* [linux-next:master 2819/11071] fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start'
@ 2023-06-20 22:56 kernel test robot
2023-06-21 21:39 ` Kees Cook
0 siblings, 1 reply; 3+ messages in thread
From: kernel test robot @ 2023-06-20 22:56 UTC (permalink / raw)
To: Kees Cook
Cc: llvm, oe-kbuild-all, Linux Memory Management List, Gustavo A. R. Silva
tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head: 9dbf40840551df336c95ce2a3adbdd25ed53c0ef
commit: df8fc4e934c12b906d08050d7779f292b9c5c6b5 [2819/11071] kbuild: Enable -fstrict-flex-arrays=3
config: riscv-randconfig-r004-20230620 (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-lkp@intel.com/config)
compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
reproduce: (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-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/202306210637.TpfYq9gM-lkp@intel.com/
All warnings (new ones prefixed by >>):
>> fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start' [-Wframe-larger-than]
static int ceph_writepages_start(struct address_space *mapping,
^
1820/2096 (86.83%) spills, 276/2096 (13.17%) variables
1 warning generated.
vim +/ceph_writepages_start +794 fs/ceph/addr.c
1d3576fd10f0d7 Sage Weil 2009-10-06 790
1d3576fd10f0d7 Sage Weil 2009-10-06 791 /*
1d3576fd10f0d7 Sage Weil 2009-10-06 792 * initiate async writeback
1d3576fd10f0d7 Sage Weil 2009-10-06 793 */
1d3576fd10f0d7 Sage Weil 2009-10-06 @794 static int ceph_writepages_start(struct address_space *mapping,
1d3576fd10f0d7 Sage Weil 2009-10-06 795 struct writeback_control *wbc)
1d3576fd10f0d7 Sage Weil 2009-10-06 796 {
1d3576fd10f0d7 Sage Weil 2009-10-06 797 struct inode *inode = mapping->host;
1d3576fd10f0d7 Sage Weil 2009-10-06 798 struct ceph_inode_info *ci = ceph_inode(inode);
fc2744aa12da71 Yan, Zheng 2013-05-31 799 struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
fc2744aa12da71 Yan, Zheng 2013-05-31 800 struct ceph_vino vino = ceph_vino(inode);
2a2d927e35dd8d Yan, Zheng 2017-09-01 801 pgoff_t index, start_index, end = -1;
80e755fedebc8d Sage Weil 2010-03-31 802 struct ceph_snap_context *snapc = NULL, *last_snapc = NULL, *pgsnapc;
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 803) struct folio_batch fbatch;
1d3576fd10f0d7 Sage Weil 2009-10-06 804 int rc = 0;
93407472a21b82 Fabian Frederick 2017-02-27 805 unsigned int wsize = i_blocksize(inode);
1d3576fd10f0d7 Sage Weil 2009-10-06 806 struct ceph_osd_request *req = NULL;
1f934b00e90752 Yan, Zheng 2017-08-30 807 struct ceph_writeback_ctl ceph_wbc;
590e9d9861f5f2 Yan, Zheng 2017-09-03 808 bool should_loop, range_whole = false;
af9cc401ce7452 Yan, Zheng 2018-03-04 809 bool done = false;
1702e79734104d Jeff Layton 2021-12-07 810 bool caching = ceph_is_cache_enabled(inode);
7d41870d65db02 Xiubo Li 2023-03-08 811 xa_mark_t tag;
1d3576fd10f0d7 Sage Weil 2009-10-06 812
503d4fa6ee28e8 NeilBrown 2022-03-22 813 if (wbc->sync_mode == WB_SYNC_NONE &&
503d4fa6ee28e8 NeilBrown 2022-03-22 814 fsc->write_congested)
503d4fa6ee28e8 NeilBrown 2022-03-22 815 return 0;
503d4fa6ee28e8 NeilBrown 2022-03-22 816
3fb99d483e614b Yanhu Cao 2017-07-21 817 dout("writepages_start %p (mode=%s)\n", inode,
1d3576fd10f0d7 Sage Weil 2009-10-06 818 wbc->sync_mode == WB_SYNC_NONE ? "NONE" :
1d3576fd10f0d7 Sage Weil 2009-10-06 819 (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD"));
1d3576fd10f0d7 Sage Weil 2009-10-06 820
5d6451b1489ad1 Jeff Layton 2021-08-31 821 if (ceph_inode_is_shutdown(inode)) {
6c93df5db628e7 Yan, Zheng 2016-04-15 822 if (ci->i_wrbuffer_ref > 0) {
6c93df5db628e7 Yan, Zheng 2016-04-15 823 pr_warn_ratelimited(
6c93df5db628e7 Yan, Zheng 2016-04-15 824 "writepage_start %p %lld forced umount\n",
6c93df5db628e7 Yan, Zheng 2016-04-15 825 inode, ceph_ino(inode));
6c93df5db628e7 Yan, Zheng 2016-04-15 826 }
a341d4df87487a Yan, Zheng 2015-07-01 827 mapping_set_error(mapping, -EIO);
1d3576fd10f0d7 Sage Weil 2009-10-06 828 return -EIO; /* we're in a forced umount, don't write! */
1d3576fd10f0d7 Sage Weil 2009-10-06 829 }
95cca2b44e54b0 Yan, Zheng 2017-07-11 830 if (fsc->mount_options->wsize < wsize)
3d14c5d2b6e15c Yehuda Sadeh 2010-04-06 831 wsize = fsc->mount_options->wsize;
1d3576fd10f0d7 Sage Weil 2009-10-06 832
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 833) folio_batch_init(&fbatch);
