| Message ID | 20241129055058.858940-6-willy@infradead.org |
|---|---|
| State | New |
| Headers | show |
| Series | Remove accesses to page->index from ceph | expand |
Hi Matthew,
kernel test robot noticed the following build warnings:
[auto build test WARNING on linus/master]
[also build test WARNING on next-20241128]
[cannot apply to ceph-client/testing ceph-client/for-linus v6.12]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Matthew-Wilcox-Oracle/ceph-Do-not-look-at-the-index-of-an-encrypted-page/20241129-135252
base: linus/master
patch link: https://lore.kernel.org/r/20241129055058.858940-6-willy%40infradead.org
patch subject: [PATCH 5/5] ceph: Use a folio in ceph_writepages_start()
config: x86_64-buildonly-randconfig-002-20241129 (https://download.01.org/0day-ci/archive/20241129/202411292322.iv4hDiEQ-lkp@intel.com/config)
compiler: clang version 19.1.3 (https://github.com/llvm/llvm-project ab51eccf88f5321e7c60591c5546b254b6afab99)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241129/202411292322.iv4hDiEQ-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202411292322.iv4hDiEQ-lkp@intel.com/
All warnings (new ones prefixed by >>):
In file included from fs/ceph/addr.c:4:
In file included from include/linux/backing-dev.h:16:
In file included from include/linux/writeback.h:13:
In file included from include/linux/blk_types.h:10:
In file included from include/linux/bvec.h:10:
In file included from include/linux/highmem.h:8:
In file included from include/linux/cacheflush.h:5:
In file included from arch/x86/include/asm/cacheflush.h:5:
In file included from include/linux/mm.h:2223:
include/linux/vmstat.h:504:43: warning: arithmetic between different enumeration types ('enum zone_stat_item' and 'enum numa_stat_item') [-Wenum-enum-conversion]
504 | return vmstat_text[NR_VM_ZONE_STAT_ITEMS +
| ~~~~~~~~~~~~~~~~~~~~~ ^
505 | item];
| ~~~~
include/linux/vmstat.h:511:43: warning: arithmetic between different enumeration types ('enum zone_stat_item' and 'enum numa_stat_item') [-Wenum-enum-conversion]
511 | return vmstat_text[NR_VM_ZONE_STAT_ITEMS +
| ~~~~~~~~~~~~~~~~~~~~~ ^
512 | NR_VM_NUMA_EVENT_ITEMS +
| ~~~~~~~~~~~~~~~~~~~~~~
include/linux/vmstat.h:518:36: warning: arithmetic between different enumeration types ('enum node_stat_item' and 'enum lru_list') [-Wenum-enum-conversion]
518 | return node_stat_name(NR_LRU_BASE + lru) + 3; // skip "nr_"
| ~~~~~~~~~~~ ^ ~~~
>> fs/ceph/addr.c:1178:39: warning: variable 'page' is uninitialized when used here [-Wuninitialized]
1178 | fscrypt_encrypt_pagecache_blocks(page,
| ^~~~
fs/ceph/addr.c:1043:20: note: initialize the variable 'page' to silence this warning
1043 | struct page *page;
| ^
| = NULL
4 warnings generated.
vim +/page +1178 fs/ceph/addr.c
1d3576fd10f0d7a Sage Weil 2009-10-06 941
1d3576fd10f0d7a Sage Weil 2009-10-06 942 /*
1d3576fd10f0d7a Sage Weil 2009-10-06 943 * initiate async writeback
1d3576fd10f0d7a Sage Weil 2009-10-06 944 */
1d3576fd10f0d7a Sage Weil 2009-10-06 945 static int ceph_writepages_start(struct address_space *mapping,
1d3576fd10f0d7a Sage Weil 2009-10-06 946 struct writeback_control *wbc)
1d3576fd10f0d7a Sage Weil 2009-10-06 947 {
1d3576fd10f0d7a Sage Weil 2009-10-06 948 struct inode *inode = mapping->host;
1d3576fd10f0d7a Sage Weil 2009-10-06 949 struct ceph_inode_info *ci = ceph_inode(inode);
5995d90d2d19f33 Xiubo Li 2023-06-12 950 struct ceph_fs_client *fsc = ceph_inode_to_fs_client(inode);
38d46409c4639a1 Xiubo Li 2023-06-12 951 struct ceph_client *cl = fsc->client;
fc2744aa12da718 Yan, Zheng 2013-05-31 952 struct ceph_vino vino = ceph_vino(inode);
2a2d927e35dd8dc Yan, Zheng 2017-09-01 953 pgoff_t index, start_index, end = -1;
80e755fedebc8de Sage Weil 2010-03-31 954 struct ceph_snap_context *snapc = NULL, *last_snapc = NULL, *pgsnapc;
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 955) struct folio_batch fbatch;
1d3576fd10f0d7a Sage Weil 2009-10-06 956 int rc = 0;
93407472a21b82f Fabian Frederick 2017-02-27 957 unsigned int wsize = i_blocksize(inode);
1d3576fd10f0d7a Sage Weil 2009-10-06 958 struct ceph_osd_request *req = NULL;
1f934b00e907527 Yan, Zheng 2017-08-30 959 struct ceph_writeback_ctl ceph_wbc;
590e9d9861f5f21 Yan, Zheng 2017-09-03 960 bool should_loop, range_whole = false;
af9cc401ce7452f Yan, Zheng 2018-03-04 961 bool done = false;
1702e79734104d7 Jeff Layton 2021-12-07 962 bool caching = ceph_is_cache_enabled(inode);
7d41870d65db028 Xiubo Li 2023-03-08 963 xa_mark_t tag;
1d3576fd10f0d7a Sage Weil 2009-10-06 964
503d4fa6ee28e8d NeilBrown 2022-03-22 965 if (wbc->sync_mode == WB_SYNC_NONE &&
503d4fa6ee28e8d NeilBrown 2022-03-22 966 fsc->write_congested)
503d4fa6ee28e8d NeilBrown 2022-03-22 967 return 0;
503d4fa6ee28e8d NeilBrown 2022-03-22 968
38d46409c4639a1 Xiubo Li 2023-06-12 969 doutc(cl, "%llx.%llx (mode=%s)\n", ceph_vinop(inode),
1d3576fd10f0d7a Sage Weil 2009-10-06 970 wbc->sync_mode == WB_SYNC_NONE ? "NONE" :
1d3576fd10f0d7a Sage Weil 2009-10-06 971 (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD"));
1d3576fd10f0d7a Sage Weil 2009-10-06 972
5d6451b1489ad17 Jeff Layton 2021-08-31 973 if (ceph_inode_is_shutdown(inode)) {
6c93df5db628e71 Yan, Zheng 2016-04-15 974 if (ci->i_wrbuffer_ref > 0) {
38d46409c4639a1 Xiubo Li 2023-06-12 975 pr_warn_ratelimited_client(cl,
38d46409c4639a1 Xiubo Li 2023-06-12 976 "%llx.%llx %lld forced umount\n",
38d46409c4639a1 Xiubo Li 2023-06-12 977 ceph_vinop(inode), ceph_ino(inode));
6c93df5db628e71 Yan, Zheng 2016-04-15 978 }
a341d4df87487ae Yan, Zheng 2015-07-01 979 mapping_set_error(mapping, -EIO);
