[PATCH v3 2/3] drm/xe/sriov: Shifting GGTT area post migration
Lis, Tomasz
tomasz.lis at intel.com
Tue Dec 3 18:22:03 UTC 2024
On 27.11.2024 23:30, Michal Wajdeczko wrote:
> On 23.11.2024 04:13, Tomasz Lis wrote:
>> We have only one GGTT for all IOV functions, with each VF having assigned
>> a range of addresses for its use. After migration, a VF can receive a
>> different range of addresses than it had initially.
>>
>> This implements shifting GGTT addresses within drm_mm nodes, so that
>> VMAs stay valid after migration. This will make the driver use new
>> addresses when accessing GGTT from the moment the shifting ends.
>>
>> By taking the ggtt->lock for the period of VMA fixups, this change
>> also adds constaint on that mutex. Any locks used during the recovery
> typo: constraint ?
ack
>> cannot ever wait for hardware response - because after migration,
>> the hardware will not do anything until fixups are finished.
>>
>> v2: Moved some functs to xe_ggtt.c; moved shift computation to just
>> after querying; improved documentation; switched some warns to asserts;
>> skipping fixups when GGTT shift eq 0; iterating through tiles (Michal)
>>
>> Signed-off-by: Tomasz Lis<tomasz.lis at intel.com>
>> ---
>> drivers/gpu/drm/xe/xe_ggtt.c | 139 ++++++++++++++++++++++
>> drivers/gpu/drm/xe/xe_ggtt.h | 3 +
>> drivers/gpu/drm/xe/xe_gt_sriov_vf.c | 25 ++++
>> drivers/gpu/drm/xe/xe_gt_sriov_vf.h | 1 +
>> drivers/gpu/drm/xe/xe_gt_sriov_vf_types.h | 2 +
>> drivers/gpu/drm/xe/xe_sriov_vf.c | 22 ++++
>> 6 files changed, 192 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c
>> index 558fac8bb6fb..44500707f8ae 100644
>> --- a/drivers/gpu/drm/xe/xe_ggtt.c
>> +++ b/drivers/gpu/drm/xe/xe_ggtt.c
>> @@ -489,6 +489,145 @@ void xe_ggtt_node_remove_balloon(struct xe_ggtt_node *node)
>> xe_ggtt_node_fini(node);
>> }
>>
>> +static u64 drm_mm_node_end(struct drm_mm_node *node)
>> +{
>> + return node->start + node->size;
>> +}
>> +
>> +u64 xe_ggtt_node_end(struct xe_ggtt_node *node)
> please add proper kernel-doc
will remove instead.
>> +{
>> + return drm_mm_node_end(&node->base);
>> +}
>> +
>> +static void xe_ggtt_mm_shift_nodes(struct xe_ggtt *ggtt, struct drm_mm_node *balloon_beg,
>> + struct drm_mm_node *balloon_fin, s64 shift)
>> +{
>> + struct drm_mm_node *node, *tmpn;
>> + int err;
>> + LIST_HEAD(temp_list_head);
> nit: please reorder vars to follow reverse-xmas-tree style
I did commented on that request in the previous round.. but ok, can do.
>> +
>> + lockdep_assert_held(&ggtt->lock);
>> +
>> + /*
>> + * Move nodes, from range previously assigned to this VF, into temp list.
> hmm, now that function is taking generic params, maybe all this comment
> should be in the VF file as full kernel-doc explaining GGTT fixup steps?
I don't have a strong opinion on docs location. If you provide details
of your idea, I can move that.
Though this is only one of the fixups applied. I hope we're not leaning
towards making kerneldoc
chapters about all of them.
>> + *
>> + * The balloon_beg and balloon_fin nodes are there to eliminate unavailable
>> + * ranges from use: first reserves the GGTT area below the range for current VF,
>> + * and second reserves area above.
>> + *
>> + * Below is a GGTT layout of example VF, with a certain address range assigned to
>> + * said VF, and inaccessible areas above and below:
>> + *
>> + * 0 4GiB
>> + * |<--------------------------- Total GGTT size ----------------------------->|
>> + * WOPCM GUC_TOP
>> + * |<-------------- Area mappable by xe_ggtt instance ---------------->|
>> + *
>> + * +---+---------------------------------+----------+----------------------+---+
>> + * |\\\|/////////////////////////////////| VF mem |//////////////////////|\\\|
>> + * +---+---------------------------------+----------+----------------------+---+
> hmm, I'm wondering if maybe after all recent GGTT refactoring it would
> be easier to trim VF's xe_ggtt size at init() instead of playing with
> two balloons?