1d3576fd10f0d7 Sage Weil 2009-10-06 834
590e9d9861f5f2 Yan, Zheng 2017-09-03 835 start_index = wbc->range_cyclic ? mapping->writeback_index : 0;
590e9d9861f5f2 Yan, Zheng 2017-09-03 836 index = start_index;
1d3576fd10f0d7 Sage Weil 2009-10-06 837
7d41870d65db02 Xiubo Li 2023-03-08 838 if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) {
7d41870d65db02 Xiubo Li 2023-03-08 839 tag = PAGECACHE_TAG_TOWRITE;
7d41870d65db02 Xiubo Li 2023-03-08 840 } else {
7d41870d65db02 Xiubo Li 2023-03-08 841 tag = PAGECACHE_TAG_DIRTY;
7d41870d65db02 Xiubo Li 2023-03-08 842 }
1d3576fd10f0d7 Sage Weil 2009-10-06 843 retry:
1d3576fd10f0d7 Sage Weil 2009-10-06 844 /* find oldest snap context with dirty data */
05455e1177f768 Yan, Zheng 2017-09-02 845 snapc = get_oldest_context(inode, &ceph_wbc, NULL);
1d3576fd10f0d7 Sage Weil 2009-10-06 846 if (!snapc) {
1d3576fd10f0d7 Sage Weil 2009-10-06 847 /* hmm, why does writepages get called when there
1d3576fd10f0d7 Sage Weil 2009-10-06 848 is no dirty data? */
1d3576fd10f0d7 Sage Weil 2009-10-06 849 dout(" no snap context with dirty data?\n");
1d3576fd10f0d7 Sage Weil 2009-10-06 850 goto out;
1d3576fd10f0d7 Sage Weil 2009-10-06 851 }
1d3576fd10f0d7 Sage Weil 2009-10-06 852 dout(" oldest snapc is %p seq %lld (%d snaps)\n",
1d3576fd10f0d7 Sage Weil 2009-10-06 853 snapc, snapc->seq, snapc->num_snaps);
fc2744aa12da71 Yan, Zheng 2013-05-31 854
2a2d927e35dd8d Yan, Zheng 2017-09-01 855 should_loop = false;
2a2d927e35dd8d Yan, Zheng 2017-09-01 856 if (ceph_wbc.head_snapc && snapc != last_snapc) {
2a2d927e35dd8d Yan, Zheng 2017-09-01 857 /* where to start/end? */
2a2d927e35dd8d Yan, Zheng 2017-09-01 858 if (wbc->range_cyclic) {
2a2d927e35dd8d Yan, Zheng 2017-09-01 859 index = start_index;
2a2d927e35dd8d Yan, Zheng 2017-09-01 860 end = -1;
2a2d927e35dd8d Yan, Zheng 2017-09-01 861 if (index > 0)
2a2d927e35dd8d Yan, Zheng 2017-09-01 862 should_loop = true;
2a2d927e35dd8d Yan, Zheng 2017-09-01 863 dout(" cyclic, start at %lu\n", index);
2a2d927e35dd8d Yan, Zheng 2017-09-01 864 } else {
2a2d927e35dd8d Yan, Zheng 2017-09-01 865 index = wbc->range_start >> PAGE_SHIFT;
2a2d927e35dd8d Yan, Zheng 2017-09-01 866 end = wbc->range_end >> PAGE_SHIFT;
2a2d927e35dd8d Yan, Zheng 2017-09-01 867 if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
2a2d927e35dd8d Yan, Zheng 2017-09-01 868 range_whole = true;
2a2d927e35dd8d Yan, Zheng 2017-09-01 869 dout(" not cyclic, %lu to %lu\n", index, end);
2a2d927e35dd8d Yan, Zheng 2017-09-01 870 }
2a2d927e35dd8d Yan, Zheng 2017-09-01 871 } else if (!ceph_wbc.head_snapc) {
2a2d927e35dd8d Yan, Zheng 2017-09-01 872 /* Do not respect wbc->range_{start,end}. Dirty pages
2a2d927e35dd8d Yan, Zheng 2017-09-01 873 * in that range can be associated with newer snapc.
2a2d927e35dd8d Yan, Zheng 2017-09-01 874 * They are not writeable until we write all dirty pages
2a2d927e35dd8d Yan, Zheng 2017-09-01 875 * associated with 'snapc' get written */
1582af2eaaf17c Yan, Zheng 2018-03-06 876 if (index > 0)
2a2d927e35dd8d Yan, Zheng 2017-09-01 877 should_loop = true;
2a2d927e35dd8d Yan, Zheng 2017-09-01 878 dout(" non-head snapc, range whole\n");
1d3576fd10f0d7 Sage Weil 2009-10-06 879 }
2a2d927e35dd8d Yan, Zheng 2017-09-01 880
7d41870d65db02 Xiubo Li 2023-03-08 881 if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
7d41870d65db02 Xiubo Li 2023-03-08 882 tag_pages_for_writeback(mapping, index, end);
7d41870d65db02 Xiubo Li 2023-03-08 883
2a2d927e35dd8d Yan, Zheng 2017-09-01 884 ceph_put_snap_context(last_snapc);
1d3576fd10f0d7 Sage Weil 2009-10-06 885 last_snapc = snapc;
1d3576fd10f0d7 Sage Weil 2009-10-06 886
af9cc401ce7452 Yan, Zheng 2018-03-04 887 while (!done && index <= end) {
5b64640cf65be4 Yan, Zheng 2016-01-07 888 int num_ops = 0, op_idx;
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 889) unsigned i, nr_folios, max_pages, locked_pages = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 890 struct page **pages = NULL, **data_pages;
1d3576fd10f0d7 Sage Weil 2009-10-06 891 struct page *page;
0e5ecac7168366 Yan, Zheng 2017-08-31 892 pgoff_t strip_unit_end = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 893 u64 offset = 0, len = 0;
a0102bda5bc099 Jeff Layton 2020-07-30 894 bool from_pool = false;
1d3576fd10f0d7 Sage Weil 2009-10-06 895
0e5ecac7168366 Yan, Zheng 2017-08-31 896 max_pages = wsize >> PAGE_SHIFT;
