From: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
To: Matthew Brost <matthew.brost@intel.com>
Cc: intel-xe@lists.freedesktop.org
Subject: Re: [Intel-xe] [PATCH 5/6] drm/xe/bo: Evict VRAM to TT rather than to system
Date: Mon, 19 Jun 2023 14:00:08 +0200 [thread overview]
Message-ID: <b9ff37eb-744c-9e96-bf14-e26c29ff1192@linux.intel.com> (raw)
In-Reply-To: <ZIzKfNQjRmy5r30C@DUT025-TGLU.fm.intel.com>
On 6/16/23 22:47, Matthew Brost wrote:
> On Fri, Jun 16, 2023 at 11:55:03AM +0200, Thomas Hellström wrote:
>> The main difference is that we don't bounce and sync on eviction, allowing
>> for pipelined eviction. Moving forward we also need to be careful with
>> dma mappings which can be released in SYSTEM but may remain in TT.
>>
>> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
>> ---
>> drivers/gpu/drm/xe/xe_bo.c | 16 ++++++++++++++++
>> 1 file changed, 16 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
>> index 77d5c5710688..fd41114f2dfb 100644
>> --- a/drivers/gpu/drm/xe/xe_bo.c
>> +++ b/drivers/gpu/drm/xe/xe_bo.c
>> @@ -40,6 +40,20 @@ static struct ttm_placement sys_placement = {
>> .busy_placement = &sys_placement_flags,
>> };
>>
>> +static const struct ttm_place tt_placement_flags = {
>> + .fpfn = 0,
>> + .lpfn = 0,
>> + .mem_type = XE_PL_TT,
>> + .flags = 0,
>> +};
>> +
>> +static struct ttm_placement tt_placement = {
>> + .num_placement = 1,
>> + .placement = &tt_placement_flags,
>> + .num_busy_placement = 1,
>> + .busy_placement = &sys_placement_flags,
>> +};
>> +
>> bool mem_type_is_vram(u32 mem_type)
>> {
>> return mem_type >= XE_PL_VRAM0 && mem_type != XE_PL_STOLEN;
>> @@ -225,6 +239,8 @@ static void xe_evict_flags(struct ttm_buffer_object *tbo,
>> case XE_PL_VRAM0:
>> case XE_PL_VRAM1:
>> case XE_PL_STOLEN:
>> + *placement = tt_placement;
>> + break;
>> case XE_PL_TT:
>> default:
>> /* for now kick out to system */
> NIT, you can probably delete this comment now.
>
> Anyways:
> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Thanks for reviewing.
>
>> --
>> 2.40.1
>>
next prev parent reply other threads:[~2023-06-19 12:00 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-16 9:54 [Intel-xe] [PATCH 0/6] drm/xe: Eviction fixes and optimizations Thomas Hellström
2023-06-16 9:54 ` [Intel-xe] [PATCH 1/6] drm/xe/bo: Don't limit the TT manager to half of system memory Thomas Hellström
2023-06-16 15:09 ` Matthew Brost
2023-06-16 22:18 ` Souza, Jose
2023-06-16 9:55 ` [Intel-xe] [PATCH 2/6] drm/xe/bo: Fix swapin when moving to VRAM Thomas Hellström
2023-06-16 15:50 ` Matthew Brost
2023-06-16 9:55 ` [Intel-xe] [PATCH 3/6] drm/xe/bo: Avoid creating a system resource when allocating a fresh VRAM bo Thomas Hellström
2023-06-16 20:32 ` Matthew Brost
2023-06-19 9:51 ` Thomas Hellström
2023-06-16 9:55 ` [Intel-xe] [PATCH 4/6] drm/xe/bo: Gracefully handle errors from ttm_bo_move_accel_cleanup() Thomas Hellström
2023-06-16 20:36 ` Matthew Brost
2023-06-19 11:59 ` Thomas Hellström
2023-06-16 9:55 ` [Intel-xe] [PATCH 5/6] drm/xe/bo: Evict VRAM to TT rather than to system Thomas Hellström
2023-06-16 20:47 ` Matthew Brost
2023-06-19 12:00 ` Thomas Hellström [this message]
2023-06-16 9:55 ` [Intel-xe] [PATCH 6/6] drm/ttm: Don't shadow the operation context Thomas Hellström
2023-06-16 20:51 ` Matthew Brost
2023-06-16 10:09 ` [Intel-xe] ✓ CI.Patch_applied: success for drm/xe: Eviction fixes and optimizations Patchwork
2023-06-16 10:09 ` [Intel-xe] ✓ CI.checkpatch: " Patchwork
2023-06-16 10:10 ` [Intel-xe] ✓ CI.KUnit: " Patchwork
2023-06-16 10:14 ` [Intel-xe] ✓ CI.Build: " Patchwork
2023-06-16 10:15 ` [Intel-xe] ✓ CI.Hooks: " Patchwork
2023-06-16 10:16 ` [Intel-xe] ✗ CI.checksparse: warning " Patchwork
2023-06-16 11:00 ` [Intel-xe] ○ CI.BAT: info " 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=b9ff37eb-744c-9e96-bf14-e26c29ff1192@linux.intel.com \
--to=thomas.hellstrom@linux.intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.