Skip to contentSkip to conversation

fscrypt,f2fs: introduce fscrypt_finalize_bounce_folio() for cleanup

PATCH In review Chao Yu 2 participants 3 messages Jun 22 – Jun 24

Subsystems (2)
lore
Share & tools
lore

Chao Yu

author maintainer
#0 lore raw

PATCH 1/1 fscrypt,f2fs: introduce fscrypt_finalize_bounce_folio() for cleanup

As part of the linux kernel's migration to folio-based APIs, introduce fscrypt_finalize_bounce_folio() as the folio equivalent of fscrypt_finalize_bounce_page(), and clean up f2fs codes with this new helper.

---

Is it worth to introduce fscrypt_finalize_bounce_folio(), then try to do clean in f2fs_write_end_bio() first, and then replace fscrypt_finalize_bounce_page() later?

 fs/f2fs/data.c          |  7 +------
 include/linux/fscrypt.h | 11 +++++++++++
 2 files changed, 12 insertions(+), 6 deletions(-)
fs/f2fs/data.c +1 -6
index ac1cf4de3d62..e0fca6c60e34 100644
@@ -362,12 +362,7 @@ static void f2fs_write_end_bio(struct bio *bio)
 		struct folio *folio = fi.folio;
 		enum count_type type;
 
-		if (fscrypt_is_bounce_folio(folio)) {
-			struct folio *io_folio = folio;
-
-			folio = fscrypt_pagecache_folio(io_folio);
-			fscrypt_free_bounce_page(&io_folio->page);
-		}
+		fscrypt_finalize_bounce_folio(&folio);
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 		if (f2fs_is_compressed_page(folio)) {
include/linux/fscrypt.h +11 -0
index 54712ec61ffb..20b59b021c94 100644
@@ -1132,4 +1132,15 @@ static inline void fscrypt_finalize_bounce_page(struct page **pagep)
 	}
 }
 
+/* If *foliop is a bounce folio, free it and set *foliop to the pagecache folio */
+static inline void fscrypt_finalize_bounce_folio(struct folio **foliop)
+{
+	struct folio *folio = *foliop;
+
+	if (fscrypt_is_bounce_folio(folio)) {
+		*foliop = fscrypt_pagecache_folio(folio);
+		fscrypt_free_bounce_page(&folio->page);
+	}
+}
+
 #endif	/* _LINUX_FSCRYPT_H */
Signature · 1 line
2.49.0
maintainer
#1 lore raw
Quoted 5 lines · Chao Yu

As part of the linux kernel's migration to folio-based APIs, introduce fscrypt_finalize_bounce_folio() as the folio equivalent of fscrypt_finalize_bounce_page(), and clean up f2fs codes with this new helper.

Quoted 8 lines

---

Is it worth to introduce fscrypt_finalize_bounce_folio(), then try to do clean in f2fs_write_end_bio() first, and then replace fscrypt_finalize_bounce_page() later?

I'm working on making ext4 and f2fs always do file contents en/decryption using fscrypt_set_bio_crypt_ctx(), which already supports large folios and doesn't require the filesystem to manage bounce buffers. I don't think these minor tweaks to the other implementation (which don't actually make it support large folios) accomplish anything useful, and we should focus on removing it instead.

- Eric

author maintainer
#2 lore raw
Quoted 15 lines · Eric Biggers
Quoted 5 lines · Chao Yu

As part of the linux kernel's migration to folio-based APIs, introduce fscrypt_finalize_bounce_folio() as the folio equivalent of fscrypt_finalize_bounce_page(), and clean up f2fs codes with this new helper.

Quoted 8 lines

---

Is it worth to introduce fscrypt_finalize_bounce_folio(), then try to do clean in f2fs_write_end_bio() first, and then replace fscrypt_finalize_bounce_page() later?

Quoted 6 lines

I'm working on making ext4 and f2fs always do file contents en/decryption using fscrypt_set_bio_crypt_ctx(), which already supports large folios and doesn't require the filesystem to manage bounce buffers. I don't think these minor tweaks to the other implementation (which don't actually make it support large folios) accomplish anything useful, and we should focus on removing it instead.

Good, I see your patchset removing those codes, let's ignore current patch.

Thanks,

Quoted 2 lines

- Eric