1d3576fd10f0d7a Sage Weil 2009-10-06 980 return -EIO; /* we're in a forced umount, don't write! */
1d3576fd10f0d7a Sage Weil 2009-10-06 981 }
95cca2b44e54b00 Yan, Zheng 2017-07-11 982 if (fsc->mount_options->wsize < wsize)
3d14c5d2b6e15c2 Yehuda Sadeh 2010-04-06 983 wsize = fsc->mount_options->wsize;
1d3576fd10f0d7a Sage Weil 2009-10-06 984
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 985) folio_batch_init(&fbatch);
1d3576fd10f0d7a Sage Weil 2009-10-06 986
590e9d9861f5f21 Yan, Zheng 2017-09-03 987 start_index = wbc->range_cyclic ? mapping->writeback_index : 0;
590e9d9861f5f21 Yan, Zheng 2017-09-03 988 index = start_index;
1d3576fd10f0d7a Sage Weil 2009-10-06 989
7d41870d65db028 Xiubo Li 2023-03-08 990 if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) {
7d41870d65db028 Xiubo Li 2023-03-08 991 tag = PAGECACHE_TAG_TOWRITE;
7d41870d65db028 Xiubo Li 2023-03-08 992 } else {
7d41870d65db028 Xiubo Li 2023-03-08 993 tag = PAGECACHE_TAG_DIRTY;
7d41870d65db028 Xiubo Li 2023-03-08 994 }
1d3576fd10f0d7a Sage Weil 2009-10-06 995 retry:
1d3576fd10f0d7a Sage Weil 2009-10-06 996 /* find oldest snap context with dirty data */
05455e1177f7684 Yan, Zheng 2017-09-02 997 snapc = get_oldest_context(inode, &ceph_wbc, NULL);
1d3576fd10f0d7a Sage Weil 2009-10-06 998 if (!snapc) {
1d3576fd10f0d7a Sage Weil 2009-10-06 999 /* hmm, why does writepages get called when there
1d3576fd10f0d7a Sage Weil 2009-10-06 1000 is no dirty data? */
38d46409c4639a1 Xiubo Li 2023-06-12 1001 doutc(cl, " no snap context with dirty data?\n");
1d3576fd10f0d7a Sage Weil 2009-10-06 1002 goto out;
1d3576fd10f0d7a Sage Weil 2009-10-06 1003 }
38d46409c4639a1 Xiubo Li 2023-06-12 1004 doutc(cl, " oldest snapc is %p seq %lld (%d snaps)\n", snapc,
38d46409c4639a1 Xiubo Li 2023-06-12 1005 snapc->seq, snapc->num_snaps);
fc2744aa12da718 Yan, Zheng 2013-05-31 1006
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1007 should_loop = false;
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1008 if (ceph_wbc.head_snapc && snapc != last_snapc) {
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1009 /* where to start/end? */
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1010 if (wbc->range_cyclic) {
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1011 index = start_index;
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1012 end = -1;
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1013 if (index > 0)
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1014 should_loop = true;
38d46409c4639a1 Xiubo Li 2023-06-12 1015 doutc(cl, " cyclic, start at %lu\n", index);
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1016 } else {
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1017 index = wbc->range_start >> PAGE_SHIFT;
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1018 end = wbc->range_end >> PAGE_SHIFT;
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1019 if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1020 range_whole = true;
38d46409c4639a1 Xiubo Li 2023-06-12 1021 doutc(cl, " not cyclic, %lu to %lu\n", index, end);
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1022 }
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1023 } else if (!ceph_wbc.head_snapc) {
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1024 /* Do not respect wbc->range_{start,end}. Dirty pages
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1025 * in that range can be associated with newer snapc.
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1026 * They are not writeable until we write all dirty pages
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1027 * associated with 'snapc' get written */
1582af2eaaf17cb Yan, Zheng 2018-03-06 1028 if (index > 0)
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1029 should_loop = true;
38d46409c4639a1 Xiubo Li 2023-06-12 1030 doutc(cl, " non-head snapc, range whole\n");
1d3576fd10f0d7a Sage Weil 2009-10-06 1031 }
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1032
7d41870d65db028 Xiubo Li 2023-03-08 1033 if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
7d41870d65db028 Xiubo Li 2023-03-08 1034 tag_pages_for_writeback(mapping, index, end);
7d41870d65db028 Xiubo Li 2023-03-08 1035
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1036 ceph_put_snap_context(last_snapc);
1d3576fd10f0d7a Sage Weil 2009-10-06 1037 last_snapc = snapc;
1d3576fd10f0d7a Sage Weil 2009-10-06 1038
af9cc401ce7452f Yan, Zheng 2018-03-04 1039 while (!done && index <= end) {
5b64640cf65be4a Yan, Zheng 2016-01-07 1040 int num_ops = 0, op_idx;
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1041) unsigned i, nr_folios, max_pages, locked_pages = 0;
5b64640cf65be4a Yan, Zheng 2016-01-07 1042 struct page **pages = NULL, **data_pages;
1d3576fd10f0d7a Sage Weil 2009-10-06 1043 struct page *page;
0e5ecac71683665 Yan, Zheng 2017-08-31 1044 pgoff_t strip_unit_end = 0;
5b64640cf65be4a Yan, Zheng 2016-01-07 1045 u64 offset = 0, len = 0;
a0102bda5bc0991 Jeff Layton 2020-07-30 1046 bool from_pool = false;
1d3576fd10f0d7a Sage Weil 2009-10-06 1047
0e5ecac71683665 Yan, Zheng 2017-08-31 1048 max_pages = wsize >> PAGE_SHIFT;
1d3576fd10f0d7a Sage Weil 2009-10-06 1049
1d3576fd10f0d7a Sage Weil 2009-10-06 1050 get_more_pages:
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1051) nr_folios = filemap_get_folios_tag(mapping, &index,
7d41870d65db028 Xiubo Li 2023-03-08 1052 end, tag, &fbatch);
38d46409c4639a1 Xiubo Li 2023-06-12 1053 doutc(cl, "pagevec_lookup_range_tag got %d\n", nr_folios);
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1054) if (!nr_folios && !locked_pages)