>
> then layout will look like:
>
> PF:
> |<---------------- xe_ggtt.size ----------------------------->|
> +---+-------------------------------------------------------------+---+
> | |/////////////////////////////////////////////////////////////| |
> +---+-------------------------------------------------------------+---+
>
>
> VF: before migration
> xe_ggtt.size |<-------->|
> +---+---------------------------+----------+----------------------+---+
> | |//////////| |
> +---+---------------------------+----------+----------------------+---+
>
> VF: after migration
> xe_ggtt.size |<-------->|
> +---+--------------------------------------------+----------+-----+---+
> | |//////////| |
> +---+--------------------------------------------+----------+-----+---+
That is a change which would require address space shifting capability
in drm_mm, and
changes outside the scope of these patches. I will try to prepare an RFC
for the drm_mm
part, then we'll see whether this is viable.
>> + *
>> + * Hardware enforced access rules before migration:
>> + *
>> + * |<------- inaccessible for VF ------->|<VF owned>|<-- inaccessible for VF ->|
>> + *
>> + * drm_mm nodes used for tracking allocations:
>> + *
>> + * |<----------- balloon ------------>|<- nodes->|<----- balloon ------>|
>> + *
>> + * After the migration, GGTT area assigned to the VF might have shifted, either
>> + * to lower or to higher address. But we expect the total size and extra areas to
>> + * be identical, as migration can only happen between matching platforms.
>> + * Below is an example of GGTT layout of the VF after migration. Content of the
>> + * GGTT for VF has been moved to a new area, and we receive its address from GuC:
>> + *
>> + * +---+----------------------+----------+---------------------------------+---+
>> + * |\\\|//////////////////////| VF mem |/////////////////////////////////|\\\|
>> + * +---+----------------------+----------+---------------------------------+---+
>> + *
>> + * Hardware enforced access rules after migration:
>> + *
>> + * |<- inaccessible for VF -->|<VF owned>|<------- inaccessible for VF ------->|
>> + *
>> + * So the VF has a new slice of GGTT assigned, and during migration process, the
>> + * memory content was copied to that new area. But the drm_mm nodes within xe kmd
>> + * are still tracking allocations using the old addresses. The nodes within VF
>> + * owned area have to be shifted, and balloon nodes need to be resized to
>> + * properly mask out areas not owned by the VF.
>> + *
>> + * Fixed drm_mm nodes used for tracking allocations:
>> + *
>> + * |<------ balloon ------>|<- nodes->|<----------- balloon ----------->|
>> + *
>> + * Due to use of GPU profiles, we do not expect the old and new GGTT ares to
>> + * overlap; but our node shifting will fix addresses properly regardless.
> whether we use profile or not, we shouldn't make any assumptions about
> overlap, so maybe just skip that comment?
The profiles are part of SRIOV architecture. Without a comment, it
should be assumed that the code
is only compliant with use cases from said architecture.
>> + *
>> + */
>> + drm_mm_for_each_node_in_range_safe(node, tmpn, &ggtt->mm,
>> + drm_mm_node_end(balloon_beg),
>> + balloon_fin->start) {
>> + drm_mm_remove_node(node);
>> + list_add(&node->node_list, &temp_list_head);
>> + }
>> +
>> + /* shift and re-add ballooning nodes */
>> + if (drm_mm_node_allocated(balloon_beg))
>> + drm_mm_remove_node(balloon_beg);
>> + if (drm_mm_node_allocated(balloon_fin))
>> + drm_mm_remove_node(balloon_fin);
>> + balloon_beg->size += shift;
>> + balloon_fin->start += shift;
>> + balloon_fin->size -= shift;
>> + if (balloon_beg->size != 0) {
>> + err = drm_mm_reserve_node(&ggtt->mm, balloon_beg);
>> + xe_tile_assert(ggtt->tile, !err);
>> + }
>> + if (balloon_fin->size != 0) {
>> + err = drm_mm_reserve_node(&ggtt->mm, balloon_fin);
>> + xe_tile_assert(ggtt->tile, !err);
>> + }
>> +
>> + /*
>> + * Now the GGTT VM contains only nodes outside of area assigned to this VF.
>> + * We can re-add all VF nodes with shifted offsets.
>> + */
>> + list_for_each_entry_safe(node, tmpn, &temp_list_head, node_list) {
>> + list_del(&node->node_list);
>> + node->start += shift;
>> + err = drm_mm_reserve_node(&ggtt->mm, node);
>> + xe_tile_assert(ggtt->tile, !err);
>> + }
>> +}
>> +
>> +void xe_ggtt_node_shift_nodes(struct xe_ggtt *ggtt, struct xe_ggtt_node *balloon_beg,
>> + struct xe_ggtt_node *balloon_fin, s64 shift)
> don't forget about kernel-doc
will add.