1d3576fd10f0d7 Sage Weil 2009-10-06 897
1d3576fd10f0d7 Sage Weil 2009-10-06 898 get_more_pages:
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 899) nr_folios = filemap_get_folios_tag(mapping, &index,
7d41870d65db02 Xiubo Li 2023-03-08 900 end, tag, &fbatch);
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 901) dout("pagevec_lookup_range_tag got %d\n", nr_folios);
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 902) if (!nr_folios && !locked_pages)
1d3576fd10f0d7 Sage Weil 2009-10-06 903 break;
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 904) for (i = 0; i < nr_folios && locked_pages < max_pages; i++) {
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 905) page = &fbatch.folios[i]->page;
1d3576fd10f0d7 Sage Weil 2009-10-06 906 dout("? %p idx %lu\n", page, page->index);
1d3576fd10f0d7 Sage Weil 2009-10-06 907 if (locked_pages == 0)
1d3576fd10f0d7 Sage Weil 2009-10-06 908 lock_page(page); /* first page */
1d3576fd10f0d7 Sage Weil 2009-10-06 909 else if (!trylock_page(page))
1d3576fd10f0d7 Sage Weil 2009-10-06 910 break;
1d3576fd10f0d7 Sage Weil 2009-10-06 911
1d3576fd10f0d7 Sage Weil 2009-10-06 912 /* only dirty pages, or our accounting breaks */
1d3576fd10f0d7 Sage Weil 2009-10-06 913 if (unlikely(!PageDirty(page)) ||
1d3576fd10f0d7 Sage Weil 2009-10-06 914 unlikely(page->mapping != mapping)) {
1d3576fd10f0d7 Sage Weil 2009-10-06 915 dout("!dirty or !mapping %p\n", page);
1d3576fd10f0d7 Sage Weil 2009-10-06 916 unlock_page(page);
0713e5f24b7deb Yan, Zheng 2017-08-31 917 continue;
1d3576fd10f0d7 Sage Weil 2009-10-06 918 }
af9cc401ce7452 Yan, Zheng 2018-03-04 919 /* only if matching snap context */
af9cc401ce7452 Yan, Zheng 2018-03-04 920 pgsnapc = page_snap_context(page);
af9cc401ce7452 Yan, Zheng 2018-03-04 921 if (pgsnapc != snapc) {
af9cc401ce7452 Yan, Zheng 2018-03-04 922 dout("page snapc %p %lld != oldest %p %lld\n",
af9cc401ce7452 Yan, Zheng 2018-03-04 923 pgsnapc, pgsnapc->seq, snapc, snapc->seq);
1582af2eaaf17c Yan, Zheng 2018-03-06 924 if (!should_loop &&
1582af2eaaf17c Yan, Zheng 2018-03-06 925 !ceph_wbc.head_snapc &&
1582af2eaaf17c Yan, Zheng 2018-03-06 926 wbc->sync_mode != WB_SYNC_NONE)
1582af2eaaf17c Yan, Zheng 2018-03-06 927 should_loop = true;
1d3576fd10f0d7 Sage Weil 2009-10-06 928 unlock_page(page);
af9cc401ce7452 Yan, Zheng 2018-03-04 929 continue;
1d3576fd10f0d7 Sage Weil 2009-10-06 930 }
1f934b00e90752 Yan, Zheng 2017-08-30 931 if (page_offset(page) >= ceph_wbc.i_size) {
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 932) struct folio *folio = page_folio(page);
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 933)
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 934) dout("folio at %lu beyond eof %llu\n",
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 935) folio->index, ceph_wbc.i_size);
c95f1c5f436bad Erqi Chen 2019-07-24 936 if ((ceph_wbc.size_stable ||
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 937) folio_pos(folio) >= i_size_read(inode)) &&
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 938) folio_clear_dirty_for_io(folio))
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 939) folio_invalidate(folio, 0,
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 940) folio_size(folio));
a628304ebe6ab0 Matthew Wilcox (Oracle 2022-02-09 941) folio_unlock(folio);
af9cc401ce7452 Yan, Zheng 2018-03-04 942 continue;
af9cc401ce7452 Yan, Zheng 2018-03-04 943 }
af9cc401ce7452 Yan, Zheng 2018-03-04 944 if (strip_unit_end && (page->index > strip_unit_end)) {
af9cc401ce7452 Yan, Zheng 2018-03-04 945 dout("end of strip unit %p\n", page);
1d3576fd10f0d7 Sage Weil 2009-10-06 946 unlock_page(page);
1d3576fd10f0d7 Sage Weil 2009-10-06 947 break;
1d3576fd10f0d7 Sage Weil 2009-10-06 948 }
1702e79734104d Jeff Layton 2021-12-07 949 if (PageWriteback(page) || PageFsCache(page)) {
0713e5f24b7deb Yan, Zheng 2017-08-31 950 if (wbc->sync_mode == WB_SYNC_NONE) {
1d3576fd10f0d7 Sage Weil 2009-10-06 951 dout("%p under writeback\n", page);
1d3576fd10f0d7 Sage Weil 2009-10-06 952 unlock_page(page);
0713e5f24b7deb Yan, Zheng 2017-08-31 953 continue;
0713e5f24b7deb Yan, Zheng 2017-08-31 954 }
0713e5f24b7deb Yan, Zheng 2017-08-31 955 dout("waiting on writeback %p\n", page);
0713e5f24b7deb Yan, Zheng 2017-08-31 956 wait_on_page_writeback(page);
