[PATCH v4 02/10] ceph: wait for pending fscache write in write_folio_nounlock()

write_folio_nounlock() marks the folio with PG_private_2 via
ceph_set_page_fscache() and starts an async write to the cache via
ceph_fscache_write_to_cache(). PG_private_2 is only cleared once that
write completes.

If the OSD write fails, the error paths return without waiting for the
cache write. On retry, ceph_find_incompatible() would wait for
writeback, but not for PG_private_2, before calling
write_folio_nounlock() again, which would trip the VM_BUG_ON_FOLIO() in
folio_start_private_2().

Wait for any pending cache write before starting writeback on the
folio. Do so before bumping the writeback congestion count and
allocating the possibly mempool-backed OSD request, so that a thread
sleeping on the cache I/O does not hold either while it waits.
Add ceph_folio_wait_fscache(), a wrapper for folio_wait_private_2(), and
use it.

Fixes: 1702e7973410 ("ceph: add fscache writeback support")
Reported-by: Sashiko <[email protected]>
Link: https://sashiko.dev/#/patchset/20260804-remove-wait-on-page-writeback-v2-0-81f0ab065284%40columbia.edu?part=7
Signed-off-by: Tal Zussman <[email protected]>
---
 fs/ceph/addr.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index d5a349daf4d0..f1f4caa8f9df 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -559,6 +559,11 @@ static void ceph_set_page_fscache(struct page *page)
 	folio_start_private_2(page_folio(page)); /* [DEPRECATED] */
 }
 
+static void ceph_folio_wait_fscache(struct folio *folio)
+{
+	folio_wait_private_2(folio); /* [DEPRECATED] */
+}
+
 static void ceph_fscache_write_terminated(void *priv, ssize_t error)
 {
 	struct inode *inode = priv;
@@ -580,6 +585,10 @@ static inline void ceph_set_page_fscache(struct page *page)
 {
 }
 
+static inline void ceph_folio_wait_fscache(struct folio *folio)
+{
+}
+
 static inline void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, bool caching)
 {
 }
@@ -783,6 +792,9 @@ static int write_folio_nounlock(struct folio *folio,
 	      ceph_vinop(inode), folio, folio->index, page_off, wlen, snapc,
 	      snapc->seq);
 
+	/* wait for a cache write left pending by a previously failed attempt */
+	ceph_folio_wait_fscache(folio);
+
 	if (atomic_long_inc_return(&fsc->writeback_count) >
 	    CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb))
 		fsc->write_congested = true;

-- 
2.39.5