]> git.ipfire.org Git - thirdparty/kernel/linux.git/commitdiff
sched_ext: scx_qmap: rename tctx to taskc
authorTejun Heo <tj@kernel.org>
Thu, 16 Apr 2026 16:41:15 +0000 (06:41 -1000)
committerTejun Heo <tj@kernel.org>
Thu, 16 Apr 2026 17:31:07 +0000 (07:31 -1000)
Rename the per-task context local variable from tctx to taskc for
consistency.

No functional change.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
tools/sched_ext/scx_qmap.bpf.c

index b68abb9e760b892f2df055649dc80c71ed57651c..a18234f3c27a6e2322224d385325685adbecf450 100644 (file)
@@ -159,22 +159,22 @@ 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;
+       struct task_ctx *taskc;
 
-       if (!(tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0))) {
+       if (!(taskc = bpf_task_storage_get(&task_ctx_stor, p, 0, 0))) {
                scx_bpf_error("task_ctx lookup failed");
                return NULL;
        }
-       return tctx;
+       return taskc;
 }
 
 s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
                   s32 prev_cpu, u64 wake_flags)
 {
-       struct task_ctx *tctx;
+       struct task_ctx *taskc;
        s32 cpu;
 
-       if (!(tctx = lookup_task_ctx(p)))
+       if (!(taskc = lookup_task_ctx(p)))
                return -ESRCH;
 
        if (p->scx.weight < 2 && !(p->flags & PF_KTHREAD))
@@ -183,7 +183,7 @@ s32 BPF_STRUCT_OPS(qmap_select_cpu, struct task_struct *p,
        cpu = pick_direct_dispatch_cpu(p, prev_cpu);
 
        if (cpu >= 0) {
-               tctx->force_local = true;
+               taskc->force_local = true;
                return cpu;
        } else {
                return prev_cpu;
@@ -208,7 +208,7 @@ static int weight_to_idx(u32 weight)
 void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
 {
        static u32 user_cnt, kernel_cnt;
-       struct task_ctx *tctx;
+       struct task_ctx *taskc;
        u32 pid = p->pid;
        int idx = weight_to_idx(p->scx.weight);
        void *ring;
@@ -231,14 +231,14 @@ void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
        if (test_error_cnt && !--test_error_cnt)
                scx_bpf_error("test triggering error");
 
-       if (!(tctx = lookup_task_ctx(p)))
+       if (!(taskc = lookup_task_ctx(p)))
                return;
 
        /*
         * All enqueued tasks must have their core_sched_seq updated for correct
         * core-sched ordering. Also, take a look at the end of qmap_dispatch().
         */
-       tctx->core_sched_seq = core_sched_tail_seqs[idx]++;
+       taskc->core_sched_seq = core_sched_tail_seqs[idx]++;
 
        /*
         * IMMED stress testing: Every immed_stress_nth'th enqueue, dispatch
@@ -249,7 +249,7 @@ void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
                static u32 immed_stress_cnt;
 
                if (!(++immed_stress_cnt % immed_stress_nth)) {
-                       tctx->force_local = false;
+                       taskc->force_local = false;
                        scx_bpf_dsq_insert(p, SCX_DSQ_LOCAL_ON | scx_bpf_task_cpu(p),
                                           slice_ns, enq_flags);
                        return;
@@ -260,8 +260,8 @@ void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
         * If qmap_select_cpu() is telling us to or this is the last runnable
         * task on the CPU, enqueue locally.
         */
-       if (tctx->force_local) {
-               tctx->force_local = false;
+       if (taskc->force_local) {
+               taskc->force_local = false;
                scx_bpf_dsq_insert(p, SCX_DSQ_LOCAL, slice_ns, enq_flags);
                return;
        }
@@ -310,7 +310,7 @@ void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
        }
 
        if (highpri_boosting && p->scx.weight >= HIGHPRI_WEIGHT) {
-               tctx->highpri = true;
+               taskc->highpri = true;
                __sync_fetch_and_add(&nr_highpri_queued, 1);
        }
        __sync_fetch_and_add(&nr_enqueued, 1);
@@ -330,10 +330,10 @@ void BPF_STRUCT_OPS(qmap_dequeue, struct task_struct *p, u64 deq_flags)
 static void update_core_sched_head_seq(struct task_struct *p)
 {
        int idx = weight_to_idx(p->scx.weight);
-       struct task_ctx *tctx;
+       struct task_ctx *taskc;
 