>> +{
>> + struct drm_mm_node *balloon_mm_beg, *balloon_mm_end;
>> + struct drm_mm_node loc_beg, loc_end;
>> +
>> + if (balloon_beg && balloon_beg->ggtt)
>> + balloon_mm_beg = &balloon_beg->base;
>> + else {
>> + loc_beg.color = 0;
>> + loc_beg.flags = 0;
>> + loc_beg.start = xe_wopcm_size(ggtt->tile->xe);
>> + loc_beg.size = 0;
>> + balloon_mm_beg = &loc_beg;
>> + }
>> +
>> + if (balloon_fin && balloon_fin->ggtt)
>> + balloon_mm_end = &balloon_fin->base;
>> + else {
>> + loc_end.color = 0;
>> + loc_end.flags = 0;
>> + loc_end.start = GUC_GGTT_TOP;
>> + loc_end.size = 0;
>> + balloon_mm_end = &loc_end;
>> + }
>> +
>> + xe_ggtt_mm_shift_nodes(ggtt, balloon_mm_beg, balloon_mm_end, shift);
>> +}
>> +
>> /**
>> * xe_ggtt_node_insert_locked - Locked version to insert a &xe_ggtt_node into the GGTT
>> * @node: the &xe_ggtt_node to be inserted
>> diff --git a/drivers/gpu/drm/xe/xe_ggtt.h b/drivers/gpu/drm/xe/xe_ggtt.h
>> index 27e7d67de004..3d2ca6f7185d 100644
>> --- a/drivers/gpu/drm/xe/xe_ggtt.h
>> +++ b/drivers/gpu/drm/xe/xe_ggtt.h
>> @@ -18,12 +18,15 @@ void xe_ggtt_node_fini(struct xe_ggtt_node *node);
>> int xe_ggtt_node_insert_balloon(struct xe_ggtt_node *node,
>> u64 start, u64 size);
>> void xe_ggtt_node_remove_balloon(struct xe_ggtt_node *node);
>> +void xe_ggtt_node_shift_nodes(struct xe_ggtt *ggtt, struct xe_ggtt_node *balloon_beg,
>> + struct xe_ggtt_node *balloon_fin, s64 shift);
>>
>> int xe_ggtt_node_insert(struct xe_ggtt_node *node, u32 size, u32 align);
>> int xe_ggtt_node_insert_locked(struct xe_ggtt_node *node,
>> u32 size, u32 align, u32 mm_flags);
>> void xe_ggtt_node_remove(struct xe_ggtt_node *node, bool invalidate);
>> bool xe_ggtt_node_allocated(const struct xe_ggtt_node *node);
>> +u64 xe_ggtt_node_end(struct xe_ggtt_node *node);
>> void xe_ggtt_map_bo(struct xe_ggtt *ggtt, struct xe_bo *bo);
>> int xe_ggtt_insert_bo(struct xe_ggtt *ggtt, struct xe_bo *bo);
>> int xe_ggtt_insert_bo_at(struct xe_ggtt *ggtt, struct xe_bo *bo,
>> diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_vf.c b/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
>> index cca5d5732802..a1841862c97e 100644
>> --- a/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
>> +++ b/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
>> @@ -389,6 +389,7 @@ static int vf_get_ggtt_info(struct xe_gt *gt)
>> xe_gt_sriov_dbg_verbose(gt, "GGTT %#llx-%#llx = %lluK\n",
>> start, start + size - 1, size / SZ_1K);
>>
>> + config->ggtt_shift = start - (s64)config->ggtt_base;
> is this explicit cast really needed ?
"needed.."
Due to the characteristics of two's complement system, not having the
cast would lead to valid s64 result on all relevant CPUs.
But according to C specification, result of arithmetic subtraction will
be unsigned if both arguments are unsigned (see "Usual arithmetic
conversions" for details). So for logic compliance with C standard, the
cast is indeed needed.
While in all relevant CPUs the unsigned underflow works as expected, the
C spec defines it as undefined behavior in the chapter mentioned before.
>> config->ggtt_base = start;
>> config->ggtt_size = size;
>>
>> @@ -912,6 +913,30 @@ int xe_gt_sriov_vf_query_runtime(struct xe_gt *gt)
>> return err;
>> }
>>
>> +/**
>> + * - Shift GGTT allocations to match assigned range.
>> + * @gt: the &xe_gt struct instance
>> + *
>> + * Since Global GTT is not virtualized, each VF has an assigned range
>> + * within the global space. This range might have changed during migration,
>> + * which requires all memory addresses pointing to GGTT to be shifted.
> Return: ?
will add.
