From: Felix Kuehling <Felix.Kuehling@amd.com>
To: amd-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Cc: alex.sierra@amd.com, Philip Yang <Philip.Yang@amd.com>
Subject: [PATCH 28/35] drm/amdkfd: add svm_bo eviction mechanism support
Date: Wed, 6 Jan 2021 22:01:20 -0500 [thread overview]
Message-ID: <20210107030127.20393-29-Felix.Kuehling@amd.com> (raw)
In-Reply-To: <20210107030127.20393-1-Felix.Kuehling@amd.com>
From: Alex Sierra <alex.sierra@amd.com>
svm_bo eviction mechanism is different from regular BOs.
Every SVM_BO created contains one eviction fence and one
worker item for eviction process.
SVM_BOs can be attached to one or more pranges.
For SVM_BO eviction mechanism, TTM will start to call
enable_signal callback for every SVM_BO until VRAM space
is available.
Here, all the ttm_evict calls are synchronous, this guarantees
that each eviction has completed and the fence has signaled before
it returns.
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 197 ++++++++++++++++++++-------
drivers/gpu/drm/amd/amdkfd/kfd_svm.h | 13 +-
2 files changed, 160 insertions(+), 50 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 7346255f7c27..63b745a06740 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -34,6 +34,7 @@
#define AMDGPU_SVM_RANGE_RESTORE_DELAY_MS 1
+static void svm_range_evict_svm_bo_worker(struct work_struct *work);
/**
* svm_range_unlink - unlink svm_range from lists and interval tree
* @prange: svm range structure to be removed
@@ -260,7 +261,15 @@ static void svm_range_bo_release(struct kref *kref)
list_del_init(&prange->svm_bo_list);
}
spin_unlock(&svm_bo->list_lock);
-
+ if (!dma_fence_is_signaled(&svm_bo->eviction_fence->base)) {
+ /* We're not in the eviction worker.
+ * Signal the fence and synchronize with any
+ * pending eviction work.
+ */
+ dma_fence_signal(&svm_bo->eviction_fence->base);
+ cancel_work_sync(&svm_bo->eviction_work);
+ }
+ dma_fence_put(&svm_bo->eviction_fence->base);
amdgpu_bo_unref(&svm_bo->bo);
kfree(svm_bo);
}
@@ -273,6 +282,62 @@ static void svm_range_bo_unref(struct svm_range_bo *svm_bo)
kref_put(&svm_bo->kref, svm_range_bo_release);
}
+static bool svm_range_validate_svm_bo(struct svm_range *prange)
+{
+ spin_lock(&prange->svm_bo_lock);
+ if (!prange->svm_bo) {
+ spin_unlock(&prange->svm_bo_lock);
+ return false;
+ }
+ if (prange->mm_nodes) {
+ /* We still have a reference, all is well */
+ spin_unlock(&prange->svm_bo_lock);
+ return true;
+ }
+ if (svm_bo_ref_unless_zero(prange->svm_bo)) {
+ if (READ_ONCE(prange->svm_bo->evicting)) {
+ struct dma_fence *f;
+ struct svm_range_bo *svm_bo;
+ /* The BO is getting evicted,
+ * we need to get a new one
+ */
+ spin_unlock(&prange->svm_bo_lock);
+ svm_bo = prange->svm_bo;
+ f = dma_fence_get(&svm_bo->eviction_fence->base);
+ svm_range_bo_unref(prange->svm_bo);
+ /* wait for the fence to avoid long spin-loop
+ * at list_empty_careful
+ */
+ dma_fence_wait(f, false);
+ dma_fence_put(f);
+ } else {
+ /* The BO was still around and we got
+ * a new reference to it
+ */
+ spin_unlock(&prange->svm_bo_lock);
+ pr_debug("reuse old bo svms 0x%p [0x%lx 0x%lx]\n",
+ prange->svms, prange->it_node.start,
+ prange->it_node.last);
+
+ prange->mm_nodes = prange->svm_bo->bo->tbo.mem.mm_node;
+ return true;
+ }
+
+ } else {
+ spin_unlock(&prange->svm_bo_lock);
+ }
+
+ /* We need a new svm_bo. Spin-loop to wait for concurrent
+ * svm_range_bo_release to finish removing this range from
+ * its range list. After this, it is safe to reuse the
+ * svm_bo pointer and svm_bo_list head.
