Re: [linux-next20250901] Boot fail on IBM Power Server

[Vlastimil Babka <vbabka@suse.cz>]

On 9/2/25 06:55, Venkat Rao Bagalkote wrote:
> Greetings!!!
> 
> 
> IBM CI has reported a boot failure with next-20250901 repo on IBM Power 
> Server.
> 
> 
> Repo: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
> 
> gcc version 11.5.0 20240719
> 
> GNU ld version 2.35.2-54.el9
> 
> 
> Attached is the .config file.
> 
> 
> Traces:
> 
> 
> [    0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
> [    0.000000] SLUB: HWalign=128, Order=0-3, MinObjects=0, CPUs=32, Nodes=32
> [    0.000000] BUG: Kernel NULL pointer dereference on read at 0x00000040
> [    0.000000] Faulting instruction address: 0xc00000000059eaac
> [    0.000000] Oops: Kernel access of bad area, sig: 7 [#1]
> [    0.000000] LE PAGE_SIZE=64K MMU=Radix  SMP NR_CPUS=8192 NUMA pSeries
> [    0.000000] Modules linked in:
> [    0.000000] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 
> 6.17.0-rc4-next-20250901 #1 VOLUNTARY
> [    0.000000] Hardware name: IBM,9080-HEX Power11 (architected) 
> 0x820200 0xf000007 of:IBM,FW1110.01 (NH1110_069) hv:phyp pSeries
> [    0.000000] NIP:  c00000000059eaac LR: c00000000059eadc CTR: 
> 0000000000000000
> [    0.000000] REGS: c000000002c7faf0 TRAP: 0300   Not tainted 
> (6.17.0-rc4-next-20250901)
> [    0.000000] MSR:  8000000002001033 <SF,VEC,ME,IR,DR,RI,LE>  CR: 
> 44008220  XER: 20040001
> [    0.000000] CFAR: c00000000059eaf8 DAR: 0000000000000040 DSISR: 
> 00080000 IRQMASK: 3
> [    0.000000] GPR00: c00000000059eab8 c000000002c7fd90 c000000001678100 
> c000000004017400
> [    0.000000] GPR04: 0000000000000cc0 0000000000000001 0000000000000000 
> c000000002cea768
> [    0.000000] GPR08: 0000000000000008 0000000000000000 0000000000000000 
> 0000000000008000
> [    0.000000] GPR12: c00000000058f8d0 c000000002ff0000 0000000000000000 
> 0000000000000000
> [    0.000000] GPR16: 0000000000000000 0000000000000000 0000000000000000 
> 0000000000000000
> [    0.000000] GPR20: 0000000000c00000 0000000000000008 0000000000000000 
> c000000002f46870
> [    0.000000] GPR24: 0000000000000100 c000000002f468c8 c000000002f425c0 
> c000000004017400
> [    0.000000] GPR28: c000000001591188 c000000002cfca00 0000000000000001 
> c0000013fd27ebe0
> [    0.000000] NIP [c00000000059eaac] do_kmem_cache_create+0x4d4/0x634
> [    0.000000] LR [c00000000059eadc] do_kmem_cache_create+0x504/0x634
> [    0.000000] Call Trace:
> [    0.000000] [c000000002c7fd90] [c00000000059eab8] 
> do_kmem_cache_create+0x4e0/0x634 (unreliable)
> [    0.000000] [c000000002c7fe40] [c00000000050e518] 
> __kmem_cache_create_args+0x198/0x434
> [    0.000000] [c000000002c7fef0] [c0000000020b3630] 
> maple_tree_init+0x60/0x98
> [    0.000000] [c000000002c7ff40] [c000000002006420] 
> start_kernel+0x288/0x60c
> [    0.000000] [c000000002c7ffe0] [c00000000000ea9c] 
> start_here_common+0x1c/0x20
> [    0.000000] Code: 4e800020 3d220167 38e92668 3d2200c0 39490a84 
> e93b0008 7fe7402a 7d5f52aa 7fe9fa14 392a001c 79291f24 7d3b482a 
> <e9290040> f93f0020 4bff15e5 38be0001
> [    0.000000] ---[ end trace 0000000000000000 ]---
> [    0.000000]
> [    0.000000] Kernel panic - not syncing: Fatal exception
> [    0.000000] Rebooting in 10 seconds..
> 
> 
> 
> If you happen to fix this please add below tag.
> 
> 
> Reported-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>

Could you check if this fixes it please?

