From: "Marek Olšák" <maraeo@gmail.com>
To: "Christian König" <ckoenig.leichtzumerken@gmail.com>
Cc: amd-gfx mailing list <amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH 1/3] drm/amdgpu: add AMDGPU_GEM_CREATE_DISCARDABLE
Date: Tue, 10 May 2022 16:43:44 -0400 [thread overview]
Message-ID: <CAAxE2A7wOfoWWh5VUFmnhyhNeCQ086trJR2BgT+nAmsYZJTbVg@mail.gmail.com> (raw)
In-Reply-To: <CAAxE2A4Bny50ywdTXi0MBV_Pb-onx0yVhPOsb2Lk9XtbJ1R4rQ@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 5170 bytes --]
A better flag name would be:
AMDGPU_GEM_CREATE_BEST_PLACEMENT_OR_DISCARD
Marek
On Tue, May 10, 2022 at 4:13 PM Marek Olšák <maraeo@gmail.com> wrote:
> Does this really guarantee VRAM placement? The code doesn't say anything
> about that.
>
> Marek
>
>
> On Fri, May 6, 2022 at 7:23 AM Christian König <
> ckoenig.leichtzumerken@gmail.com> wrote:
>
>> Add a AMDGPU_GEM_CREATE_DISCARDABLE flag to note that the content of a BO
>> doesn't needs to be preserved during eviction.
>>
>> KFD was already using a similar functionality for SVM BOs so replace the
>> internal flag with the new UAPI.
>>
>> Only compile tested!
>>
>> Signed-off-by: Christian König <christian.koenig@amd.com>
>> ---
>> drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 4 ++--
>> drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 1 +
>> drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 1 -
>> drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 2 +-
>> drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 2 +-
>> include/uapi/drm/amdgpu_drm.h | 4 ++++
>> 6 files changed, 9 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
>> index 2e16484bf606..bf97d8f07f57 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
>> @@ -302,8 +302,8 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev,
>> void *data,
>> AMDGPU_GEM_CREATE_VRAM_CLEARED |
>> AMDGPU_GEM_CREATE_VM_ALWAYS_VALID |
>> AMDGPU_GEM_CREATE_EXPLICIT_SYNC |
>> - AMDGPU_GEM_CREATE_ENCRYPTED))
>> -
>> + AMDGPU_GEM_CREATE_ENCRYPTED |
>> + AMDGPU_GEM_CREATE_DISCARDABLE))
>> return -EINVAL;
>>
>> /* reject invalid gem domains */
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
>> index 8b7ee1142d9a..1944ef37a61e 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
>> @@ -567,6 +567,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
>> bp->domain;
>> bo->allowed_domains = bo->preferred_domains;
>> if (bp->type != ttm_bo_type_kernel &&
>> + !(bp->flags & AMDGPU_GEM_CREATE_DISCARDABLE) &&
>> bo->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM)
>> bo->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
>> index 4c9cbdc66995..147b79c10cbb 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h
>> @@ -41,7 +41,6 @@
>>
>> /* BO flag to indicate a KFD userptr BO */
>> #define AMDGPU_AMDKFD_CREATE_USERPTR_BO (1ULL << 63)
>> -#define AMDGPU_AMDKFD_CREATE_SVM_BO (1ULL << 62)
>>
>> #define to_amdgpu_bo_user(abo) container_of((abo), struct
>> amdgpu_bo_user, bo)
>> #define to_amdgpu_bo_vm(abo) container_of((abo), struct amdgpu_bo_vm, bo)
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
>> index 41d6f604813d..ba3221a25e75 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
>> @@ -117,7 +117,7 @@ static void amdgpu_evict_flags(struct
>> ttm_buffer_object *bo,
>> }
>>