+ */
+ while (!list_empty_careful(&prange->svm_bo_list))
+ ;
+
+ return false;
+}
+
static struct svm_range_bo *svm_range_bo_new(void)
{
struct svm_range_bo *svm_bo;
@@ -292,71 +357,54 @@ int
svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
bool clear)
{
- struct amdkfd_process_info *process_info;
struct amdgpu_bo_param bp;
struct svm_range_bo *svm_bo;
struct amdgpu_bo *bo;
struct kfd_process *p;
+ struct mm_struct *mm;
int r;
- pr_debug("[0x%lx 0x%lx]\n", prange->it_node.start,
- prange->it_node.last);
- spin_lock(&prange->svm_bo_lock);
- if (prange->svm_bo) {
- if (prange->mm_nodes) {
- /* We still have a reference, all is well */
- spin_unlock(&prange->svm_bo_lock);
- return 0;
- }
- if (svm_bo_ref_unless_zero(prange->svm_bo)) {
- /* The BO was still around and we got
- * a new reference to it
- */
- spin_unlock(&prange->svm_bo_lock);
- pr_debug("reuse old bo [0x%lx 0x%lx]\n",
- prange->it_node.start, prange->it_node.last);
-
- prange->mm_nodes = prange->svm_bo->bo->tbo.mem.mm_node;
- return 0;
- }
-
- spin_unlock(&prange->svm_bo_lock);
-
- /* We need a new svm_bo. Spin-loop to wait for concurrent
- * svm_range_bo_release to finish removing this range from
- * its range list. After this, it is safe to reuse the
- * svm_bo pointer and svm_bo_list head.
- */
- while (!list_empty_careful(&prange->svm_bo_list))
- ;
+ p = container_of(prange->svms, struct kfd_process, svms);
+ pr_debug("pasid: %x svms 0x%p [0x%lx 0x%lx]\n", p->pasid, prange->svms,
+ prange->it_node.start, prange->it_node.last);
- } else {
- spin_unlock(&prange->svm_bo_lock);
- }
+ if (svm_range_validate_svm_bo(prange))
+ return 0;
svm_bo = svm_range_bo_new();
if (!svm_bo) {
pr_debug("failed to alloc svm bo\n");
return -ENOMEM;
}
-
+ mm = get_task_mm(p->lead_thread);
+ if (!mm) {
+ pr_debug("failed to get mm\n");
+ kfree(svm_bo);
+ return -ESRCH;
+ }
+ svm_bo->svms = prange->svms;
+ svm_bo->eviction_fence =
+ amdgpu_amdkfd_fence_create(dma_fence_context_alloc(1),
+ mm,
+ svm_bo);
+ mmput(mm);
+ INIT_WORK(&svm_bo->eviction_work, svm_range_evict_svm_bo_worker);
+ svm_bo->evicting = 0;
memset(&bp, 0, sizeof(bp));
bp.size = prange->npages * PAGE_SIZE;
bp.byte_align = PAGE_SIZE;
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.type = ttm_bo_type_device;
bp.resv = NULL;
r = amdgpu_bo_create(adev, &bp, &bo);
if (r) {
pr_debug("failed %d to create bo\n", r);
- kfree(svm_bo);
- return r;
+ goto create_bo_failed;
}
-
- p = container_of(prange->svms, struct kfd_process, svms);
r = amdgpu_bo_reserve(bo, true);
if (r) {
pr_debug("failed %d to reserve bo\n", r);
@@ -369,8 +417,7 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
amdgpu_bo_unreserve(bo);
goto reserve_bo_failed;
}
- process_info = p->kgd_process_info;
- amdgpu_bo_fence(bo, &process_info->eviction_fence->base, true);
+ amdgpu_bo_fence(bo, &svm_bo->eviction_fence->base, true);
amdgpu_bo_unreserve(bo);
@@ -380,14 +427,16 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
prange->offset = 0;
spin_lock(&svm_bo->list_lock);
- list_add(&prange->svm_bo_list, &svm_bo->range_list);
+ list_add_rcu(&prange->svm_bo_list, &svm_bo->range_list);
