Skip to contentSkip to conversation

tools/sched_ext: scx_qmap: Silence task_ctx lookup miss

PATCH Superseded Tejun Heo 3 participants 5 messages Apr 21 Scheduler - sched_ext

lore
Share & tools
lore

Tejun Heo

author reviewer
#0 lore raw

PATCH 1/1 tools/sched_ext: scx_qmap: Silence task_ctx lookup miss

scx_fork() dispatches ops.init_task to exactly one scheduler - the one owning the forking task's cgroup. A task forked inside a sub-scheduler's cgroup is init'd into the sub only; the root scheduler has no task_ctx entry for it. When that task later appears as @prev in the root's qmap_dispatch() (or flows through core-sched comparison via task_qdist), the bpf_task_storage_get() legitimately misses.

qmap treated those misses as fatal via scx_bpf_error("task_ctx lookup failed") and aborted the scheduler as soon as the first cross-sched task hit the root. Drop the error in the three sites where the miss is
legitimate: lookup_task_ctx() (helper; callers already check for NULL), qmap_dispatch()'s @prev branch (bookkeeping-only), and task_qdist() (returns 0 which makes the comparison a no-op). The existing scx_error was a paranoid guard from the pre-sub-sched world where every task was owned by the one and only scheduler.

Fixes 4f8b122848db ("sched_ext: Add basic building blocks for nested sub-scheduler dispatching")
---
 tools/sched_ext/scx_qmap.bpf.c | 22 +++++-----------------
 1 file changed, 5 insertions(+), 17 deletions(-)
tools/sched_ext/scx_qmap.bpf.c +5 -17
index b68abb9e760b..0330ba15210d 100644
@@ -159,13 +159,7 @@ static s32 pick_direct_dispatch_cpu(struct task_struct *p, s32 prev_cpu)
 
 static struct task_ctx *lookup_task_ctx(struct task_struct *p)
 {
-	struct task_ctx *tctx;
-
-	if (!(tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0))) {
-		scx_bpf_error("task_ctx lookup failed");
-		return NULL;
-	}
-	return tctx;
+	return bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
 }
 
 s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
@@ -540,13 +534,9 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
 	 */
 	if (prev) {
 		tctx = bpf_task_storage_get(&task_ctx_stor, prev, 0, 0);
-		if (!tctx) {
-			scx_bpf_error("task_ctx lookup failed");
-			return;
-		}
-
-		tctx->core_sched_seq =
-			core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
+		if (tctx)
+			tctx->core_sched_seq =
+				core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
 	}
 }
 
@@ -584,10 +574,8 @@ static s64 task_qdist(struct task_struct *p)
 	s64 qdist;
 
 	tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
-	if (!tctx) {
-		scx_bpf_error("task_ctx lookup failed");
+	if (!tctx)
 		return 0;
-	}
 
 	qdist = tctx->core_sched_seq - core_sched_head_seqs[idx];
 

--
2.53.0

reviewer Reviewed-by on v1 1/1
#1 lore raw

Hi Tejun,

Quoted 16 lines · Tejun Heo

scx_fork() dispatches ops.init_task to exactly one scheduler - the one owning the forking task's cgroup. A task forked inside a sub-scheduler's cgroup is init'd into the sub only; the root scheduler has no task_ctx entry for it. When that task later appears as @prev in the root's qmap_dispatch() (or flows through core-sched comparison via task_qdist), the bpf_task_storage_get() legitimately misses.

qmap treated those misses as fatal via scx_bpf_error("task_ctx lookup failed") and aborted the scheduler as soon as the first cross-sched task hit the root. Drop the error in the three sites where the miss is
legitimate: lookup_task_ctx() (helper; callers already check for NULL), qmap_dispatch()'s @prev branch (bookkeeping-only), and task_qdist() (returns 0 which makes the comparison a no-op). The existing scx_error was a paranoid guard from the pre-sub-sched world where every task was owned by the one and only scheduler.

Quoted 2 lines
Fixes 4f8b122848db ("sched_ext: Add basic building blocks for nested sub-scheduler dispatching")

Should we return prev_cpu in qmap_select_cpu() instead of -ESRCH when lookup_task_ctx() returns NULL? Otherwise the scheduler would error.

Apart than that looks good to me.

Thanks,
-Andrea

Quoted 54 lines
---
 tools/sched_ext/scx_qmap.bpf.c | 22 +++++-----------------
 1 file changed, 5 insertions(+), 17 deletions(-)
tools/sched_ext/scx_qmap.bpf.c +5 -18
index b68abb9e760b..0330ba15210d 100644
@@ -159,13 +159,7 @@ static s32 pick_direct_dispatch_cpu(struct task_struct *p, s32 prev_cpu)
 
 static struct task_ctx *lookup_task_ctx(struct task_struct *p)
 {
-	struct task_ctx *tctx;
-
-	if (!(tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0))) {
-		scx_bpf_error("task_ctx lookup failed");
-		return NULL;
-	}
-	return tctx;
+	return bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
 }
 
 s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
@@ -540,13 +534,9 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
 	 */
 	if (prev) {
 		tctx = bpf_task_storage_get(&task_ctx_stor, prev, 0, 0);
-		if (!tctx) {
-			scx_bpf_error("task_ctx lookup failed");
-			return;
-		}
-
-		tctx->core_sched_seq =
-			core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
+		if (tctx)
+			tctx->core_sched_seq =
+				core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
 	}
 }
 