From 345b6b43ebf2dacae570a6971ab147be0f724fa1 Mon Sep 17 00:00:00 2001
From: Vlastimil Babka <vbabka@suse.cz>
Date: Tue, 2 Sep 2025 14:36:44 +0200
Subject: [PATCH] slub: don't cache barn pointers in pcs

Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
---
 mm/slub.c | 69 +++++++++++++++++++++++++++++++++----------------------
 1 file changed, 42 insertions(+), 27 deletions(-)

diff --git a/mm/slub.c b/mm/slub.c
index e47032ea38e0..cb0ba1cf0ee4 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -481,7 +481,6 @@ struct slub_percpu_sheaves {
 	struct slab_sheaf *main; /* never NULL when unlocked */
 	struct slab_sheaf *spare; /* empty or full, may be NULL */
 	struct slab_sheaf *rcu_free; /* for batching kfree_rcu() */
-	struct node_barn *barn;
 };
 
 /*
@@ -504,6 +503,12 @@ static inline struct kmem_cache_node *get_node(struct kmem_cache *s, int node)
 	return s->node[node];
 }
 
+/* Get the barn of the current cpu's memory node */
+static inline struct node_barn *get_barn(struct kmem_cache *s)
+{
+	return get_node(s, numa_mem_id())->barn;
+}
+
 /*
  * Iterator over all nodes. The body will be executed for each node that has
  * a kmem_cache_node structure allocated (which is true for all online nodes)
@@ -4722,6 +4727,7 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 {
 	struct slab_sheaf *empty = NULL;
 	struct slab_sheaf *full;
+	struct node_barn *barn;
 	bool can_alloc;
 
 	lockdep_assert_held(this_cpu_ptr(&s->cpu_sheaves->lock));
@@ -4731,7 +4737,9 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 		return pcs;
 	}
 
-	full = barn_replace_empty_sheaf(pcs->barn, pcs->main);
+	barn = get_barn(s);
+
+	full = barn_replace_empty_sheaf(barn, pcs->main);
 
 	if (full) {
 		stat(s, BARN_GET);
@@ -4748,7 +4756,7 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 			empty = pcs->spare;
 			pcs->spare = NULL;
 		} else {
-			empty = barn_get_empty_sheaf(pcs->barn);
+			empty = barn_get_empty_sheaf(barn);
 		}
 	}
 
@@ -4780,6 +4788,7 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 	 */
 	local_lock(&s->cpu_sheaves->lock);
 	pcs = this_cpu_ptr(s->cpu_sheaves);
+	barn = get_barn(s);
 
 	/*
 	 * If we are returning empty sheaf, we either got it from the
@@ -4790,7 +4799,7 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 	 */
 
 	if (pcs->main->size == 0) {
-		barn_put_empty_sheaf(pcs->barn, pcs->main);
+		barn_put_empty_sheaf(barn, pcs->main);
 		pcs->main = full;
 		return pcs;
 	}
@@ -4801,12 +4810,12 @@ __pcs_replace_empty_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs,
 	}
 
 	if (pcs->spare->size == 0) {
-		barn_put_empty_sheaf(pcs->barn, pcs->spare);
+		barn_put_empty_sheaf(barn, pcs->spare);
 		pcs->spare = full;
 		return pcs;
 	}
 
-	barn_put_full_sheaf(pcs->barn, full);
+	barn_put_full_sheaf(barn, full);
 	stat(s, BARN_PUT);
 
 	return pcs;
@@ -4907,7 +4916,7 @@ unsigned int alloc_from_pcs_bulk(struct kmem_cache *s, size_t size, void **p)
 			goto do_alloc;
 		}
 