1d3576fd10f0d7a Sage Weil 2009-10-06 1055 break;
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1056) for (i = 0; i < nr_folios && locked_pages < max_pages; i++) {
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1057) struct folio *folio = fbatch.folios[i];
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1058)
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1059) doutc(cl, "? %p idx %lu\n", folio, folio->index);
1d3576fd10f0d7a Sage Weil 2009-10-06 1060 if (locked_pages == 0)
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1061) folio_lock(folio); /* first page */
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1062) else if (!folio_trylock(folio))
1d3576fd10f0d7a Sage Weil 2009-10-06 1063 break;
1d3576fd10f0d7a Sage Weil 2009-10-06 1064
1d3576fd10f0d7a Sage Weil 2009-10-06 1065 /* only dirty pages, or our accounting breaks */
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1066) if (unlikely(!folio_test_dirty(folio)) ||
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1067) unlikely(folio->mapping != mapping)) {
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1068) doutc(cl, "!dirty or !mapping %p\n", folio);
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1069) folio_unlock(folio);
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1070 continue;
1d3576fd10f0d7a Sage Weil 2009-10-06 1071 }
af9cc401ce7452f Yan, Zheng 2018-03-04 1072 /* only if matching snap context */
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1073) pgsnapc = page_snap_context(&folio->page);
af9cc401ce7452f Yan, Zheng 2018-03-04 1074 if (pgsnapc != snapc) {
38d46409c4639a1 Xiubo Li 2023-06-12 1075 doutc(cl, "page snapc %p %lld != oldest %p %lld\n",
af9cc401ce7452f Yan, Zheng 2018-03-04 1076 pgsnapc, pgsnapc->seq, snapc, snapc->seq);
1582af2eaaf17cb Yan, Zheng 2018-03-06 1077 if (!should_loop &&
1582af2eaaf17cb Yan, Zheng 2018-03-06 1078 !ceph_wbc.head_snapc &&
1582af2eaaf17cb Yan, Zheng 2018-03-06 1079 wbc->sync_mode != WB_SYNC_NONE)
1582af2eaaf17cb Yan, Zheng 2018-03-06 1080 should_loop = true;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1081) folio_unlock(folio);
af9cc401ce7452f Yan, Zheng 2018-03-04 1082 continue;
1d3576fd10f0d7a Sage Weil 2009-10-06 1083 }
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1084) if (folio_pos(folio) >= ceph_wbc.i_size) {
38d46409c4639a1 Xiubo Li 2023-06-12 1085 doutc(cl, "folio at %lu beyond eof %llu\n",
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1086) folio->index, ceph_wbc.i_size);
c95f1c5f436badb Erqi Chen 2019-07-24 1087 if ((ceph_wbc.size_stable ||
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1088) folio_pos(folio) >= i_size_read(inode)) &&
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1089) folio_clear_dirty_for_io(folio))
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1090) folio_invalidate(folio, 0,
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1091) folio_size(folio));
a628304ebe6ab02 Matthew Wilcox (Oracle 2022-02-09 1092) folio_unlock(folio);
af9cc401ce7452f Yan, Zheng 2018-03-04 1093 continue;
af9cc401ce7452f Yan, Zheng 2018-03-04 1094 }
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1095) if (strip_unit_end && (folio->index > strip_unit_end)) {
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1096) doutc(cl, "end of strip unit %p\n", folio);
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1097) folio_unlock(folio);
1d3576fd10f0d7a Sage Weil 2009-10-06 1098 break;
1d3576fd10f0d7a Sage Weil 2009-10-06 1099 }
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1100) if (folio_test_writeback(folio) ||
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1101) folio_test_private_2(folio) /* [DEPRECATED] */) {
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1102 if (wbc->sync_mode == WB_SYNC_NONE) {
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1103) doutc(cl, "%p under writeback\n", folio);
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1104) folio_unlock(folio);
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1105 continue;
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1106 }
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1107) doutc(cl, "waiting on writeback %p\n", folio);
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1108) folio_wait_writeback(folio);
fd15ba4cb00a43f Matthew Wilcox (Oracle 2024-10-02 1109) folio_wait_private_2(folio); /* [DEPRECATED] */
1d3576fd10f0d7a Sage Weil 2009-10-06 1110 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1111
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1112) if (!folio_clear_dirty_for_io(folio)) {
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1113) doutc(cl, "%p !clear_page_dirty_for_io\n", folio);
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1114) folio_unlock(folio);
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1115 continue;
1d3576fd10f0d7a Sage Weil 2009-10-06 1116 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1117
e5975c7c8eb6aea Alex Elder 2013-03-14 1118 /*
e5975c7c8eb6aea Alex Elder 2013-03-14 1119 * We have something to write. If this is
e5975c7c8eb6aea Alex Elder 2013-03-14 1120 * the first locked page this time through,
5b64640cf65be4a Yan, Zheng 2016-01-07 1121 * calculate max possinle write size and
5b64640cf65be4a Yan, Zheng 2016-01-07 1122 * allocate a page array
e5975c7c8eb6aea Alex Elder 2013-03-14 1123 */
1d3576fd10f0d7a Sage Weil 2009-10-06 1124 if (locked_pages == 0) {
5b64640cf65be4a Yan, Zheng 2016-01-07 1125 u64 objnum;
5b64640cf65be4a Yan, Zheng 2016-01-07 1126 u64 objoff;
dccbf08005df800 Ilya Dryomov 2018-02-17 1127 u32 xlen;
