,0 +1,246 @@
>> +// SPDX-License-Identifier: MIT
>> +/*
>> + * Copyright © 2021 Intel Corporation
>> + */
>> +
>> +#include <linux/slab.h>
>> +
>> +#include <drm/ttm/ttm_bo_driver.h>
>> +#include <drm/ttm/ttm_placement.h>
>> +
>> +#include "i915_ttm_buddy_manager.h"
>> +
>> +#include "i915_buddy.h"
>> +#include "i915_gem.h"
>> +
>> +struct i915_ttm_buddy_manager {
>> +       struct ttm_resource_manager manager;
>> +       struct i915_buddy_mm mm;
>> +       struct list_head reserved;
>> +       struct mutex lock;
>> +};
>> +
>> +static inline struct i915_ttm_buddy_manager *
> 
> "inline" shouldn't be needed here.
> 
>> +to_buddy_manager(struct ttm_resource_manager *man)
>> +{
>> +       return container_of(man, struct i915_ttm_buddy_manager,
>> manager);
>> +}
>> +
>> +static int i915_ttm_buddy_man_alloc(struct ttm_resource_manager
>> *man,
>> +                                   struct ttm_buffer_object *bo,
>> +                                   const struct ttm_place *place,
>> +                                   struct ttm_resource **res)
>> +{
>> +       struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
>> +       struct i915_ttm_buddy_resource *bman_res;
>> +       struct i915_buddy_mm *mm = &bman->mm;
>> +       unsigned long n_pages;
>> +       unsigned int min_order;
>> +       u64 size;
>> +       int err;
>> +
>> +       GEM_BUG_ON(place->fpfn || place->lpfn);
>> +       GEM_BUG_ON(bo->page_alignment < mm->chunk_size);
>> +
>> +       bman_res = kzalloc(sizeof(*bman_res), GFP_KERNEL);
>> +       if (!bman_res)
>> +               return -ENOMEM;
>> +
>> +       ttm_resource_init(bo, place, &bman_res->base);
>> +       INIT_LIST_HEAD(&bman_res->blocks);
>> +       bman_res->mm = mm;
>> +
>> +       GEM_BUG_ON(!bman_res->base.num_pages);
>> +       size = bman_res->base.num_pages << PAGE_SHIFT;
>> +
>> +       min_order = ilog2(bo->page_alignment) - ilog2(mm-
>>> chunk_size);
>> +       if (place->flags & TTM_PL_FLAG_CONTIGUOUS) {
>> +               size = roundup_pow_of_two(size);
>> +               min_order = ilog2(size) - ilog2(mm->chunk_size);
>> +       }
>> +
>> +       if (size > mm->size) {
>> +               err = -E2BIG;
>> +               goto err_free_res;
>> +       }
>> +
>> +       n_pages = size >> ilog2(mm->chunk_size);
>> +
>> +       do {
>> +               struct i915_buddy_block *block;
>> +               unsigned int order;
>> +
>> +               order = fls(n_pages) - 1;
>> +               GEM_BUG_ON(order > mm->max_order);
>> +               GEM_BUG_ON(order < min_order);
>> +
>> +               do {
>> +                       mutex_lock(&bman->lock);
>> +                       block = i915_buddy_alloc(mm, order);
>> +                       mutex_unlock(&bman->lock);
>> +                       if (!IS_ERR(block))
>> +                               break;
>> +
>> +                       if (order-- == min_order) {
>> +                               err = -ENXIO;
> 
> IIRC, TTM relies on -ENOSPC to retry with evictions.

Ah, right. We convert that back to -ENXIO in the upper levels somewhere?

> 
>> +                               goto err_free_blocks;
>> +                       }
>> +               } while (1);
>> +
>> +               n_pages -= BIT(order);
>> +
>> +               list_add_tail(&block->link, &bman_res->blocks);
>> +
>> +               if (!n_pages)
>> +                       break;
>> +       } while (1);
>> +
>> +       *res = &bman_res->base;
>> +       return 0;
>> +
>> +err_free_blocks:
>> +       mutex_lock(&bman->lock);
>> +       i915_buddy_free_list(mm, &bman_res->blocks);
>> +       mutex_unlock(&bman->lock);
>> +err_free_res:
>> +       kfree(bman_res);
>> +       return err;
>> +}
>> +
> 
> /Thomas
> 
> 
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2021-06-08  8:11 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-07 18:22 [Intel-gfx] [PATCH 0/6] Add back the buddy allocator Matthew Auld
2021-06-07 18:22 ` [Intel-gfx] [PATCH 1/6] drm/i915/ttm: add ttm_buddy_man Matthew Auld
2021-06-08  7:11   ` Thomas Hellström
2021-06-08  8:11     ` Matthew Auld [this message]
2021-06-08  8:15       ` Thomas Hellström
2021-06-07 18:22 ` [Intel-gfx] [PATCH 2/6] drm/i915/ttm: add i915_sg_from_buddy_resource Matthew Auld
2021-06-08  7:19   ` Thomas Hellström
2021-06-07 18:22 ` [Intel-gfx] [PATCH 3/6] drm/i915/ttm: Calculate the object placement at get_pages time Matthew Auld
2021-06-07 18:22 ` [Intel-gfx] [PATCH 4/6] drm/i915/ttm: pass along the I915_BO_ALLOC_CONTIGUOUS Matthew Auld
2021-06-08  7:26   ` Thomas Hellström
2021-06-08  8:08     ` Matthew Auld
2021-06-07 18:22 ` [Intel-gfx] [PATCH 5/6] drm/i915/ttm: switch over to ttm_buddy_man Matthew Auld
2021-06-08  7:34   ` Thomas Hellström
2021-06-08  7:39   ` Thomas Hellström
2021-06-08  8:13     ` Matthew Auld
2021-06-07 18:22 ` [Intel-gfx] [PATCH 6/6] drm/i915/ttm: restore min_page_size behaviour Matthew Auld
2021-06-08  7:41   ` Thomas Hellström
2021-06-07 20:08 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Add back the buddy allocator 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=125c067c-b94a-c218-5ebc-a57b7d75402d@intel.com \
    --to=matthew.auld@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --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