-		full = barn_replace_empty_sheaf(pcs->barn, pcs->main);
+		full = barn_replace_empty_sheaf(get_barn(s), pcs->main);
 
 		if (full) {
 			stat(s, BARN_GET);
@@ -5104,7 +5113,7 @@ kmem_cache_prefill_sheaf(struct kmem_cache *s, gfp_t gfp, unsigned int size)
 		stat(s, SHEAF_PREFILL_FAST);
 	} else {
 		stat(s, SHEAF_PREFILL_SLOW);
-		sheaf = barn_get_full_or_empty_sheaf(pcs->barn);
+		sheaf = barn_get_full_or_empty_sheaf(get_barn(s));
 		if (sheaf && sheaf->size)
 			stat(s, BARN_GET);
 		else
@@ -5155,6 +5164,7 @@ void kmem_cache_return_sheaf(struct kmem_cache *s, gfp_t gfp,
 
 	local_lock(&s->cpu_sheaves->lock);
 	pcs = this_cpu_ptr(s->cpu_sheaves);
+	barn = get_barn(s);
 
 	if (!pcs->spare) {
 		pcs->spare = sheaf;
@@ -5169,14 +5179,11 @@ void kmem_cache_return_sheaf(struct kmem_cache *s, gfp_t gfp,
 
 	stat(s, SHEAF_RETURN_SLOW);
 
-	/* Accessing pcs->barn outside local_lock is safe. */
-	barn = pcs->barn;
-
 	/*
 	 * If the barn has too many full sheaves or we fail to refill the sheaf,
 	 * simply flush and free it.
 	 */
-	if (data_race(pcs->barn->nr_full) >= MAX_FULL_SHEAVES ||
+	if (data_race(barn->nr_full) >= MAX_FULL_SHEAVES ||
 	    refill_sheaf(s, sheaf, gfp)) {
 		sheaf_flush_unused(s, sheaf);
 		free_empty_sheaf(s, sheaf);
@@ -5601,6 +5608,8 @@ static void __slab_free(struct kmem_cache *s, struct slab *slab,
 static void __pcs_install_empty_sheaf(struct kmem_cache *s,
 		struct slub_percpu_sheaves *pcs, struct slab_sheaf *empty)
 {
+	struct node_barn *barn;
+
 	lockdep_assert_held(this_cpu_ptr(&s->cpu_sheaves->lock));
 
 	/* This is what we expect to find if nobody interrupted us. */
@@ -5610,19 +5619,21 @@ static void __pcs_install_empty_sheaf(struct kmem_cache *s,
 		return;
 	}
 
+	barn = get_barn(s);
+
 	/*
 	 * Unlikely because if the main sheaf had space, we would have just
 	 * freed to it. Get rid of our empty sheaf.
 	 */
 	if (pcs->main->size < s->sheaf_capacity) {
-		barn_put_empty_sheaf(pcs->barn, empty);
+		barn_put_empty_sheaf(barn, empty);
 		return;
 	}
 
 	/* Also unlikely for the same reason */
 	if (pcs->spare->size < s->sheaf_capacity) {
 		swap(pcs->main, pcs->spare);
-		barn_put_empty_sheaf(pcs->barn, empty);
+		barn_put_empty_sheaf(barn, empty);
 		return;
 	}
 
@@ -5630,7 +5641,7 @@ static void __pcs_install_empty_sheaf(struct kmem_cache *s,
 	 * We probably failed barn_replace_full_sheaf() due to no empty sheaf
 	 * available there, but we allocated one, so finish the job.
 	 */
-	barn_put_full_sheaf(pcs->barn, pcs->main);
+	barn_put_full_sheaf(barn, pcs->main);
 	stat(s, BARN_PUT);
 	pcs->main = empty;
 }
@@ -5647,15 +5658,17 @@ static struct slub_percpu_sheaves *
 __pcs_replace_full_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs)
 {
 	struct slab_sheaf *empty;
+	struct node_barn *barn;
 	bool put_fail;
 
 restart:
 	lockdep_assert_held(this_cpu_ptr(&s->cpu_sheaves->lock));
 
+	barn = get_barn(s);
 	put_fail = false;
 
 	if (!pcs->spare) {
-		empty = barn_get_empty_sheaf(pcs->barn);
+		empty = barn_get_empty_sheaf(barn);
 		if (empty) {
 			pcs->spare = pcs->main;
 			pcs->main = empty;
@@ -5669,7 +5682,7 @@ __pcs_replace_full_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs)
 		return pcs;
 	}
 