5b64640cf65be4a Yan, Zheng 2016-01-07 1128
1d3576fd10f0d7a Sage Weil 2009-10-06 1129 /* prepare async write request */
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1130) offset = folio_pos(folio);
dccbf08005df800 Ilya Dryomov 2018-02-17 1131 ceph_calc_file_object_mapping(&ci->i_layout,
dccbf08005df800 Ilya Dryomov 2018-02-17 1132 offset, wsize,
5b64640cf65be4a Yan, Zheng 2016-01-07 1133 &objnum, &objoff,
dccbf08005df800 Ilya Dryomov 2018-02-17 1134 &xlen);
dccbf08005df800 Ilya Dryomov 2018-02-17 1135 len = xlen;
8c71897be2ddfd8 Henry C Chang 2011-05-03 1136
3fb99d483e614bc Yanhu Cao 2017-07-21 1137 num_ops = 1;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1138) strip_unit_end = folio->index +
09cbfeaf1a5a67b Kirill A. Shutemov 2016-04-01 1139 ((len - 1) >> PAGE_SHIFT);
88486957f9fbf52 Alex Elder 2013-03-14 1140
5b64640cf65be4a Yan, Zheng 2016-01-07 1141 BUG_ON(pages);
88486957f9fbf52 Alex Elder 2013-03-14 1142 max_pages = calc_pages_for(0, (u64)len);
6da2ec56059c3c7 Kees Cook 2018-06-12 1143 pages = kmalloc_array(max_pages,
6da2ec56059c3c7 Kees Cook 2018-06-12 1144 sizeof(*pages),
fc2744aa12da718 Yan, Zheng 2013-05-31 1145 GFP_NOFS);
88486957f9fbf52 Alex Elder 2013-03-14 1146 if (!pages) {
a0102bda5bc0991 Jeff Layton 2020-07-30 1147 from_pool = true;
a0102bda5bc0991 Jeff Layton 2020-07-30 1148 pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS);
e5975c7c8eb6aea Alex Elder 2013-03-14 1149 BUG_ON(!pages);
88486957f9fbf52 Alex Elder 2013-03-14 1150 }
5b64640cf65be4a Yan, Zheng 2016-01-07 1151
5b64640cf65be4a Yan, Zheng 2016-01-07 1152 len = 0;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1153) } else if (folio->index !=
09cbfeaf1a5a67b Kirill A. Shutemov 2016-04-01 1154 (offset + len) >> PAGE_SHIFT) {
a0102bda5bc0991 Jeff Layton 2020-07-30 1155 if (num_ops >= (from_pool ? CEPH_OSD_SLAB_OPS :
5b64640cf65be4a Yan, Zheng 2016-01-07 1156 CEPH_OSD_MAX_OPS)) {
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1157) folio_redirty_for_writepage(wbc, folio);
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1158) folio_unlock(folio);
5b64640cf65be4a Yan, Zheng 2016-01-07 1159 break;
5b64640cf65be4a Yan, Zheng 2016-01-07 1160 }
5b64640cf65be4a Yan, Zheng 2016-01-07 1161
5b64640cf65be4a Yan, Zheng 2016-01-07 1162 num_ops++;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1163) offset = folio_pos(folio);
5b64640cf65be4a Yan, Zheng 2016-01-07 1164 len = 0;
1d3576fd10f0d7a Sage Weil 2009-10-06 1165 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1166
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1167) /* note position of first page in fbatch */
38d46409c4639a1 Xiubo Li 2023-06-12 1168 doutc(cl, "%llx.%llx will write page %p idx %lu\n",
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1169) ceph_vinop(inode), folio, folio->index);
2baba25019ec564 Yehuda Sadeh 2009-12-18 1170
5b64640cf65be4a Yan, Zheng 2016-01-07 1171 if (atomic_long_inc_return(&fsc->writeback_count) >
5b64640cf65be4a Yan, Zheng 2016-01-07 1172 CONGESTION_ON_THRESH(
503d4fa6ee28e8d NeilBrown 2022-03-22 1173 fsc->mount_options->congestion_kb))
503d4fa6ee28e8d NeilBrown 2022-03-22 1174 fsc->write_congested = true;
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1175
d55207717ded95c Jeff Layton 2022-08-25 1176 if (IS_ENCRYPTED(inode)) {
d55207717ded95c Jeff Layton 2022-08-25 1177 pages[locked_pages] =
d55207717ded95c Jeff Layton 2022-08-25 @1178 fscrypt_encrypt_pagecache_blocks(page,
d55207717ded95c Jeff Layton 2022-08-25 1179 PAGE_SIZE, 0,
d55207717ded95c Jeff Layton 2022-08-25 1180 locked_pages ? GFP_NOWAIT : GFP_NOFS);
d55207717ded95c Jeff Layton 2022-08-25 1181 if (IS_ERR(pages[locked_pages])) {
d55207717ded95c Jeff Layton 2022-08-25 1182 if (PTR_ERR(pages[locked_pages]) == -EINVAL)
38d46409c4639a1 Xiubo Li 2023-06-12 1183 pr_err_client(cl,
38d46409c4639a1 Xiubo Li 2023-06-12 1184 "inode->i_blkbits=%hhu\n",
38d46409c4639a1 Xiubo Li 2023-06-12 1185 inode->i_blkbits);
d55207717ded95c Jeff Layton 2022-08-25 1186 /* better not fail on first page! */
d55207717ded95c Jeff Layton 2022-08-25 1187 BUG_ON(locked_pages == 0);
d55207717ded95c Jeff Layton 2022-08-25 1188 pages[locked_pages] = NULL;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1189) folio_redirty_for_writepage(wbc, folio);
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1190) folio_unlock(folio);
d55207717ded95c Jeff Layton 2022-08-25 1191 break;
d55207717ded95c Jeff Layton 2022-08-25 1192 }
d55207717ded95c Jeff Layton 2022-08-25 1193 ++locked_pages;
d55207717ded95c Jeff Layton 2022-08-25 1194 } else {
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1195) pages[locked_pages++] = &folio->page;
d55207717ded95c Jeff Layton 2022-08-25 1196 }
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1197
d55207717ded95c Jeff Layton 2022-08-25 1198 fbatch.folios[i] = NULL;
a38da1eafb17612 Matthew Wilcox (Oracle 2024-11-29 1199) len += folio_size(folio);
1d3576fd10f0d7a Sage Weil 2009-10-06 1200 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1201
1d3576fd10f0d7a Sage Weil 2009-10-06 1202 /* did we get anything? */
1d3576fd10f0d7a Sage Weil 2009-10-06 1203 if (!locked_pages)
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1204) goto release_folios;
1d3576fd10f0d7a Sage Weil 2009-10-06 1205 if (i) {
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1206 unsigned j, n = 0;
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1207) /* shift unused page to beginning of fbatch */
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1208) for (j = 0; j < nr_folios; j++) {