spin_unlock(&svm_bo->list_lock);
return 0;
reserve_bo_failed:
- kfree(svm_bo);
amdgpu_bo_unref(&bo);
+create_bo_failed:
+ dma_fence_put(&svm_bo->eviction_fence->base);
+ kfree(svm_bo);
prange->mm_nodes = NULL;
return r;
@@ -621,7 +670,7 @@ svm_range_split_nodes(struct svm_range *new, struct svm_range *old,
new->mm_nodes = old->mm_nodes;
spin_lock(&new->svm_bo->list_lock);
- list_add(&new->svm_bo_list, &new->svm_bo->range_list);
+ list_add_rcu(&new->svm_bo_list, &new->svm_bo->range_list);
spin_unlock(&new->svm_bo->list_lock);
return 0;
@@ -1353,7 +1402,7 @@ struct svm_range *svm_range_clone(struct svm_range *old)
new->offset = old->offset;
new->svm_bo = svm_range_bo_ref(old->svm_bo);
spin_lock(&new->svm_bo->list_lock);
- list_add(&new->svm_bo_list, &new->svm_bo->range_list);
+ list_add_rcu(&new->svm_bo_list, &new->svm_bo->range_list);
spin_unlock(&new->svm_bo->list_lock);
}
new->flags = old->flags;
@@ -1964,6 +2013,62 @@ svm_range_trigger_migration(struct mm_struct *mm, struct svm_range *prange,
return r;
}
+int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence)
+{
+ if (!fence)
+ return -EINVAL;
+
+ if (dma_fence_is_signaled(&fence->base))
+ return 0;
+
+ if (fence->svm_bo) {
+ WRITE_ONCE(fence->svm_bo->evicting, 1);
+ schedule_work(&fence->svm_bo->eviction_work);
+ }
+
+ return 0;
+}
+
+static void svm_range_evict_svm_bo_worker(struct work_struct *work)
+{
+ struct svm_range_bo *svm_bo;
+ struct svm_range *prange;
+ struct kfd_process *p;
+ struct mm_struct *mm;
+ int srcu_idx;
+
+ svm_bo = container_of(work, struct svm_range_bo, eviction_work);
+ if (!svm_bo_ref_unless_zero(svm_bo))
+ return; /* svm_bo was freed while eviction was pending */
+
+ /* svm_range_bo_release destroys this worker thread. So during
+ * the lifetime of this thread, kfd_process and mm will be valid.
+ */
+ p = container_of(svm_bo->svms, struct kfd_process, svms);
+ mm = p->mm;
+ if (!mm)
+ return;
+
+ mmap_read_lock(mm);
+ srcu_idx = srcu_read_lock(&svm_bo->svms->srcu);
+ list_for_each_entry_rcu(prange, &svm_bo->range_list, svm_bo_list) {
+ pr_debug("svms 0x%p [0x%lx 0x%lx]\n", prange->svms,
+ prange->it_node.start, prange->it_node.last);
+ mutex_lock(&prange->mutex);
+ svm_migrate_vram_to_ram(prange, svm_bo->eviction_fence->mm);
+ mutex_unlock(&prange->mutex);
+ }
+ srcu_read_unlock(&svm_bo->svms->srcu, srcu_idx);
+ mmap_read_unlock(mm);
+
+ dma_fence_signal(&svm_bo->eviction_fence->base);
+ /* This is the last reference to svm_bo, after svm_range_vram_node_free
+ * has been called in svm_migrate_vram_to_ram
+ */
+ WARN_ONCE(kref_read(&svm_bo->kref) != 1, "This was not the last reference\n");
+ svm_range_bo_unref(svm_bo);
+}
+
static int
svm_range_set_attr(struct kfd_process *p, uint64_t start, uint64_t size,
uint32_t nattr, struct kfd_ioctl_svm_attribute *attrs)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
index e546f36ef709..143573621956 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.h
@@ -33,10 +33,14 @@
#include "kfd_priv.h"
struct svm_range_bo {
- struct amdgpu_bo *bo;
- struct kref kref;
- struct list_head range_list; /* all svm ranges shared this bo */