1702e79734104d Jeff Layton 2021-12-07 957 wait_on_page_fscache(page);
1d3576fd10f0d7 Sage Weil 2009-10-06 958 }
1d3576fd10f0d7 Sage Weil 2009-10-06 959
1d3576fd10f0d7 Sage Weil 2009-10-06 960 if (!clear_page_dirty_for_io(page)) {
1d3576fd10f0d7 Sage Weil 2009-10-06 961 dout("%p !clear_page_dirty_for_io\n", page);
1d3576fd10f0d7 Sage Weil 2009-10-06 962 unlock_page(page);
0713e5f24b7deb Yan, Zheng 2017-08-31 963 continue;
1d3576fd10f0d7 Sage Weil 2009-10-06 964 }
1d3576fd10f0d7 Sage Weil 2009-10-06 965
e5975c7c8eb6ae Alex Elder 2013-03-14 966 /*
e5975c7c8eb6ae Alex Elder 2013-03-14 967 * We have something to write. If this is
e5975c7c8eb6ae Alex Elder 2013-03-14 968 * the first locked page this time through,
5b64640cf65be4 Yan, Zheng 2016-01-07 969 * calculate max possinle write size and
5b64640cf65be4 Yan, Zheng 2016-01-07 970 * allocate a page array
e5975c7c8eb6ae Alex Elder 2013-03-14 971 */
1d3576fd10f0d7 Sage Weil 2009-10-06 972 if (locked_pages == 0) {
5b64640cf65be4 Yan, Zheng 2016-01-07 973 u64 objnum;
5b64640cf65be4 Yan, Zheng 2016-01-07 974 u64 objoff;
dccbf08005df80 Ilya Dryomov 2018-02-17 975 u32 xlen;
5b64640cf65be4 Yan, Zheng 2016-01-07 976
1d3576fd10f0d7 Sage Weil 2009-10-06 977 /* prepare async write request */
6285bc23127741 Alex Elder 2012-10-02 978 offset = (u64)page_offset(page);
dccbf08005df80 Ilya Dryomov 2018-02-17 979 ceph_calc_file_object_mapping(&ci->i_layout,
dccbf08005df80 Ilya Dryomov 2018-02-17 980 offset, wsize,
5b64640cf65be4 Yan, Zheng 2016-01-07 981 &objnum, &objoff,
dccbf08005df80 Ilya Dryomov 2018-02-17 982 &xlen);
dccbf08005df80 Ilya Dryomov 2018-02-17 983 len = xlen;
8c71897be2ddfd Henry C Chang 2011-05-03 984
3fb99d483e614b Yanhu Cao 2017-07-21 985 num_ops = 1;
5b64640cf65be4 Yan, Zheng 2016-01-07 986 strip_unit_end = page->index +
09cbfeaf1a5a67 Kirill A. Shutemov 2016-04-01 987 ((len - 1) >> PAGE_SHIFT);
88486957f9fbf5 Alex Elder 2013-03-14 988
5b64640cf65be4 Yan, Zheng 2016-01-07 989 BUG_ON(pages);
88486957f9fbf5 Alex Elder 2013-03-14 990 max_pages = calc_pages_for(0, (u64)len);
6da2ec56059c3c Kees Cook 2018-06-12 991 pages = kmalloc_array(max_pages,
6da2ec56059c3c Kees Cook 2018-06-12 992 sizeof(*pages),
fc2744aa12da71 Yan, Zheng 2013-05-31 993 GFP_NOFS);
88486957f9fbf5 Alex Elder 2013-03-14 994 if (!pages) {
a0102bda5bc099 Jeff Layton 2020-07-30 995 from_pool = true;
a0102bda5bc099 Jeff Layton 2020-07-30 996 pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS);
e5975c7c8eb6ae Alex Elder 2013-03-14 997 BUG_ON(!pages);
88486957f9fbf5 Alex Elder 2013-03-14 998 }
5b64640cf65be4 Yan, Zheng 2016-01-07 999
5b64640cf65be4 Yan, Zheng 2016-01-07 1000 len = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 1001 } else if (page->index !=
09cbfeaf1a5a67 Kirill A. Shutemov 2016-04-01 1002 (offset + len) >> PAGE_SHIFT) {
a0102bda5bc099 Jeff Layton 2020-07-30 1003 if (num_ops >= (from_pool ? CEPH_OSD_SLAB_OPS :
5b64640cf65be4 Yan, Zheng 2016-01-07 1004 CEPH_OSD_MAX_OPS)) {
5b64640cf65be4 Yan, Zheng 2016-01-07 1005 redirty_page_for_writepage(wbc, page);
5b64640cf65be4 Yan, Zheng 2016-01-07 1006 unlock_page(page);
5b64640cf65be4 Yan, Zheng 2016-01-07 1007 break;
5b64640cf65be4 Yan, Zheng 2016-01-07 1008 }
5b64640cf65be4 Yan, Zheng 2016-01-07 1009
5b64640cf65be4 Yan, Zheng 2016-01-07 1010 num_ops++;
5b64640cf65be4 Yan, Zheng 2016-01-07 1011 offset = (u64)page_offset(page);
5b64640cf65be4 Yan, Zheng 2016-01-07 1012 len = 0;
1d3576fd10f0d7 Sage Weil 2009-10-06 1013 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1014
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1015) /* note position of first page in fbatch */
1d3576fd10f0d7 Sage Weil 2009-10-06 1016 dout("%p will write page %p idx %lu\n",
1d3576fd10f0d7 Sage Weil 2009-10-06 1017 inode, page, page->index);
2baba25019ec56 Yehuda Sadeh 2009-12-18 1018
5b64640cf65be4 Yan, Zheng 2016-01-07 1019 if (atomic_long_inc_return(&fsc->writeback_count) >
5b64640cf65be4 Yan, Zheng 2016-01-07 1020 CONGESTION_ON_THRESH(
503d4fa6ee28e8 NeilBrown 2022-03-22 1021 fsc->mount_options->congestion_kb))
503d4fa6ee28e8 NeilBrown 2022-03-22 1022 fsc->write_congested = true;
0713e5f24b7deb Yan, Zheng 2017-08-31 1023
0713e5f24b7deb Yan, Zheng 2017-08-31 1024 pages[locked_pages++] = page;
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1025) fbatch.folios[i] = NULL;