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1209) if (!fbatch.folios[j])
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1210 continue;
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1211 if (n < j)
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1212) fbatch.folios[n] = fbatch.folios[j];
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1213 n++;
0713e5f24b7deb8 Yan, Zheng 2017-08-31 1214 }
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1215) fbatch.nr = n;
1d3576fd10f0d7a Sage Weil 2009-10-06 1216
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1217) if (nr_folios && i == nr_folios &&
1d3576fd10f0d7a Sage Weil 2009-10-06 1218 locked_pages < max_pages) {
38d46409c4639a1 Xiubo Li 2023-06-12 1219 doutc(cl, "reached end fbatch, trying for more\n");
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1220) folio_batch_release(&fbatch);
1d3576fd10f0d7a Sage Weil 2009-10-06 1221 goto get_more_pages;
1d3576fd10f0d7a Sage Weil 2009-10-06 1222 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1223 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1224
5b64640cf65be4a Yan, Zheng 2016-01-07 1225 new_request:
d55207717ded95c Jeff Layton 2022-08-25 1226 offset = ceph_fscrypt_page_offset(pages[0]);
5b64640cf65be4a Yan, Zheng 2016-01-07 1227 len = wsize;
5b64640cf65be4a Yan, Zheng 2016-01-07 1228
5b64640cf65be4a Yan, Zheng 2016-01-07 1229 req = ceph_osdc_new_request(&fsc->client->osdc,
5b64640cf65be4a Yan, Zheng 2016-01-07 1230 &ci->i_layout, vino,
5b64640cf65be4a Yan, Zheng 2016-01-07 1231 offset, &len, 0, num_ops,
1f934b00e907527 Yan, Zheng 2017-08-30 1232 CEPH_OSD_OP_WRITE, CEPH_OSD_FLAG_WRITE,
1f934b00e907527 Yan, Zheng 2017-08-30 1233 snapc, ceph_wbc.truncate_seq,
1f934b00e907527 Yan, Zheng 2017-08-30 1234 ceph_wbc.truncate_size, false);
5b64640cf65be4a Yan, Zheng 2016-01-07 1235 if (IS_ERR(req)) {
5b64640cf65be4a Yan, Zheng 2016-01-07 1236 req = ceph_osdc_new_request(&fsc->client->osdc,
5b64640cf65be4a Yan, Zheng 2016-01-07 1237 &ci->i_layout, vino,
5b64640cf65be4a Yan, Zheng 2016-01-07 1238 offset, &len, 0,
5b64640cf65be4a Yan, Zheng 2016-01-07 1239 min(num_ops,
5b64640cf65be4a Yan, Zheng 2016-01-07 1240 CEPH_OSD_SLAB_OPS),
5b64640cf65be4a Yan, Zheng 2016-01-07 1241 CEPH_OSD_OP_WRITE,
54ea0046b6fe36e Ilya Dryomov 2017-02-11 1242 CEPH_OSD_FLAG_WRITE,
1f934b00e907527 Yan, Zheng 2017-08-30 1243 snapc, ceph_wbc.truncate_seq,
1f934b00e907527 Yan, Zheng 2017-08-30 1244 ceph_wbc.truncate_size, true);
5b64640cf65be4a Yan, Zheng 2016-01-07 1245 BUG_ON(IS_ERR(req));
5b64640cf65be4a Yan, Zheng 2016-01-07 1246 }
d55207717ded95c Jeff Layton 2022-08-25 1247 BUG_ON(len < ceph_fscrypt_page_offset(pages[locked_pages - 1]) +
d55207717ded95c Jeff Layton 2022-08-25 1248 thp_size(pages[locked_pages - 1]) - offset);
5b64640cf65be4a Yan, Zheng 2016-01-07 1249
1464de9f813e355 Xiubo Li 2023-05-09 1250 if (!ceph_inc_osd_stopping_blocker(fsc->mdsc)) {
1464de9f813e355 Xiubo Li 2023-05-09 1251 rc = -EIO;
1464de9f813e355 Xiubo Li 2023-05-09 1252 goto release_folios;
1464de9f813e355 Xiubo Li 2023-05-09 1253 }
5b64640cf65be4a Yan, Zheng 2016-01-07 1254 req->r_callback = writepages_finish;
5b64640cf65be4a Yan, Zheng 2016-01-07 1255 req->r_inode = inode;
5b64640cf65be4a Yan, Zheng 2016-01-07 1256
5b64640cf65be4a Yan, Zheng 2016-01-07 1257 /* Format the osd request message and submit the write */
5b64640cf65be4a Yan, Zheng 2016-01-07 1258 len = 0;
5b64640cf65be4a Yan, Zheng 2016-01-07 1259 data_pages = pages;
5b64640cf65be4a Yan, Zheng 2016-01-07 1260 op_idx = 0;
5b64640cf65be4a Yan, Zheng 2016-01-07 1261 for (i = 0; i < locked_pages; i++) {
d55207717ded95c Jeff Layton 2022-08-25 1262 struct page *page = ceph_fscrypt_pagecache_page(pages[i]);
d55207717ded95c Jeff Layton 2022-08-25 1263
d55207717ded95c Jeff Layton 2022-08-25 1264 u64 cur_offset = page_offset(page);
1702e79734104d7 Jeff Layton 2021-12-07 1265 /*
1702e79734104d7 Jeff Layton 2021-12-07 1266 * Discontinuity in page range? Ceph can handle that by just passing
1702e79734104d7 Jeff Layton 2021-12-07 1267 * multiple extents in the write op.
1702e79734104d7 Jeff Layton 2021-12-07 1268 */
5b64640cf65be4a Yan, Zheng 2016-01-07 1269 if (offset + len != cur_offset) {
1702e79734104d7 Jeff Layton 2021-12-07 1270 /* If it's full, stop here */
3fb99d483e614bc Yanhu Cao 2017-07-21 1271 if (op_idx + 1 == req->r_num_ops)
5b64640cf65be4a Yan, Zheng 2016-01-07 1272 break;
1702e79734104d7 Jeff Layton 2021-12-07 1273
1702e79734104d7 Jeff Layton 2021-12-07 1274 /* Kick off an fscache write with what we have so far. */
1702e79734104d7 Jeff Layton 2021-12-07 1275 ceph_fscache_write_to_cache(inode, offset, len, caching);
1702e79734104d7 Jeff Layton 2021-12-07 1276
1702e79734104d7 Jeff Layton 2021-12-07 1277 /* Start a new extent */
5b64640cf65be4a Yan, Zheng 2016-01-07 1278 osd_req_op_extent_dup_last(req, op_idx,
5b64640cf65be4a Yan, Zheng 2016-01-07 1279 cur_offset - offset);
38d46409c4639a1 Xiubo Li 2023-06-12 1280 doutc(cl, "got pages at %llu~%llu\n", offset,
38d46409c4639a1 Xiubo Li 2023-06-12 1281 len);
5b64640cf65be4a Yan, Zheng 2016-01-07 1282 osd_req_op_extent_osd_data_pages(req, op_idx,
5b64640cf65be4a Yan, Zheng 2016-01-07 1283 data_pages, len, 0,
a0102bda5bc0991 Jeff Layton 2020-07-30 1284 from_pool, false);
5b64640cf65be4a Yan, Zheng 2016-01-07 1285 osd_req_op_extent_update(req, op_idx, len);
5b64640cf65be4a Yan, Zheng 2016-01-07 1286
5b64640cf65be4a Yan, Zheng 2016-01-07 1287 len = 0;
5b64640cf65be4a Yan, Zheng 2016-01-07 1288 offset = cur_offset;
5b64640cf65be4a Yan, Zheng 2016-01-07 1289 data_pages = pages + i;
