percpu: allow pcpu_alloc() to be called with IRQs off
pcpu_alloc() and pcpu_extend_area_map() perform a series of
spin_lock_irq()/spin_unlock_irq() calls, which make them unsafe
with respect to being called from contexts which have IRQs off.
This patch converts the code to perform save/restore of flags instead,
making pcpu_alloc() (or __alloc_percpu() respectively) to be called
from early kernel startup stage, where IRQs are off.
This is needed for proper initialization of per-cpu rq_weight data from
sched_init().
tj: added comment explaining why irqsave/restore is used in alloc path.
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Tejun Heo <tj@kernel.org>
diff --git a/mm/percpu.c b/mm/percpu.c
index 6af78c1e..d907971 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -153,7 +153,10 @@
*
* During allocation, pcpu_alloc_mutex is kept locked all the time and
* pcpu_lock is grabbed and released as necessary. All actual memory
- * allocations are done using GFP_KERNEL with pcpu_lock released.
+ * allocations are done using GFP_KERNEL with pcpu_lock released. In
+ * general, percpu memory can't be allocated with irq off but
+ * irqsave/restore are still used in alloc path so that it can be used
+ * from early init path - sched_init() specifically.
*
* Free path accesses and alters only the index data structures, so it
* can be safely called from atomic context. When memory needs to be
@@ -366,7 +369,7 @@
* RETURNS:
* 0 if noop, 1 if successfully extended, -errno on failure.
*/
-static int pcpu_extend_area_map(struct pcpu_chunk *chunk)
+static int pcpu_extend_area_map(struct pcpu_chunk *chunk, unsigned long *flags)
{
int new_alloc;
int *new;
@@ -376,7 +379,7 @@
if (chunk->map_alloc >= chunk->map_used + 2)
return 0;
- spin_unlock_irq(&pcpu_lock);
+ spin_unlock_irqrestore(&pcpu_lock, *flags);
new_alloc = PCPU_DFL_MAP_ALLOC;
while (new_alloc < chunk->map_used + 2)
@@ -384,7 +387,7 @@
new = pcpu_mem_alloc(new_alloc * sizeof(new[0]));
if (!new) {
- spin_lock_irq(&pcpu_lock);
+ spin_lock_irqsave(&pcpu_lock, *flags);
return -ENOMEM;
}
@@ -393,7 +396,7 @@
* could have happened inbetween, so map_used couldn't have
* grown.
*/
- spin_lock_irq(&pcpu_lock);
+ spin_lock_irqsave(&pcpu_lock, *flags);
BUG_ON(new_alloc < chunk->map_used + 2);
size = chunk->map_alloc * sizeof(chunk->map[0]);
@@ -1047,6 +1050,7 @@
struct pcpu_chunk *chunk;
const char *err;
int slot, off;
+ unsigned long flags;
if (unlikely(!size || size > PCPU_MIN_UNIT_SIZE || align > PAGE_SIZE)) {
WARN(true, "illegal size (%zu) or align (%zu) for "
@@ -1055,13 +1059,13 @@
}
mutex_lock(&pcpu_alloc_mutex);
- spin_lock_irq(&pcpu_lock);
+ spin_lock_irqsave(&pcpu_lock, flags);
/* serve reserved allocations from the reserved chunk if available */
if (reserved && pcpu_reserved_chunk) {
chunk = pcpu_reserved_chunk;
if (size > chunk->contig_hint ||
- pcpu_extend_area_map(chunk) < 0) {
+ pcpu_extend_area_map(chunk, &flags) < 0) {
err = "failed to extend area map of reserved chunk";
goto fail_unlock;
}
@@ -1079,7 +1083,7 @@
if (size > chunk->contig_hint)
continue;
- switch (pcpu_extend_area_map(chunk)) {
+ switch (pcpu_extend_area_map(chunk, &flags)) {
case 0:
break;
case 1:
@@ -1096,7 +1100,7 @@
}
/* hmmm... no space left, create a new chunk */
- spin_unlock_irq(&pcpu_lock);
+ spin_unlock_irqrestore(&pcpu_lock, flags);
chunk = alloc_pcpu_chunk();
if (!chunk) {
@@ -1104,16 +1108,16 @@
goto fail_unlock_mutex;
}
- spin_lock_irq(&pcpu_lock);
+ spin_lock_irqsave(&pcpu_lock, flags);
pcpu_chunk_relocate(chunk, -1);
goto restart;
area_found:
- spin_unlock_irq(&pcpu_lock);
+ spin_unlock_irqrestore(&pcpu_lock, flags);
/* populate, map and clear the area */
if (pcpu_populate_chunk(chunk, off, size)) {
- spin_lock_irq(&pcpu_lock);
+ spin_lock_irqsave(&pcpu_lock, flags);
pcpu_free_area(chunk, off);
err = "failed to populate";
goto fail_unlock;
@@ -1125,7 +1129,7 @@
return __addr_to_pcpu_ptr(chunk->base_addr + off);
fail_unlock:
- spin_unlock_irq(&pcpu_lock);
+ spin_unlock_irqrestore(&pcpu_lock, flags);
fail_unlock_mutex:
mutex_unlock(&pcpu_alloc_mutex);
if (warn_limit) {