0713e5f24b7deb Yan, Zheng 2017-08-31 1026
8ff2d290c8ce77 Jeff Layton 2021-04-05 1027 len += thp_size(page);
1d3576fd10f0d7 Sage Weil 2009-10-06 1028 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1029
1d3576fd10f0d7 Sage Weil 2009-10-06 1030 /* did we get anything? */
1d3576fd10f0d7 Sage Weil 2009-10-06 1031 if (!locked_pages)
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1032) goto release_folios;
1d3576fd10f0d7 Sage Weil 2009-10-06 1033 if (i) {
0713e5f24b7deb Yan, Zheng 2017-08-31 1034 unsigned j, n = 0;
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1035) /* shift unused page to beginning of fbatch */
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1036) for (j = 0; j < nr_folios; j++) {
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1037) if (!fbatch.folios[j])
0713e5f24b7deb Yan, Zheng 2017-08-31 1038 continue;
0713e5f24b7deb Yan, Zheng 2017-08-31 1039 if (n < j)
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1040) fbatch.folios[n] = fbatch.folios[j];
0713e5f24b7deb Yan, Zheng 2017-08-31 1041 n++;
0713e5f24b7deb Yan, Zheng 2017-08-31 1042 }
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1043) fbatch.nr = n;
1d3576fd10f0d7 Sage Weil 2009-10-06 1044
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1045) if (nr_folios && i == nr_folios &&
1d3576fd10f0d7 Sage Weil 2009-10-06 1046 locked_pages < max_pages) {
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1047) dout("reached end fbatch, trying for more\n");
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1048) folio_batch_release(&fbatch);
1d3576fd10f0d7 Sage Weil 2009-10-06 1049 goto get_more_pages;
1d3576fd10f0d7 Sage Weil 2009-10-06 1050 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1051 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1052
5b64640cf65be4 Yan, Zheng 2016-01-07 1053 new_request:
e5975c7c8eb6ae Alex Elder 2013-03-14 1054 offset = page_offset(pages[0]);
5b64640cf65be4 Yan, Zheng 2016-01-07 1055 len = wsize;
5b64640cf65be4 Yan, Zheng 2016-01-07 1056
5b64640cf65be4 Yan, Zheng 2016-01-07 1057 req = ceph_osdc_new_request(&fsc->client->osdc,
5b64640cf65be4 Yan, Zheng 2016-01-07 1058 &ci->i_layout, vino,
5b64640cf65be4 Yan, Zheng 2016-01-07 1059 offset, &len, 0, num_ops,
1f934b00e90752 Yan, Zheng 2017-08-30 1060 CEPH_OSD_OP_WRITE, CEPH_OSD_FLAG_WRITE,
1f934b00e90752 Yan, Zheng 2017-08-30 1061 snapc, ceph_wbc.truncate_seq,
1f934b00e90752 Yan, Zheng 2017-08-30 1062 ceph_wbc.truncate_size, false);
5b64640cf65be4 Yan, Zheng 2016-01-07 1063 if (IS_ERR(req)) {
5b64640cf65be4 Yan, Zheng 2016-01-07 1064 req = ceph_osdc_new_request(&fsc->client->osdc,
5b64640cf65be4 Yan, Zheng 2016-01-07 1065 &ci->i_layout, vino,
5b64640cf65be4 Yan, Zheng 2016-01-07 1066 offset, &len, 0,
5b64640cf65be4 Yan, Zheng 2016-01-07 1067 min(num_ops,
5b64640cf65be4 Yan, Zheng 2016-01-07 1068 CEPH_OSD_SLAB_OPS),
5b64640cf65be4 Yan, Zheng 2016-01-07 1069 CEPH_OSD_OP_WRITE,
54ea0046b6fe36 Ilya Dryomov 2017-02-11 1070 CEPH_OSD_FLAG_WRITE,
1f934b00e90752 Yan, Zheng 2017-08-30 1071 snapc, ceph_wbc.truncate_seq,
1f934b00e90752 Yan, Zheng 2017-08-30 1072 ceph_wbc.truncate_size, true);
5b64640cf65be4 Yan, Zheng 2016-01-07 1073 BUG_ON(IS_ERR(req));
5b64640cf65be4 Yan, Zheng 2016-01-07 1074 }
5b64640cf65be4 Yan, Zheng 2016-01-07 1075 BUG_ON(len < page_offset(pages[locked_pages - 1]) +
8ff2d290c8ce77 Jeff Layton 2021-04-05 1076 thp_size(page) - offset);
5b64640cf65be4 Yan, Zheng 2016-01-07 1077
5b64640cf65be4 Yan, Zheng 2016-01-07 1078 req->r_callback = writepages_finish;
5b64640cf65be4 Yan, Zheng 2016-01-07 1079 req->r_inode = inode;
5b64640cf65be4 Yan, Zheng 2016-01-07 1080
5b64640cf65be4 Yan, Zheng 2016-01-07 1081 /* Format the osd request message and submit the write */
5b64640cf65be4 Yan, Zheng 2016-01-07 1082 len = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 1083 data_pages = pages;
5b64640cf65be4 Yan, Zheng 2016-01-07 1084 op_idx = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 1085 for (i = 0; i < locked_pages; i++) {
5b64640cf65be4 Yan, Zheng 2016-01-07 1086 u64 cur_offset = page_offset(pages[i]);
1702e79734104d Jeff Layton 2021-12-07 1087 /*
1702e79734104d Jeff Layton 2021-12-07 1088 * Discontinuity in page range? Ceph can handle that by just passing
1702e79734104d Jeff Layton 2021-12-07 1089 * multiple extents in the write op.