5b64640cf65be4a Yan, Zheng 2016-01-07 1290 op_idx++;
5b64640cf65be4a Yan, Zheng 2016-01-07 1291 }
5b64640cf65be4a Yan, Zheng 2016-01-07 1292
d55207717ded95c Jeff Layton 2022-08-25 1293 set_page_writeback(page);
8e5ced7804cb918 David Howells 2024-07-30 1294 if (caching)
8e5ced7804cb918 David Howells 2024-07-30 1295 ceph_set_page_fscache(page);
8ff2d290c8ce77c Jeff Layton 2021-04-05 1296 len += thp_size(page);
5b64640cf65be4a Yan, Zheng 2016-01-07 1297 }
1702e79734104d7 Jeff Layton 2021-12-07 1298 ceph_fscache_write_to_cache(inode, offset, len, caching);
5b64640cf65be4a Yan, Zheng 2016-01-07 1299
1f934b00e907527 Yan, Zheng 2017-08-30 1300 if (ceph_wbc.size_stable) {
1f934b00e907527 Yan, Zheng 2017-08-30 1301 len = min(len, ceph_wbc.i_size - offset);
5b64640cf65be4a Yan, Zheng 2016-01-07 1302 } else if (i == locked_pages) {
e1966b49446a439 Yan, Zheng 2015-06-18 1303 /* writepages_finish() clears writeback pages
e1966b49446a439 Yan, Zheng 2015-06-18 1304 * according to the data length, so make sure
e1966b49446a439 Yan, Zheng 2015-06-18 1305 * data length covers all locked pages */
8ff2d290c8ce77c Jeff Layton 2021-04-05 1306 u64 min_len = len + 1 - thp_size(page);
1f934b00e907527 Yan, Zheng 2017-08-30 1307 len = get_writepages_data_length(inode, pages[i - 1],
1f934b00e907527 Yan, Zheng 2017-08-30 1308 offset);
5b64640cf65be4a Yan, Zheng 2016-01-07 1309 len = max(len, min_len);
e1966b49446a439 Yan, Zheng 2015-06-18 1310 }
d55207717ded95c Jeff Layton 2022-08-25 1311 if (IS_ENCRYPTED(inode))
d55207717ded95c Jeff Layton 2022-08-25 1312 len = round_up(len, CEPH_FSCRYPT_BLOCK_SIZE);
d55207717ded95c Jeff Layton 2022-08-25 1313
38d46409c4639a1 Xiubo Li 2023-06-12 1314 doutc(cl, "got pages at %llu~%llu\n", offset, len);
1d3576fd10f0d7a Sage Weil 2009-10-06 1315
d55207717ded95c Jeff Layton 2022-08-25 1316 if (IS_ENCRYPTED(inode) &&
d55207717ded95c Jeff Layton 2022-08-25 1317 ((offset | len) & ~CEPH_FSCRYPT_BLOCK_MASK))
38d46409c4639a1 Xiubo Li 2023-06-12 1318 pr_warn_client(cl,
38d46409c4639a1 Xiubo Li 2023-06-12 1319 "bad encrypted write offset=%lld len=%llu\n",
38d46409c4639a1 Xiubo Li 2023-06-12 1320 offset, len);
d55207717ded95c Jeff Layton 2022-08-25 1321
5b64640cf65be4a Yan, Zheng 2016-01-07 1322 osd_req_op_extent_osd_data_pages(req, op_idx, data_pages, len,
a0102bda5bc0991 Jeff Layton 2020-07-30 1323 0, from_pool, false);
5b64640cf65be4a Yan, Zheng 2016-01-07 1324 osd_req_op_extent_update(req, op_idx, len);
e5975c7c8eb6aea Alex Elder 2013-03-14 1325
5b64640cf65be4a Yan, Zheng 2016-01-07 1326 BUG_ON(op_idx + 1 != req->r_num_ops);
e5975c7c8eb6aea Alex Elder 2013-03-14 1327
a0102bda5bc0991 Jeff Layton 2020-07-30 1328 from_pool = false;
5b64640cf65be4a Yan, Zheng 2016-01-07 1329 if (i < locked_pages) {
5b64640cf65be4a Yan, Zheng 2016-01-07 1330 BUG_ON(num_ops <= req->r_num_ops);
5b64640cf65be4a Yan, Zheng 2016-01-07 1331 num_ops -= req->r_num_ops;
5b64640cf65be4a Yan, Zheng 2016-01-07 1332 locked_pages -= i;
5b64640cf65be4a Yan, Zheng 2016-01-07 1333
5b64640cf65be4a Yan, Zheng 2016-01-07 1334 /* allocate new pages array for next request */
5b64640cf65be4a Yan, Zheng 2016-01-07 1335 data_pages = pages;
6da2ec56059c3c7 Kees Cook 2018-06-12 1336 pages = kmalloc_array(locked_pages, sizeof(*pages),
5b64640cf65be4a Yan, Zheng 2016-01-07 1337 GFP_NOFS);
5b64640cf65be4a Yan, Zheng 2016-01-07 1338 if (!pages) {
a0102bda5bc0991 Jeff Layton 2020-07-30 1339 from_pool = true;
a0102bda5bc0991 Jeff Layton 2020-07-30 1340 pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS);
5b64640cf65be4a Yan, Zheng 2016-01-07 1341 BUG_ON(!pages);
5b64640cf65be4a Yan, Zheng 2016-01-07 1342 }
5b64640cf65be4a Yan, Zheng 2016-01-07 1343 memcpy(pages, data_pages + i,
5b64640cf65be4a Yan, Zheng 2016-01-07 1344 locked_pages * sizeof(*pages));
5b64640cf65be4a Yan, Zheng 2016-01-07 1345 memset(data_pages + i, 0,
5b64640cf65be4a Yan, Zheng 2016-01-07 1346 locked_pages * sizeof(*pages));
5b64640cf65be4a Yan, Zheng 2016-01-07 1347 } else {
9182bf4af1f077d Matthew Wilcox (Oracle 2024-11-29 1348) struct folio *folio;
9182bf4af1f077d Matthew Wilcox (Oracle 2024-11-29 1349)
5b64640cf65be4a Yan, Zheng 2016-01-07 1350 BUG_ON(num_ops != req->r_num_ops);
9182bf4af1f077d Matthew Wilcox (Oracle 2024-11-29 1351) folio = ceph_fscrypt_pagecache_folio(pages[i - 1]);
9182bf4af1f077d Matthew Wilcox (Oracle 2024-11-29 1352) index = folio->index + 1;
5b64640cf65be4a Yan, Zheng 2016-01-07 1353 /* request message now owns the pages array */
5b64640cf65be4a Yan, Zheng 2016-01-07 1354 pages = NULL;
5b64640cf65be4a Yan, Zheng 2016-01-07 1355 }
e5975c7c8eb6aea Alex Elder 2013-03-14 1356
c453bdb535341b5 Jeff Layton 2023-10-04 1357 req->r_mtime = inode_get_mtime(inode);
a8af0d682ae0c9c Jeff Layton 2022-06-30 1358 ceph_osdc_start_request(&fsc->client->osdc, req);
1d3576fd10f0d7a Sage Weil 2009-10-06 1359 req = NULL;
1d3576fd10f0d7a Sage Weil 2009-10-06 1360
5b64640cf65be4a Yan, Zheng 2016-01-07 1361 wbc->nr_to_write -= i;
5b64640cf65be4a Yan, Zheng 2016-01-07 1362 if (pages)
5b64640cf65be4a Yan, Zheng 2016-01-07 1363 goto new_request;
5b64640cf65be4a Yan, Zheng 2016-01-07 1364
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1365 /*
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1366 * We stop writing back only if we are not doing
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1367 * integrity sync. In case of integrity sync we have to
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1368 * keep going until we have written all the pages
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1369 * we tagged for writeback prior to entering this loop.
