| Message ID | 20210619093151.1492174-7-linmiaohe@huawei.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | Cleanup and fixup for z3fold | expand |
On Sat, 19 Jun 2021 17:31:51 +0800 Miaohe Lin wrote: > We should use release_z3fold_page_locked() to release z3fold page when it's > locked, although it looks harmless to use release_z3fold_page() now. > > Fixes: dcf5aedb24f8 ("z3fold: stricter locking and more careful reclaim") > Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> > --- > mm/z3fold.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/mm/z3fold.c b/mm/z3fold.c > index 196d886a3436..b3c0577b8095 100644 > --- a/mm/z3fold.c > +++ b/mm/z3fold.c > @@ -1372,7 +1372,7 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries) > if (zhdr->foreign_handles || > test_and_set_bit(PAGE_CLAIMED, &page->private)) { > if (kref_put(&zhdr->refcount, > - release_z3fold_page)) > + release_z3fold_page_locked)) > atomic64_dec(&pool->pages_nr); LGTM. JFYI other issue in z3fold was reported [1] and if the fix proposed there makes any sense to you feel free to pick it up and ask Mike for his tests. [1] https://lore.kernel.org/linux-mm/20210316061351.1649-1-hdanton@sina.com/
On 2021/6/20 8:26, Hillf Danton wrote: > On Sat, 19 Jun 2021 17:31:51 +0800 Miaohe Lin wrote: >> We should use release_z3fold_page_locked() to release z3fold page when it's >> locked, although it looks harmless to use release_z3fold_page() now. >> >> Fixes: dcf5aedb24f8 ("z3fold: stricter locking and more careful reclaim") >> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> >> --- >> mm/z3fold.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/mm/z3fold.c b/mm/z3fold.c >> index 196d886a3436..b3c0577b8095 100644 >> --- a/mm/z3fold.c >> +++ b/mm/z3fold.c >> @@ -1372,7 +1372,7 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries) >> if (zhdr->foreign_handles || >> test_and_set_bit(PAGE_CLAIMED, &page->private)) { >> if (kref_put(&zhdr->refcount, >> - release_z3fold_page)) >> + release_z3fold_page_locked)) >> atomic64_dec(&pool->pages_nr); > > LGTM. JFYI other issue in z3fold was reported [1] and if the fix proposed there > makes any sense to you feel free to pick it up and ask Mike for his tests. > Thank you for review and reply. I browsed [1] several times but I failed to figure out what's the root cause. And I found some bugs and possible race windows from previous code inspection. I think we can try fix these first and see whether [1] is (hopefully) fixed. :) Thanks again. > [1] https://lore.kernel.org/linux-mm/20210316061351.1649-1-hdanton@sina.com/ > . >
diff --git a/mm/z3fold.c b/mm/z3fold.c index 196d886a3436..b3c0577b8095 100644 --- a/mm/z3fold.c +++ b/mm/z3fold.c @@ -1372,7 +1372,7 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries) if (zhdr->foreign_handles || test_and_set_bit(PAGE_CLAIMED, &page->private)) { if (kref_put(&zhdr->refcount, - release_z3fold_page)) + release_z3fold_page_locked)) atomic64_dec(&pool->pages_nr); else z3fold_page_unlock(zhdr);
We should use release_z3fold_page_locked() to release z3fold page when it's locked, although it looks harmless to use release_z3fold_page() now. Fixes: dcf5aedb24f8 ("z3fold: stricter locking and more careful reclaim") Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> --- mm/z3fold.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)