1702e79734104d Jeff Layton 2021-12-07 1090 */
5b64640cf65be4 Yan, Zheng 2016-01-07 1091 if (offset + len != cur_offset) {
1702e79734104d Jeff Layton 2021-12-07 1092 /* If it's full, stop here */
3fb99d483e614b Yanhu Cao 2017-07-21 1093 if (op_idx + 1 == req->r_num_ops)
5b64640cf65be4 Yan, Zheng 2016-01-07 1094 break;
1702e79734104d Jeff Layton 2021-12-07 1095
1702e79734104d Jeff Layton 2021-12-07 1096 /* Kick off an fscache write with what we have so far. */
1702e79734104d Jeff Layton 2021-12-07 1097 ceph_fscache_write_to_cache(inode, offset, len, caching);
1702e79734104d Jeff Layton 2021-12-07 1098
1702e79734104d Jeff Layton 2021-12-07 1099 /* Start a new extent */
5b64640cf65be4 Yan, Zheng 2016-01-07 1100 osd_req_op_extent_dup_last(req, op_idx,
5b64640cf65be4 Yan, Zheng 2016-01-07 1101 cur_offset - offset);
5b64640cf65be4 Yan, Zheng 2016-01-07 1102 dout("writepages got pages at %llu~%llu\n",
5b64640cf65be4 Yan, Zheng 2016-01-07 1103 offset, len);
5b64640cf65be4 Yan, Zheng 2016-01-07 1104 osd_req_op_extent_osd_data_pages(req, op_idx,
5b64640cf65be4 Yan, Zheng 2016-01-07 1105 data_pages, len, 0,
a0102bda5bc099 Jeff Layton 2020-07-30 1106 from_pool, false);
5b64640cf65be4 Yan, Zheng 2016-01-07 1107 osd_req_op_extent_update(req, op_idx, len);
5b64640cf65be4 Yan, Zheng 2016-01-07 1108
5b64640cf65be4 Yan, Zheng 2016-01-07 1109 len = 0;
5b64640cf65be4 Yan, Zheng 2016-01-07 1110 offset = cur_offset;
5b64640cf65be4 Yan, Zheng 2016-01-07 1111 data_pages = pages + i;
5b64640cf65be4 Yan, Zheng 2016-01-07 1112 op_idx++;
5b64640cf65be4 Yan, Zheng 2016-01-07 1113 }
5b64640cf65be4 Yan, Zheng 2016-01-07 1114
5b64640cf65be4 Yan, Zheng 2016-01-07 1115 set_page_writeback(pages[i]);
1702e79734104d Jeff Layton 2021-12-07 1116 if (caching)
1702e79734104d Jeff Layton 2021-12-07 1117 ceph_set_page_fscache(pages[i]);
8ff2d290c8ce77 Jeff Layton 2021-04-05 1118 len += thp_size(page);
5b64640cf65be4 Yan, Zheng 2016-01-07 1119 }
1702e79734104d Jeff Layton 2021-12-07 1120 ceph_fscache_write_to_cache(inode, offset, len, caching);
5b64640cf65be4 Yan, Zheng 2016-01-07 1121
1f934b00e90752 Yan, Zheng 2017-08-30 1122 if (ceph_wbc.size_stable) {
1f934b00e90752 Yan, Zheng 2017-08-30 1123 len = min(len, ceph_wbc.i_size - offset);
5b64640cf65be4 Yan, Zheng 2016-01-07 1124 } else if (i == locked_pages) {
e1966b49446a43 Yan, Zheng 2015-06-18 1125 /* writepages_finish() clears writeback pages
e1966b49446a43 Yan, Zheng 2015-06-18 1126 * according to the data length, so make sure
e1966b49446a43 Yan, Zheng 2015-06-18 1127 * data length covers all locked pages */
8ff2d290c8ce77 Jeff Layton 2021-04-05 1128 u64 min_len = len + 1 - thp_size(page);
1f934b00e90752 Yan, Zheng 2017-08-30 1129 len = get_writepages_data_length(inode, pages[i - 1],
1f934b00e90752 Yan, Zheng 2017-08-30 1130 offset);
5b64640cf65be4 Yan, Zheng 2016-01-07 1131 len = max(len, min_len);
e1966b49446a43 Yan, Zheng 2015-06-18 1132 }
5b64640cf65be4 Yan, Zheng 2016-01-07 1133 dout("writepages got pages at %llu~%llu\n", offset, len);
1d3576fd10f0d7 Sage Weil 2009-10-06 1134
5b64640cf65be4 Yan, Zheng 2016-01-07 1135 osd_req_op_extent_osd_data_pages(req, op_idx, data_pages, len,
a0102bda5bc099 Jeff Layton 2020-07-30 1136 0, from_pool, false);
5b64640cf65be4 Yan, Zheng 2016-01-07 1137 osd_req_op_extent_update(req, op_idx, len);
e5975c7c8eb6ae Alex Elder 2013-03-14 1138
5b64640cf65be4 Yan, Zheng 2016-01-07 1139 BUG_ON(op_idx + 1 != req->r_num_ops);
e5975c7c8eb6ae Alex Elder 2013-03-14 1140
a0102bda5bc099 Jeff Layton 2020-07-30 1141 from_pool = false;
5b64640cf65be4 Yan, Zheng 2016-01-07 1142 if (i < locked_pages) {
5b64640cf65be4 Yan, Zheng 2016-01-07 1143 BUG_ON(num_ops <= req->r_num_ops);
5b64640cf65be4 Yan, Zheng 2016-01-07 1144 num_ops -= req->r_num_ops;
5b64640cf65be4 Yan, Zheng 2016-01-07 1145 locked_pages -= i;
5b64640cf65be4 Yan, Zheng 2016-01-07 1146
5b64640cf65be4 Yan, Zheng 2016-01-07 1147 /* allocate new pages array for next request */
5b64640cf65be4 Yan, Zheng 2016-01-07 1148 data_pages = pages;
6da2ec56059c3c Kees Cook 2018-06-12 1149 pages = kmalloc_array(locked_pages, sizeof(*pages),
5b64640cf65be4 Yan, Zheng 2016-01-07 1150 GFP_NOFS);
5b64640cf65be4 Yan, Zheng 2016-01-07 1151 if (!pages) {
a0102bda5bc099 Jeff Layton 2020-07-30 1152 from_pool = true;
a0102bda5bc099 Jeff Layton 2020-07-30 1153 pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS);
5b64640cf65be4 Yan, Zheng 2016-01-07 1154 BUG_ON(!pages);
5b64640cf65be4 Yan, Zheng 2016-01-07 1155 }