>> abo = ttm_to_amdgpu_bo(bo);
>> - if (abo->flags & AMDGPU_AMDKFD_CREATE_SVM_BO) {
>> + if (abo->flags & AMDGPU_GEM_CREATE_DISCARDABLE) {
>> placement->num_placement = 0;
>> placement->num_busy_placement = 0;
>> return;
>> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
>> b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
>> index 5ed8d9b549a4..835b5187f0b8 100644
>> --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
>> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
>> @@ -531,7 +531,7 @@ svm_range_vram_node_new(struct amdgpu_device *adev,
>> struct svm_range *prange,
>> bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
>> bp.flags = AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
>> bp.flags |= clear ? AMDGPU_GEM_CREATE_VRAM_CLEARED : 0;
>> - bp.flags |= AMDGPU_AMDKFD_CREATE_SVM_BO;
>> + bp.flags |= AMDGPU_GEM_CREATE_DISCARDABLE;
>> bp.type = ttm_bo_type_device;
>> bp.resv = NULL;
>>
>> diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
>> index 9a1d210d135d..57b9d8f0133a 100644
>> --- a/include/uapi/drm/amdgpu_drm.h
>> +++ b/include/uapi/drm/amdgpu_drm.h
>> @@ -140,6 +140,10 @@ extern "C" {
>> * not require GTT memory accounting
>> */
>> #define AMDGPU_GEM_CREATE_PREEMPTIBLE (1 << 11)
>> +/* Flag that BO can be discarded under memory pressure without keeping
>> the
>> + * content.
>> + */
>> +#define AMDGPU_GEM_CREATE_DISCARDABLE (1 << 12)
>>
>> struct drm_amdgpu_gem_create_in {
>> /** the requested memory size */
>> --
>> 2.25.1
>>
>>
[-- Attachment #2: Type: text/html, Size: 6288 bytes --]
next prev parent reply other threads:[~2022-05-10 20:44 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-06 11:23 [PATCH 1/3] drm/amdgpu: add AMDGPU_GEM_CREATE_DISCARDABLE Christian König
2022-05-06 11:23 ` [PATCH 2/3] drm/amdgpu: add AMDGPU_VM_NOALLOC Christian König
2022-05-10 21:21 ` Marek Olšák
2022-05-11 6:06 ` Christian König
2022-05-11 6:15 ` Lazar, Lijo
2022-05-11 7:22 ` Marek Olšák
2022-05-11 7:43 ` Christian König
2022-05-11 18:55 ` Marek Olšák
2022-05-16 11:06 ` Marek Olšák
2022-05-16 11:10 ` Marek Olšák
2022-05-16 11:53 ` Christian König
2022-05-16 12:56 ` Marek Olšák
2022-05-16 16:13 ` Christian König
2022-05-17 0:12 ` Marek Olšák
2022-05-17 6:33 ` Christian König
2022-05-18 8:28 ` Christian König
2022-05-06 11:23 ` [PATCH 3/3] drm/amdgpu: bump minor version number Christian König
2022-05-06 13:34 ` Alex Deucher
2022-05-12 2:38 ` Marek Olšák
2022-05-06 15:04 ` [PATCH 1/3] drm/amdgpu: add AMDGPU_GEM_CREATE_DISCARDABLE Felix Kuehling
2022-05-10 20:13 ` Marek Olšák
2022-05-10 20:43 ` Marek Olšák [this message]
2022-05-11 6:04 ` Christian König
2022-05-11 7:08 ` Marek Olšák
2022-05-11 21:58 ` Marek Olšák
2022-05-11 22:06 ` Marek Olšák
2022-05-12 7:25 ` Christian König
2022-05-12 22:17 ` Marek Olšák
2022-05-13 11:24 ` Christian König
[not found] ` <62165c7c-892a-5b35-79dd-b90414ccb5cd@damsy.net>
2022-05-13 11:26 ` Christian König
2022-07-08 14:58 ` Marek Olšák
2022-07-11 10:15 ` Christian König
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=CAAxE2A7wOfoWWh5VUFmnhyhNeCQ086trJR2BgT+nAmsYZJTbVg@mail.gmail.com \
--to=maraeo@gmail.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=ckoenig.leichtzumerken@gmail.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).