-       if ((tctx = lookup_task_ctx(p)))
-               core_sched_head_seqs[idx] = tctx->core_sched_seq;
+       if ((taskc = lookup_task_ctx(p)))
+               core_sched_head_seqs[idx] = taskc->core_sched_seq;
 }
 
 /*
@@ -354,12 +354,12 @@ static bool dispatch_highpri(bool from_timer)
        /* scan SHARED_DSQ and move highpri tasks to HIGHPRI_DSQ */
        bpf_for_each(scx_dsq, p, SHARED_DSQ, 0) {
                static u64 highpri_seq;
-               struct task_ctx *tctx;
+               struct task_ctx *taskc;
 
-               if (!(tctx = lookup_task_ctx(p)))
+               if (!(taskc = lookup_task_ctx(p)))
                        return false;
 
-               if (tctx->highpri) {
+               if (taskc->highpri) {
                        /* exercise the set_*() and vtime interface too */
                        scx_bpf_dsq_move_set_slice(BPF_FOR_EACH_ITER, slice_ns * 2);
                        scx_bpf_dsq_move_set_vtime(BPF_FOR_EACH_ITER, highpri_seq++);
@@ -405,7 +405,7 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
 {
        struct task_struct *p;
        struct cpu_ctx *cpuc;
-       struct task_ctx *tctx;
+       struct task_ctx *taskc;
        u32 zero = 0, batch = dsp_batch ?: 1;
        void *fifo;
        s32 i, pid;
@@ -450,7 +450,7 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
 
                /* Dispatch or advance. */
                bpf_repeat(BPF_MAX_LOOPS) {
-                       struct task_ctx *tctx;
+                       struct task_ctx *taskc;
 
                        if (bpf_map_pop_elem(fifo, &pid))
                                break;
@@ -459,12 +459,12 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
                        if (!p)
                                continue;
 
-                       if (!(tctx = lookup_task_ctx(p))) {
+                       if (!(taskc = lookup_task_ctx(p))) {
                                bpf_task_release(p);
                                return;
                        }
 
-                       if (tctx->highpri)
+                       if (taskc->highpri)
                                __sync_fetch_and_sub(&nr_highpri_queued, 1);
 
                        update_core_sched_head_seq(p);
@@ -539,13 +539,13 @@ void BPF_STRUCT_OPS(qmap_dispatch, s32 cpu, struct task_struct *prev)
         * if the task were enqueued and dispatched immediately.
         */
        if (prev) {
-               tctx = bpf_task_storage_get(&task_ctx_stor, prev, 0, 0);
-               if (!tctx) {
+               taskc = bpf_task_storage_get(&task_ctx_stor, prev, 0, 0);
+               if (!taskc) {
                        scx_bpf_error("task_ctx lookup failed");
                        return;
                }
 
-               tctx->core_sched_seq =
+               taskc->core_sched_seq =
                        core_sched_tail_seqs[weight_to_idx(prev->scx.weight)]++;
        }
 }
@@ -580,16 +580,16 @@ void BPF_STRUCT_OPS(qmap_tick, struct task_struct *p)
 static s64 task_qdist(struct task_struct *p)
 {
        int idx = weight_to_idx(p->scx.weight);
-       struct task_ctx *tctx;
+       struct task_ctx *taskc;
        s64 qdist;
 
-       tctx = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
-       if (!tctx) {
+       taskc = bpf_task_storage_get(&task_ctx_stor, p, 0, 0);
+       if (!taskc) {
                scx_bpf_error("task_ctx lookup failed");
                return 0;
        }
 
-       qdist = tctx->core_sched_seq - core_sched_head_seqs[idx];
+       qdist = taskc->core_sched_seq - core_sched_head_seqs[idx];
 
        /*
         * As queue index increments, the priority doubles. The queue w/ index 3