5b64640cf65be4 Yan, Zheng 2016-01-07 1156 memcpy(pages, data_pages + i,
5b64640cf65be4 Yan, Zheng 2016-01-07 1157 locked_pages * sizeof(*pages));
5b64640cf65be4 Yan, Zheng 2016-01-07 1158 memset(data_pages + i, 0,
5b64640cf65be4 Yan, Zheng 2016-01-07 1159 locked_pages * sizeof(*pages));
5b64640cf65be4 Yan, Zheng 2016-01-07 1160 } else {
5b64640cf65be4 Yan, Zheng 2016-01-07 1161 BUG_ON(num_ops != req->r_num_ops);
5b64640cf65be4 Yan, Zheng 2016-01-07 1162 index = pages[i - 1]->index + 1;
5b64640cf65be4 Yan, Zheng 2016-01-07 1163 /* request message now owns the pages array */
5b64640cf65be4 Yan, Zheng 2016-01-07 1164 pages = NULL;
5b64640cf65be4 Yan, Zheng 2016-01-07 1165 }
e5975c7c8eb6ae Alex Elder 2013-03-14 1166
fac02ddf910814 Arnd Bergmann 2018-07-13 1167 req->r_mtime = inode->i_mtime;
a8af0d682ae0c9 Jeff Layton 2022-06-30 1168 ceph_osdc_start_request(&fsc->client->osdc, req);
1d3576fd10f0d7 Sage Weil 2009-10-06 1169 req = NULL;
1d3576fd10f0d7 Sage Weil 2009-10-06 1170
5b64640cf65be4 Yan, Zheng 2016-01-07 1171 wbc->nr_to_write -= i;
5b64640cf65be4 Yan, Zheng 2016-01-07 1172 if (pages)
5b64640cf65be4 Yan, Zheng 2016-01-07 1173 goto new_request;
5b64640cf65be4 Yan, Zheng 2016-01-07 1174
2a2d927e35dd8d Yan, Zheng 2017-09-01 1175 /*
2a2d927e35dd8d Yan, Zheng 2017-09-01 1176 * We stop writing back only if we are not doing
2a2d927e35dd8d Yan, Zheng 2017-09-01 1177 * integrity sync. In case of integrity sync we have to
2a2d927e35dd8d Yan, Zheng 2017-09-01 1178 * keep going until we have written all the pages
2a2d927e35dd8d Yan, Zheng 2017-09-01 1179 * we tagged for writeback prior to entering this loop.
2a2d927e35dd8d Yan, Zheng 2017-09-01 1180 */
2a2d927e35dd8d Yan, Zheng 2017-09-01 1181 if (wbc->nr_to_write <= 0 && wbc->sync_mode == WB_SYNC_NONE)
af9cc401ce7452 Yan, Zheng 2018-03-04 1182 done = true;
1d3576fd10f0d7 Sage Weil 2009-10-06 1183
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1184) release_folios:
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1185) dout("folio_batch release on %d folios (%p)\n", (int)fbatch.nr,
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1186) fbatch.nr ? fbatch.folios[0] : NULL);
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1187) folio_batch_release(&fbatch);
1d3576fd10f0d7 Sage Weil 2009-10-06 1188 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1189
1d3576fd10f0d7 Sage Weil 2009-10-06 1190 if (should_loop && !done) {
1d3576fd10f0d7 Sage Weil 2009-10-06 1191 /* more to do; loop back to beginning of file */
1d3576fd10f0d7 Sage Weil 2009-10-06 1192 dout("writepages looping back to beginning of file\n");
2a2d927e35dd8d Yan, Zheng 2017-09-01 1193 end = start_index - 1; /* OK even when start_index == 0 */
f275635ee0b664 Yan, Zheng 2017-09-01 1194
f275635ee0b664 Yan, Zheng 2017-09-01 1195 /* to write dirty pages associated with next snapc,
f275635ee0b664 Yan, Zheng 2017-09-01 1196 * we need to wait until current writes complete */
f275635ee0b664 Yan, Zheng 2017-09-01 1197 if (wbc->sync_mode != WB_SYNC_NONE &&
f275635ee0b664 Yan, Zheng 2017-09-01 1198 start_index == 0 && /* all dirty pages were checked */
f275635ee0b664 Yan, Zheng 2017-09-01 1199 !ceph_wbc.head_snapc) {
f275635ee0b664 Yan, Zheng 2017-09-01 1200 struct page *page;
f275635ee0b664 Yan, Zheng 2017-09-01 1201 unsigned i, nr;
f275635ee0b664 Yan, Zheng 2017-09-01 1202 index = 0;
f275635ee0b664 Yan, Zheng 2017-09-01 1203 while ((index <= end) &&
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1204) (nr = filemap_get_folios_tag(mapping, &index,
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1205) (pgoff_t)-1,
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1206) PAGECACHE_TAG_WRITEBACK,
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1207) &fbatch))) {
f275635ee0b664 Yan, Zheng 2017-09-01 1208 for (i = 0; i < nr; i++) {
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1209) page = &fbatch.folios[i]->page;
f275635ee0b664 Yan, Zheng 2017-09-01 1210 if (page_snap_context(page) != snapc)
f275635ee0b664 Yan, Zheng 2017-09-01 1211 continue;
f275635ee0b664 Yan, Zheng 2017-09-01 1212 wait_on_page_writeback(page);
f275635ee0b664 Yan, Zheng 2017-09-01 1213 }
590a2b5f0a9b74 Vishal Moola (Oracle 2023-01-04 1214) folio_batch_release(&fbatch);
f275635ee0b664 Yan, Zheng 2017-09-01 1215 cond_resched();
f275635ee0b664 Yan, Zheng 2017-09-01 1216 }
f275635ee0b664 Yan, Zheng 2017-09-01 1217 }
f275635ee0b664 Yan, Zheng 2017-09-01 1218
2a2d927e35dd8d Yan, Zheng 2017-09-01 1219 start_index = 0;