- spinlock_t list_lock;
+ struct amdgpu_bo *bo;
+ struct kref kref;
+ struct list_head range_list; /* all svm ranges shared this bo */
+ spinlock_t list_lock;
+ struct amdgpu_amdkfd_fence *eviction_fence;
+ struct work_struct eviction_work;
+ struct svm_range_list *svms;
+ uint32_t evicting;
};
/**
* struct svm_range - shared virtual memory range
@@ -123,5 +127,6 @@ int svm_range_split_by_granularity(struct kfd_process *p, unsigned long addr,
struct list_head *list);
int svm_range_restore_pages(struct amdgpu_device *adev,
unsigned int pasid, uint64_t addr);
+int svm_range_schedule_evict_svm_bo(struct amdgpu_amdkfd_fence *fence);
#endif /* KFD_SVM_H_ */
--
2.29.2
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2021-01-07 3:03 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-07 3:00 [PATCH 00/35] Add HMM-based SVM memory manager to KFD Felix Kuehling
2021-01-07 3:00 ` [PATCH 01/35] drm/amdkfd: select kernel DEVICE_PRIVATE option Felix Kuehling
2021-01-07 3:00 ` [PATCH 02/35] drm/amdgpu: replace per_device_list by array Felix Kuehling
2021-01-07 3:00 ` [PATCH 03/35] drm/amdkfd: helper to convert gpu id and idx Felix Kuehling
2021-01-07 3:00 ` [PATCH 04/35] drm/amdkfd: add svm ioctl API Felix Kuehling
2021-01-07 3:00 ` [PATCH 05/35] drm/amdkfd: Add SVM API support capability bits Felix Kuehling
2021-01-07 3:00 ` [PATCH 06/35] drm/amdkfd: register svm range Felix Kuehling
2021-01-07 3:00 ` [PATCH 07/35] drm/amdkfd: add svm ioctl GET_ATTR op Felix Kuehling
2021-01-07 3:01 ` [PATCH 08/35] drm/amdgpu: add common HMM get pages function Felix Kuehling
2021-01-07 10:53 ` Christian König
2021-01-07 3:01 ` [PATCH 09/35] drm/amdkfd: validate svm range system memory Felix Kuehling
2021-01-07 3:01 ` [PATCH 10/35] drm/amdkfd: register overlap system memory range Felix Kuehling
2021-01-07 3:01 ` [PATCH 11/35] drm/amdkfd: deregister svm range Felix Kuehling
2021-01-07 3:01 ` [PATCH 12/35] drm/amdgpu: export vm update mapping interface Felix Kuehling
2021-01-07 10:54 ` Christian König
2021-01-07 3:01 ` [PATCH 13/35] drm/amdkfd: map svm range to GPUs Felix Kuehling
2021-01-07 3:01 ` [PATCH 14/35] drm/amdkfd: svm range eviction and restore Felix Kuehling
2021-01-07 3:01 ` [PATCH 15/35] drm/amdkfd: add xnack enabled flag to kfd_process Felix Kuehling
2021-01-07 3:01 ` [PATCH 16/35] drm/amdkfd: add ioctl to configure and query xnack retries Felix Kuehling
2021-01-07 3:01 ` [PATCH 17/35] drm/amdkfd: register HMM device private zone Felix Kuehling
2021-03-01 8:32 ` Daniel Vetter
2021-03-01 8:46 ` Thomas Hellström (Intel)
2021-03-01 8:58 ` Daniel Vetter
2021-03-01 9:30 ` Thomas Hellström (Intel)
2021-03-04 17:58 ` Felix Kuehling
2021-03-11 12:24 ` Thomas Hellström (Intel)
2021-01-07 3:01 ` [PATCH 18/35] drm/amdkfd: validate vram svm range from TTM Felix Kuehling
2021-01-07 3:01 ` [PATCH 19/35] drm/amdkfd: support xgmi same hive mapping Felix Kuehling
2021-01-07 3:01 ` [PATCH 20/35] drm/amdkfd: copy memory through gart table Felix Kuehling
2021-01-07 3:01 ` [PATCH 21/35] drm/amdkfd: HMM migrate ram to vram Felix Kuehling