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1370 */
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1371 if (wbc->nr_to_write <= 0 && wbc->sync_mode == WB_SYNC_NONE)
af9cc401ce7452f Yan, Zheng 2018-03-04 1372 done = true;
1d3576fd10f0d7a Sage Weil 2009-10-06 1373
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1374) release_folios:
38d46409c4639a1 Xiubo Li 2023-06-12 1375 doutc(cl, "folio_batch release on %d folios (%p)\n",
38d46409c4639a1 Xiubo Li 2023-06-12 1376 (int)fbatch.nr, fbatch.nr ? fbatch.folios[0] : NULL);
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1377) folio_batch_release(&fbatch);
1d3576fd10f0d7a Sage Weil 2009-10-06 1378 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1379
1d3576fd10f0d7a Sage Weil 2009-10-06 1380 if (should_loop && !done) {
1d3576fd10f0d7a Sage Weil 2009-10-06 1381 /* more to do; loop back to beginning of file */
38d46409c4639a1 Xiubo Li 2023-06-12 1382 doutc(cl, "looping back to beginning of file\n");
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1383 end = start_index - 1; /* OK even when start_index == 0 */
f275635ee0b6641 Yan, Zheng 2017-09-01 1384
f275635ee0b6641 Yan, Zheng 2017-09-01 1385 /* to write dirty pages associated with next snapc,
f275635ee0b6641 Yan, Zheng 2017-09-01 1386 * we need to wait until current writes complete */
f275635ee0b6641 Yan, Zheng 2017-09-01 1387 if (wbc->sync_mode != WB_SYNC_NONE &&
f275635ee0b6641 Yan, Zheng 2017-09-01 1388 start_index == 0 && /* all dirty pages were checked */
f275635ee0b6641 Yan, Zheng 2017-09-01 1389 !ceph_wbc.head_snapc) {
f275635ee0b6641 Yan, Zheng 2017-09-01 1390 struct page *page;
f275635ee0b6641 Yan, Zheng 2017-09-01 1391 unsigned i, nr;
f275635ee0b6641 Yan, Zheng 2017-09-01 1392 index = 0;
f275635ee0b6641 Yan, Zheng 2017-09-01 1393 while ((index <= end) &&
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1394) (nr = filemap_get_folios_tag(mapping, &index,
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1395) (pgoff_t)-1,
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1396) PAGECACHE_TAG_WRITEBACK,
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1397) &fbatch))) {
f275635ee0b6641 Yan, Zheng 2017-09-01 1398 for (i = 0; i < nr; i++) {
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1399) page = &fbatch.folios[i]->page;
f275635ee0b6641 Yan, Zheng 2017-09-01 1400 if (page_snap_context(page) != snapc)
f275635ee0b6641 Yan, Zheng 2017-09-01 1401 continue;
f275635ee0b6641 Yan, Zheng 2017-09-01 1402 wait_on_page_writeback(page);
f275635ee0b6641 Yan, Zheng 2017-09-01 1403 }
590a2b5f0a9b740 Vishal Moola (Oracle 2023-01-04 1404) folio_batch_release(&fbatch);
f275635ee0b6641 Yan, Zheng 2017-09-01 1405 cond_resched();
f275635ee0b6641 Yan, Zheng 2017-09-01 1406 }
f275635ee0b6641 Yan, Zheng 2017-09-01 1407 }
f275635ee0b6641 Yan, Zheng 2017-09-01 1408
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1409 start_index = 0;
1d3576fd10f0d7a Sage Weil 2009-10-06 1410 index = 0;
1d3576fd10f0d7a Sage Weil 2009-10-06 1411 goto retry;
1d3576fd10f0d7a Sage Weil 2009-10-06 1412 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1413
1d3576fd10f0d7a Sage Weil 2009-10-06 1414 if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
1d3576fd10f0d7a Sage Weil 2009-10-06 1415 mapping->writeback_index = index;
1d3576fd10f0d7a Sage Weil 2009-10-06 1416
1d3576fd10f0d7a Sage Weil 2009-10-06 1417 out:
1d3576fd10f0d7a Sage Weil 2009-10-06 1418 ceph_osdc_put_request(req);
2a2d927e35dd8dc Yan, Zheng 2017-09-01 1419 ceph_put_snap_context(last_snapc);
38d46409c4639a1 Xiubo Li 2023-06-12 1420 doutc(cl, "%llx.%llx dend - startone, rc = %d\n", ceph_vinop(inode),
38d46409c4639a1 Xiubo Li 2023-06-12 1421 rc);
1d3576fd10f0d7a Sage Weil 2009-10-06 1422 return rc;
1d3576fd10f0d7a Sage Weil 2009-10-06 1423 }
1d3576fd10f0d7a Sage Weil 2009-10-06 1424
On Fri, Nov 29, 2024 at 11:42:44PM +0800, kernel test robot wrote: > >> fs/ceph/addr.c:1178:39: warning: variable 'page' is uninitialized when used here [-Wuninitialized] > 1178 | fscrypt_encrypt_pagecache_blocks(page, This will fix the bug. I'll send a v2 with this folded in. diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 1e0c982f5abe..b065b971a9f0 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1040,7 +1040,7 @@ static int ceph_writepages_start(struct address_space *mapping, int num_ops = 0, op_idx; unsigned i, nr_folios, max_pages, locked_pages = 0; struct page **pages = NULL, **data_pages; - struct page *page; + struct folio *folio; pgoff_t strip_unit_end = 0; u64 offset = 0, len = 0; bool from_pool = false; @@ -1054,7 +1054,7 @@ static int ceph_writepages_start(struct address_space *mapping, if (!nr_folios && !locked_pages) break; for (i = 0; i < nr_folios && locked_pages < max_pages; i++) { - struct folio *folio = fbatch.folios[i]; + folio = fbatch.folios[i]; doutc(cl, "? %p idx %lu\n", folio, folio->index); if (locked_pages == 0) @@ -1175,7 +1175,7 @@ static int ceph_writepages_start(struct address_space *mapping, if (IS_ENCRYPTED(inode)) { pages[locked_pages] = - fscrypt_encrypt_pagecache_blocks(page, + fscrypt_encrypt_pagecache_blocks(&folio->page, PAGE_SIZE, 0, locked_pages ? GFP_NOWAIT : GFP_NOFS); if (IS_ERR(pages[locked_pages])) { @@ -1303,7 +1303,7 @@ static int ceph_writepages_start(struct address_space *mapping, /* writepages_finish() clears writeback pages * according to the data length, so make sure * data length covers all locked pages */ - u64 min_len = len + 1 - thp_size(page); + u64 min_len = len + 1 - folio_size(folio); len = get_writepages_data_length(inode, pages[i - 1], offset); len = max(len, min_len);
