diff options
| author |   Thomas Gleixner <tglx@linutronix.de> | 2023-05-25 14:57:07 +0200 |
|---|---|---|
| committer |   Andrew Morton <akpm@linux-foundation.org> | 2023-06-09 16:25:41 -0700 |
| commit | 43d7650234c62201ba3ca5b731226b0b189989a8 (patch) | |
| tree | 39e61f957966f409b0010dc8c80ea026f975dcc5 /mm/vmalloc.c | |
| parent | mm/vmalloc: prevent flushing dirty space over and over (diff) | |
| download | wireguard-linux-43d7650234c62201ba3ca5b731226b0b189989a8.tar.xz wireguard-linux-43d7650234c62201ba3ca5b731226b0b189989a8.zip | |
mm/vmalloc: check free space in vmap_block lockless
vb_alloc() unconditionally locks a vmap_block on the free list to check
the free space.
This can be done locklessly because vmap_block::free never increases, it's
only decreased on allocations.
Check the free space lockless and only if that succeeds, recheck under the
lock.
Link: https://lkml.kernel.org/r/20230525124504.750481992@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Lorenzo Stoakes <lstoakes@gmail.com>
Reviewed-by: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Diffstat (limited to 'mm/vmalloc.c')
| -rw-r--r-- | mm/vmalloc.c | 5 |
1 files changed, 4 insertions, 1 deletions
diff --git a/mm/vmalloc.c b/mm/vmalloc.c index ad9a1d9e314f..679112e2ffd2 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2168,6 +2168,9 @@ static void *vb_alloc(unsigned long size, gfp_t gfp_mask) list_for_each_entry_rcu(vb, &vbq->free, free_list) { unsigned long pages_off; + if (READ_ONCE(vb->free) < (1UL << order)) + continue; + spin_lock(&vb->lock); if (vb->free < (1UL << order)) { spin_unlock(&vb->lock); @@ -2176,7 +2179,7 @@ static void *vb_alloc(unsigned long size, gfp_t gfp_mask) pages_off = VMAP_BBMAP_BITS - vb->free; vaddr = vmap_block_vaddr(vb->va->va_start, pages_off); - vb->free -= 1UL << order; + WRITE_ONCE(vb->free, vb->free - (1UL << order)); bitmap_set(vb->used_map, pages_off, (1UL << order)); if (vb->free == 0) { spin_lock(&vbq->lock); |