2021-01-07 3:01 ` [PATCH 22/35] drm/amdkfd: HMM migrate vram to ram Felix Kuehling
2021-01-07 3:01 ` [PATCH 23/35] drm/amdkfd: invalidate tables on page retry fault Felix Kuehling
2021-01-07 3:01 ` [PATCH 24/35] drm/amdkfd: page table restore through svm API Felix Kuehling
2021-01-07 3:01 ` [PATCH 25/35] drm/amdkfd: SVM API call to restore page tables Felix Kuehling
2021-01-07 3:01 ` [PATCH 26/35] drm/amdkfd: add svm_bo reference for eviction fence Felix Kuehling
2021-01-07 3:01 ` [PATCH 27/35] drm/amdgpu: add param bit flag to create SVM BOs Felix Kuehling
2021-01-07 3:01 ` Felix Kuehling [this message]
2021-01-07 3:01 ` [PATCH 29/35] drm/amdgpu: svm bo enable_signal call condition Felix Kuehling
2021-01-07 10:56 ` Christian König
2021-01-07 16:16 ` Felix Kuehling
2021-01-07 16:28 ` Christian König
2021-01-07 16:53 ` Felix Kuehling
2021-01-07 3:01 ` [PATCH 30/35] drm/amdgpu: add svm_bo eviction to enable_signal cb Felix Kuehling
2021-01-07 3:01 ` [PATCH 31/35] drm/amdgpu: reserve fence slot to update page table Felix Kuehling
2021-01-07 10:57 ` Christian König
2021-01-07 3:01 ` [PATCH 32/35] drm/amdgpu: enable retry fault wptr overflow Felix Kuehling
2021-01-07 11:01 ` Christian König
2021-01-07 3:01 ` [PATCH 33/35] drm/amdkfd: refine migration policy with xnack on Felix Kuehling
2021-01-07 3:01 ` [PATCH 34/35] drm/amdkfd: add svm range validate timestamp Felix Kuehling
2021-01-07 3:01 ` [PATCH 35/35] drm/amdkfd: multiple gpu migrate vram to vram Felix Kuehling
2021-01-07 9:23 ` [PATCH 00/35] Add HMM-based SVM memory manager to KFD Daniel Vetter
2021-01-07 16:25 ` Felix Kuehling
2021-01-08 14:40 ` Daniel Vetter
2021-01-08 14:45 ` Christian König
2021-01-08 15:58 ` Felix Kuehling
2021-01-08 16:06 ` Daniel Vetter
2021-01-08 16:36 ` Felix Kuehling
2021-01-08 16:53 ` Daniel Vetter
2021-01-08 17:56 ` Felix Kuehling
2021-01-11 16:29 ` Daniel Vetter
2021-01-14 5:34 ` Felix Kuehling
2021-01-14 12:19 ` Christian König
2021-01-13 16:56 ` Jerome Glisse
2021-01-13 20:31 ` Daniel Vetter
2021-01-14 3:27 ` Jerome Glisse
2021-01-14 9:26 ` Daniel Vetter
2021-01-14 10:39 ` Daniel Vetter
2021-01-14 10:49 ` Christian König
2021-01-14 11:52 ` Daniel Vetter
2021-01-14 13:37 ` HMM fence (was Re: [PATCH 00/35] Add HMM-based SVM memory manager to KFD) Christian König
2021-01-14 13:57 ` Daniel Vetter
2021-01-14 14:13 ` Christian König
2021-01-14 14:23 ` Daniel Vetter
2021-01-14 15:08 ` Christian König
2021-01-14 15:40 ` Daniel Vetter
2021-01-14 16:01 ` Christian König
2021-01-14 16:36 ` Daniel Vetter
2021-01-14 19:08 ` Christian König
2021-01-14 20:09 ` Daniel Vetter
2021-01-14 16:51 ` Jerome Glisse
2021-01-14 21:13 ` Felix Kuehling
2021-01-15 7:47 ` Christian König
2021-01-13 16:47 ` [PATCH 00/35] Add HMM-based SVM memory manager to KFD Jerome Glisse
2021-01-14 0:06 ` Felix Kuehling
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=20210107030127.20393-29-Felix.Kuehling@amd.com \
--to=felix.kuehling@amd.com \
--cc=Philip.Yang@amd.com \
--cc=alex.sierra@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=dri-devel@lists.freedesktop.org \
/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).