mm/mmap: free vm_area_struct without call_rcu in exit_mmap

call_rcu() can take a long time when callback offloading is enabled.  Its
use in the vm_area_free can cause regressions in the exit path when
multiple VMAs are being freed.

Because exit_mmap() is called only after the last mm user drops its
refcount, the page fault handlers can't be racing with it.  Any other
possible user like oom-reaper or process_mrelease are already synchronized
using mmap_lock.  Therefore exit_mmap() can free VMAs directly, without
the use of call_rcu().

Expose __vm_area_free() and use it from exit_mmap() to avoid possible
call_rcu() floods and performance regressions caused by it.

Link: https://lkml.kernel.org/r/20230227173632.3292573-33-surenb@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This commit is contained in:
Suren Baghdasaryan
2023-02-27 09:36:31 -08:00
committed by Andrew Morton
parent 70d4cbc80c
commit 0d2ebf9c3f
3 changed files with 10 additions and 5 deletions

View File

@@ -480,7 +480,7 @@ struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig)
return new;
}
static void __vm_area_free(struct vm_area_struct *vma)
void __vm_area_free(struct vm_area_struct *vma)
{
free_anon_vma_name(vma);
kmem_cache_free(vm_area_cachep, vma);