>> + */
>> +int xe_gt_sriov_vf_fixup_ggtt_nodes(struct xe_gt *gt)
>> +{
>> + struct xe_gt_sriov_vf_selfconfig *config = >->sriov.vf.self_config;
>> + struct xe_tile *tile = gt_to_tile(gt);
>> + struct xe_ggtt *ggtt = tile->mem.ggtt;
>> + s64 ggtt_shift;
>> +
>> + mutex_lock(&ggtt->lock);
>> + ggtt_shift = config->ggtt_shift;
>> + if (ggtt_shift)
>> + xe_ggtt_node_shift_nodes(ggtt, tile->sriov.vf.ggtt_balloon[0],
>> + tile->sriov.vf.ggtt_balloon[1], ggtt_shift);
>> + mutex_unlock(&ggtt->lock);
>> + return ggtt_shift ? 0 : ENODATA;
>> +}
>> +
>> static int vf_runtime_reg_cmp(const void *a, const void *b)
>> {
>> const struct vf_runtime_reg *ra = a;
>> diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_vf.h b/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
>> index 912d20814261..49af35484a57 100644
>> --- a/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
>> +++ b/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
>> @@ -17,6 +17,7 @@ int xe_gt_sriov_vf_query_config(struct xe_gt *gt);
>> int xe_gt_sriov_vf_connect(struct xe_gt *gt);
>> int xe_gt_sriov_vf_query_runtime(struct xe_gt *gt);
>> int xe_gt_sriov_vf_prepare_ggtt(struct xe_gt *gt);
>> +int xe_gt_sriov_vf_fixup_ggtt_nodes(struct xe_gt *gt);
>> int xe_gt_sriov_vf_notify_resfix_done(struct xe_gt *gt);
>> void xe_gt_sriov_vf_migrated_event_handler(struct xe_gt *gt);
>>
>> diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_vf_types.h b/drivers/gpu/drm/xe/xe_gt_sriov_vf_types.h
>> index a57f13b5afcd..5ccbdf8d08b6 100644
>> --- a/drivers/gpu/drm/xe/xe_gt_sriov_vf_types.h
>> +++ b/drivers/gpu/drm/xe/xe_gt_sriov_vf_types.h
>> @@ -40,6 +40,8 @@ struct xe_gt_sriov_vf_selfconfig {
>> u64 ggtt_base;
>> /** @ggtt_size: assigned size of the GGTT region. */
>> u64 ggtt_size;
>> + /** @ggtt_shift: difference in ggtt_base on last migration */
>> + s64 ggtt_shift;
>> /** @lmem_size: assigned size of the LMEM. */
>> u64 lmem_size;
>> /** @num_ctxs: assigned number of GuC submission context IDs. */
>> diff --git a/drivers/gpu/drm/xe/xe_sriov_vf.c b/drivers/gpu/drm/xe/xe_sriov_vf.c
>> index c1275e64aa9c..4ee8fc70a744 100644
>> --- a/drivers/gpu/drm/xe/xe_sriov_vf.c
>> +++ b/drivers/gpu/drm/xe/xe_sriov_vf.c
>> @@ -7,6 +7,7 @@
>>
>> #include "xe_assert.h"
>> #include "xe_device.h"
>> +#include "xe_gt.h"
>> #include "xe_gt_sriov_printk.h"
>> #include "xe_gt_sriov_vf.h"
>> #include "xe_pm.h"
>> @@ -170,6 +171,26 @@ static bool vf_post_migration_imminent(struct xe_device *xe)
>> work_pending(&xe->sriov.vf.migration.worker);
>> }
>>
>> +static int vf_post_migration_fixup_ggtt_nodes(struct xe_device *xe)
>> +{
>> + struct xe_tile *tile;
>> + unsigned int id;
>> + int err;
>> +
>> + for_each_tile(tile, xe, id) {
>> + struct xe_gt *gt = tile->primary_gt;
>> + int ret;
>> +
>> + /* media doesn't have its own ggtt */
>> + if (xe_gt_is_media_type(gt))
>> + continue;
>> + ret = xe_gt_sriov_vf_fixup_ggtt_nodes(gt);
>> + if (ret != ENODATA)
>> + err = ret;
>> + }
>> + return err;
>> +}
>> +
>> /*
>> * Notify all GuCs about resource fixups apply finished.
>> */
>> @@ -201,6 +222,7 @@ static void vf_post_migration_recovery(struct xe_device *xe)
>> if (unlikely(err))
>> goto fail;
>>
>> + err = vf_post_migration_fixup_ggtt_nodes(xe);
>> /* FIXME: add the recovery steps */
> nit: ... add *more* recovery steps ... ?
the idea of this comment is to alter the code around it, not to continue
adding only slightly relevant fixes to the temporary comment itself.
-Tomasz
>> vf_post_migration_notify_resfix_done(xe);
>> xe_pm_runtime_put(xe);
More information about the Intel-xe
mailing list