tjh.dev / kernel
Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux
1
fork

Configure Feed

Select the types of activity you want to include in your feed.

mm: use folio_xor_flags_has_waiters() in folio_end_writeback()

Match how folio_unlock() works by combining the test for PG_waiters with
the clearing of PG_writeback. This should have a small performance win,
and removes the last user of folio_wake().

Link: https://lkml.kernel.org/r/20231004165317.1061855-18-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Richard Henderson <richard.henderson@linaro.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

authored by

Matthew Wilcox (Oracle) and committed by
Andrew Morton 2580d554 7d0795d0

+10 -16
+3 -12
mm/filemap.c
··· 1175 1175 spin_unlock_irqrestore(&q->lock, flags); 1176 1176 } 1177 1177 1178 - static void folio_wake(struct folio *folio, int bit) 1179 - { 1180 - if (!folio_test_waiters(folio)) 1181 - return; 1182 - folio_wake_bit(folio, bit); 1183 - } 1184 - 1185 1178 /* 1186 1179 * A choice of three behaviors for folio_wait_bit_common(): 1187 1180 */ ··· 1611 1618 * Writeback does not hold a folio reference of its own, relying 1612 1619 * on truncation to wait for the clearing of PG_writeback. 1613 1620 * But here we must make sure that the folio is not freed and 1614 - * reused before the folio_wake(). 1621 + * reused before the folio_wake_bit(). 1615 1622 */ 1616 1623 folio_get(folio); 1617 - __folio_end_writeback(folio); 1618 - 1619 - smp_mb__after_atomic(); 1620 - folio_wake(folio, PG_writeback); 1624 + if (__folio_end_writeback(folio)) 1625 + folio_wake_bit(folio, PG_writeback); 1621 1626 acct_reclaim_writeback(folio); 1622 1627 folio_put(folio); 1623 1628 }
+1 -1
mm/internal.h
··· 105 105 106 106 vm_fault_t do_swap_page(struct vm_fault *vmf); 107 107 void folio_rotate_reclaimable(struct folio *folio); 108 - void __folio_end_writeback(struct folio *folio); 108 + bool __folio_end_writeback(struct folio *folio); 109 109 void deactivate_file_folio(struct folio *folio); 110 110 void folio_activate(struct folio *folio); 111 111
+6 -3
mm/page-writeback.c
··· 2940 2940 spin_unlock_irqrestore(&wb->work_lock, flags); 2941 2941 } 2942 2942 2943 - void __folio_end_writeback(struct folio *folio) 2943 + bool __folio_end_writeback(struct folio *folio) 2944 2944 { 2945 2945 long nr = folio_nr_pages(folio); 2946 2946 struct address_space *mapping = folio_mapping(folio); 2947 + bool ret; 2947 2948 2948 2949 folio_memcg_lock(folio); 2949 2950 if (mapping && mapping_use_writeback_tags(mapping)) { ··· 2953 2952 unsigned long flags; 2954 2953 2955 2954 xa_lock_irqsave(&mapping->i_pages, flags); 2956 - folio_test_clear_writeback(folio); 2955 + ret = folio_xor_flags_has_waiters(folio, 1 << PG_writeback); 2957 2956 __xa_clear_mark(&mapping->i_pages, folio_index(folio), 2958 2957 PAGECACHE_TAG_WRITEBACK); 2959 2958 if (bdi->capabilities & BDI_CAP_WRITEBACK_ACCT) { ··· 2971 2970 2972 2971 xa_unlock_irqrestore(&mapping->i_pages, flags); 2973 2972 } else { 2974 - folio_test_clear_writeback(folio); 2973 + ret = folio_xor_flags_has_waiters(folio, 1 << PG_writeback); 2975 2974 } 2976 2975 2977 2976 lruvec_stat_mod_folio(folio, NR_WRITEBACK, -nr); 2978 2977 zone_stat_mod_folio(folio, NR_ZONE_WRITE_PENDING, -nr); 2979 2978 node_stat_mod_folio(folio, NR_WRITTEN, nr); 2980 2979 folio_memcg_unlock(folio); 2980 + 2981 + return ret; 2981 2982 } 2982 2983 2983 2984 bool __folio_start_writeback(struct folio *folio, bool keep_write)