@@ -584,10 +574,8 @@ static s64 task_qdist(struct task_struct *p)
 	s64 qdist;
 
 	tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
-	if (!tctx) {
-		scx_bpf_error("task_ctx lookup failed");
+	if (!tctx)
 		return 0;
-	}
 
 	qdist = tctx->core_sched_seq - core_sched_head_seqs[idx];
 
--

2.53.0

author reviewer
#2 lore raw

PATCH v2 1/1 tools/sched_ext: scx_qmap: Silence task_ctx lookup miss

In reply to Tejun Heo

scx_fork() dispatches ops.init_task to exactly one scheduler - the one owning the forking task's cgroup. A task forked inside a sub-scheduler's cgroup is init'd into the sub only; the root scheduler has no task_ctx entry for it. When that task later appears as @prev in the root's qmap_dispatch() (or flows through core-sched comparison via task_qdist), the bpf_task_storage_get() legitimately misses.

qmap treated those misses as fatal via scx_bpf_error("task_ctx lookup failed") and aborted the scheduler as soon as the first cross-sched task hit the root. Drop the error in the sites where the miss is
legitimate: lookup_task_ctx() (helper; callers already check for NULL), qmap_dispatch()'s @prev branch (bookkeeping-only), task_qdist() (returns 0 which makes the comparison a no-op), and qmap_select_cpu() (returns prev_cpu as a no-op fallback instead of -ESRCH). The existing scx_error was a paranoid guard from the pre-sub-sched world where every task was owned by the one and only scheduler.

v2: qmap_select_cpu() returns prev_cpu on NULL instead of -ESRCH, so
the root scheduler doesn't error on cross-sched tasks that pass through it (Andrea Righi).

Fixes 4f8b122848db ("sched_ext: Add basic building blocks for nested sub-scheduler dispatching")
---
 tools/sched_ext/scx_qmap.bpf.c | 24 ++++++------------------
 1 file changed, 6 insertions(+), 18 deletions(-)
tools/sched_ext/scx_qmap.bpf.c +6 -18
index b68abb9e760b..aad698fe294b 100644
@@ -159,13 +159,7 @@ static s32 pick_direct_dispatch_cpu(struct task_struct *p, s32 prev_cpu)
 
 static struct task_ctx *lookup_task_ctx(struct task_struct *p)
 {
-	struct task_ctx *tctx;
-
-	if (!(tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0))) {
-		scx_bpf_error("task_ctx lookup failed");
-		return NULL;
-	}
-	return tctx;
+	return bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
 }
 
 s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
@@ -175,7 +169,7 @@ s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
 	s32 cpu;
 
 	if (!(tctx = lookup_task_ctx(p)))
-		return -ESRCH;
+		return prev_cpu;
 
 	if (p->scx.weight < 2 && !(p->flags & PF_KTHREAD))
 		return prev_cpu;
@@ -540,13 +534,9 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
 	 */
 	if (prev) {
 		tctx = bpf_task_storage_get(&task_ctx_stor, prev, 0, 0);
-		if (!tctx) {
-			scx_bpf_error("task_ctx lookup failed");
-			return;
-		}
-
-		tctx->core_sched_seq =
-			core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
+		if (tctx)
+			tctx->core_sched_seq =
+				core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
 	}
 }
 
@@ -584,10 +574,8 @@ static s64 task_qdist(struct task_struct *p)
 	s64 qdist;
 
 	tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
-	if (!tctx) {
-		scx_bpf_error("task_ctx lookup failed");
+	if (!tctx)
 		return 0;
-	}
 
 	qdist = tctx->core_sched_seq - core_sched_head_seqs[idx];
 
Signature · 1 line
2.53.0
Reviewed-by on v2 1/1
#3 lore raw
Quoted 6 lines · Tejun Heo
tools/sched_ext/scx_qmap.bpf.c +0 -0
index b68abb9e760b..aad698fe294b 100644
@@ -584,10 +574,8 @@ static s64 task_qdist(struct task_struct *p)
 	s64 qdist;

tools/sched_ext/scx_qmap.bpf.c static s64 task_qdist(struct task_struct *p) · line 581

Quoted 10 lines
tools/sched_ext/scx_qmap.bpf.c +1 -3
 
 	tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
-	if (!tctx) {
-		scx_bpf_error("task_ctx lookup failed");
+	if (!tctx)
 		return 0;
-	}
 
 	qdist = tctx->core_sched_seq - core_sched_head_seqs[idx];
 

Hi Tejun,
I met this error also on the lastest for-next branch, trying to prepare a fix and find your patch. Looks good to me.

Signature · 1 line
Zhao Mengmeng <[email protected]>