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index aba8d55bd533..1e0c982f5abe 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1056,22 +1056,21 @@ static int ceph_writepages_start(struct address_space *mapping, for (i = 0; i < nr_folios && locked_pages < max_pages; i++) { struct folio *folio = fbatch.folios[i]; - page = &folio->page; - doutc(cl, "? %p idx %lu\n", page, page->index); + doutc(cl, "? %p idx %lu\n", folio, folio->index); if (locked_pages == 0) - lock_page(page); /* first page */ - else if (!trylock_page(page)) + folio_lock(folio); /* first page */ + else if (!folio_trylock(folio)) break; /* only dirty pages, or our accounting breaks */ - if (unlikely(!PageDirty(page)) || - unlikely(page->mapping != mapping)) { - doutc(cl, "!dirty or !mapping %p\n", page); - unlock_page(page); + if (unlikely(!folio_test_dirty(folio)) || + unlikely(folio->mapping != mapping)) { + doutc(cl, "!dirty or !mapping %p\n", folio); + folio_unlock(folio); continue; } /* only if matching snap context */ - pgsnapc = page_snap_context(page); + pgsnapc = page_snap_context(&folio->page); if (pgsnapc != snapc) { doutc(cl, "page snapc %p %lld != oldest %p %lld\n", pgsnapc, pgsnapc->seq, snapc, snapc->seq); @@ -1079,10 +1078,10 @@ static int ceph_writepages_start(struct address_space *mapping, !ceph_wbc.head_snapc && wbc->sync_mode != WB_SYNC_NONE) should_loop = true; - unlock_page(page); + folio_unlock(folio); continue; } - if (page_offset(page) >= ceph_wbc.i_size) { + if (folio_pos(folio) >= ceph_wbc.i_size) { doutc(cl, "folio at %lu beyond eof %llu\n", folio->index, ceph_wbc.i_size); if ((ceph_wbc.size_stable || @@ -1093,9 +1092,9 @@ static int ceph_writepages_start(struct address_space *mapping, folio_unlock(folio); continue; } - if (strip_unit_end && (page->index > strip_unit_end)) { - doutc(cl, "end of strip unit %p\n", page); - unlock_page(page); + if (strip_unit_end && (folio->index > strip_unit_end)) { + doutc(cl, "end of strip unit %p\n", folio); + folio_unlock(folio); break; } if (folio_test_writeback(folio) || @@ -1110,9 +1109,9 @@ static int ceph_writepages_start(struct address_space *mapping, folio_wait_private_2(folio); /* [DEPRECATED] */ } - if (!clear_page_dirty_for_io(page)) { - doutc(cl, "%p !clear_page_dirty_for_io\n", page); - unlock_page(page); + if (!folio_clear_dirty_for_io(folio)) { + doutc(cl, "%p !clear_page_dirty_for_io\n", folio); + folio_unlock(folio); continue; } @@ -1128,7 +1127,7 @@ static int ceph_writepages_start(struct address_space *mapping, u32 xlen; /* prepare async write request */ - offset = (u64)page_offset(page); + offset = folio_pos(folio); ceph_calc_file_object_mapping(&ci->i_layout, offset, wsize, &objnum, &objoff, @@ -1136,7 +1135,7 @@ static int ceph_writepages_start(struct address_space *mapping, len = xlen; num_ops = 1; - strip_unit_end = page->index + + strip_unit_end = folio->index + ((len - 1) >> PAGE_SHIFT); BUG_ON(pages); @@ -1151,23 +1150,23 @@ static int ceph_writepages_start(struct address_space *mapping, } len = 0; - } else if (page->index != + } else if (folio->index != (offset + len) >> PAGE_SHIFT) { if (num_ops >= (from_pool ? CEPH_OSD_SLAB_OPS : CEPH_OSD_MAX_OPS)) { - redirty_page_for_writepage(wbc, page); - unlock_page(page); + folio_redirty_for_writepage(wbc, folio); + folio_unlock(folio); break; } num_ops++; - offset = (u64)page_offset(page); + offset = folio_pos(folio); len = 0; } /* note position of first page in fbatch */ doutc(cl, "%llx.%llx will write page %p idx %lu\n", - ceph_vinop(inode), page, page->index); + ceph_vinop(inode), folio, folio->index); if (atomic_long_inc_return(&fsc->writeback_count) > CONGESTION_ON_THRESH( @@ -1187,17 +1186,17 @@ static int ceph_writepages_start(struct address_space *mapping, /* better not fail on first page! */ BUG_ON(locked_pages == 0); pages[locked_pages] = NULL; - redirty_page_for_writepage(wbc, page); - unlock_page(page); + folio_redirty_for_writepage(wbc, folio); + folio_unlock(folio); break; } ++locked_pages; } else { - pages[locked_pages++] = page; + pages[locked_pages++] = &folio->page; } fbatch.folios[i] = NULL; - len += thp_size(page); + len += folio_size(folio); } /* did we get anything? */
We currently convert the folio returned from filemap_get_folios_tag() to a page and operate on that page. Remove this and operate on the folio. Removes a lot of calls to obsolete functions and references to page->index and page->mapping. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> --- fs/ceph/addr.c | 55 +++++++++++++++++++++++++------------------------- 1 file changed, 27 insertions(+), 28 deletions(-)