-	empty = barn_replace_full_sheaf(pcs->barn, pcs->main);
+	empty = barn_replace_full_sheaf(barn, pcs->main);
 
 	if (!IS_ERR(empty)) {
 		stat(s, BARN_PUT);
@@ -5729,7 +5742,7 @@ __pcs_replace_full_main(struct kmem_cache *s, struct slub_percpu_sheaves *pcs)
 
 got_empty:
 	if (!local_trylock(&s->cpu_sheaves->lock)) {
-		barn_put_empty_sheaf(pcs->barn, empty);
+		barn_put_empty_sheaf(barn, empty);
 		return NULL;
 	}
 
@@ -5832,6 +5845,7 @@ bool __kfree_rcu_sheaf(struct kmem_cache *s, void *obj)
 	if (unlikely(!pcs->rcu_free)) {
 
 		struct slab_sheaf *empty;
+		struct node_barn *barn;
 
 		if (pcs->spare && pcs->spare->size == 0) {
 			pcs->rcu_free = pcs->spare;
@@ -5839,7 +5853,9 @@ bool __kfree_rcu_sheaf(struct kmem_cache *s, void *obj)
 			goto do_free;
 		}
 
-		empty = barn_get_empty_sheaf(pcs->barn);
+		barn = get_barn(s);
+
+		empty = barn_get_empty_sheaf(barn);
 
 		if (empty) {
 			pcs->rcu_free = empty;
@@ -5854,14 +5870,14 @@ bool __kfree_rcu_sheaf(struct kmem_cache *s, void *obj)
 			goto fail;
 
 		if (!local_trylock(&s->cpu_sheaves->lock)) {
-			barn_put_empty_sheaf(pcs->barn, empty);
+			barn_put_empty_sheaf(barn, empty);
 			goto fail;
 		}
 
 		pcs = this_cpu_ptr(s->cpu_sheaves);
 
 		if (unlikely(pcs->rcu_free))
-			barn_put_empty_sheaf(pcs->barn, empty);
+			barn_put_empty_sheaf(get_barn(s), empty);
 		else
 			pcs->rcu_free = empty;
 	}
@@ -5906,6 +5922,7 @@ static void free_to_pcs_bulk(struct kmem_cache *s, size_t size, void **p)
 	void *remote_objects[PCS_BATCH_MAX];
 	unsigned int remote_nr = 0;
 	int node = numa_mem_id();
+	struct node_barn *barn;
 
 next_remote_batch:
 	while (i < size) {
@@ -5941,8 +5958,10 @@ static void free_to_pcs_bulk(struct kmem_cache *s, size_t size, void **p)
 	if (likely(pcs->main->size < s->sheaf_capacity))
 		goto do_free;
 
+	barn = get_barn(s);
+
 	if (!pcs->spare) {
-		empty = barn_get_empty_sheaf(pcs->barn);
+		empty = barn_get_empty_sheaf(barn);
 		if (!empty)
 			goto no_empty;
 
@@ -5956,7 +5975,7 @@ static void free_to_pcs_bulk(struct kmem_cache *s, size_t size, void **p)
 		goto do_free;
 	}
 
-	empty = barn_replace_full_sheaf(pcs->barn, pcs->main);
+	empty = barn_replace_full_sheaf(barn, pcs->main);
 	if (IS_ERR(empty)) {
 		stat(s, BARN_PUT_FAIL);
 		goto no_empty;
@@ -7041,15 +7060,11 @@ static int init_percpu_sheaves(struct kmem_cache *s)
 
 	for_each_possible_cpu(cpu) {
 		struct slub_percpu_sheaves *pcs;
-		int nid;
 
 		pcs = per_cpu_ptr(s->cpu_sheaves, cpu);
 
 		local_trylock_init(&pcs->lock);
 
-		nid = cpu_to_mem(cpu);
-
-		pcs->barn = get_node(s, nid)->barn;
 		pcs->main = alloc_empty_sheaf(s, GFP_KERNEL);
 
 		if (!pcs->main)
-- 
2.51.0