1d3576fd10f0d7 Sage Weil 2009-10-06 1220 index = 0;
1d3576fd10f0d7 Sage Weil 2009-10-06 1221 goto retry;
1d3576fd10f0d7 Sage Weil 2009-10-06 1222 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1223
1d3576fd10f0d7 Sage Weil 2009-10-06 1224 if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
1d3576fd10f0d7 Sage Weil 2009-10-06 1225 mapping->writeback_index = index;
1d3576fd10f0d7 Sage Weil 2009-10-06 1226
1d3576fd10f0d7 Sage Weil 2009-10-06 1227 out:
1d3576fd10f0d7 Sage Weil 2009-10-06 1228 ceph_osdc_put_request(req);
2a2d927e35dd8d Yan, Zheng 2017-09-01 1229 ceph_put_snap_context(last_snapc);
2a2d927e35dd8d Yan, Zheng 2017-09-01 1230 dout("writepages dend - startone, rc = %d\n", rc);
1d3576fd10f0d7 Sage Weil 2009-10-06 1231 return rc;
1d3576fd10f0d7 Sage Weil 2009-10-06 1232 }
1d3576fd10f0d7 Sage Weil 2009-10-06 1233
:::::: The code at line 794 was first introduced by commit
:::::: 1d3576fd10f0d7a104204267b81cf84a07028dad ceph: address space operations
:::::: TO: Sage Weil <sage@newdream.net>
:::::: CC: Sage Weil <sage@newdream.net>
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [linux-next:master 2819/11071] fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start'
2023-06-20 22:56 [linux-next:master 2819/11071] fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start' kernel test robot
@ 2023-06-21 21:39 ` Kees Cook
2023-06-21 21:44 ` Kees Cook
0 siblings, 1 reply; 3+ messages in thread
From: Kees Cook @ 2023-06-21 21:39 UTC (permalink / raw)
To: kernel test robot
Cc: llvm, oe-kbuild-all, Linux Memory Management List,
Gustavo A. R. Silva, linux-hardening, Nick Desaulniers
On Wed, Jun 21, 2023 at 06:56:36AM +0800, kernel test robot wrote:
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> head: 9dbf40840551df336c95ce2a3adbdd25ed53c0ef
> commit: df8fc4e934c12b906d08050d7779f292b9c5c6b5 [2819/11071] kbuild: Enable -fstrict-flex-arrays=3
This looks very much like:
https://lore.kernel.org/all/202306100035.VTusNhm4-lkp@intel.com/
> config: riscv-randconfig-r004-20230620 (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-lkp@intel.com/config)
> compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
> reproduce: (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-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/202306210637.TpfYq9gM-lkp@intel.com/
>
> All warnings (new ones prefixed by >>):
>
> >> fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start' [-Wframe-larger-than]
> static int ceph_writepages_start(struct address_space *mapping,
> ^
> 1820/2096 (86.83%) spills, 276/2096 (13.17%) variables
> 1 warning generated.
This appears to be a clang stack slot reuse issue. It continues to be
investigated.
-Kees
--
Kees Cook
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [linux-next:master 2819/11071] fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start'
2023-06-21 21:39 ` Kees Cook
@ 2023-06-21 21:44 ` Kees Cook
0 siblings, 0 replies; 3+ messages in thread
From: Kees Cook @ 2023-06-21 21:44 UTC (permalink / raw)
To: kernel test robot
Cc: llvm, oe-kbuild-all, Linux Memory Management List,
Gustavo A. R. Silva, linux-hardening, Nick Desaulniers
On Wed, Jun 21, 2023 at 02:39:39PM -0700, Kees Cook wrote:
> On Wed, Jun 21, 2023 at 06:56:36AM +0800, kernel test robot wrote:
> > tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> > head: 9dbf40840551df336c95ce2a3adbdd25ed53c0ef
> > commit: df8fc4e934c12b906d08050d7779f292b9c5c6b5 [2819/11071] kbuild: Enable -fstrict-flex-arrays=3
>
> This looks very much like:
> https://lore.kernel.org/all/202306100035.VTusNhm4-lkp@intel.com/
>
> > config: riscv-randconfig-r004-20230620 (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-lkp@intel.com/config)
> > compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
> > reproduce: (https://download.01.org/0day-ci/archive/20230621/202306210637.TpfYq9gM-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/202306210637.TpfYq9gM-lkp@intel.com/
> >
> > All warnings (new ones prefixed by >>):
> >
> > >> fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start' [-Wframe-larger-than]
> > static int ceph_writepages_start(struct address_space *mapping,
> > ^
> > 1820/2096 (86.83%) spills, 276/2096 (13.17%) variables
> > 1 warning generated.
>
> This appears to be a clang stack slot reuse issue. It continues to be
> investigated.
We're currently tracking it here, though we may need to split the bug
up:
https://github.com/ClangBuiltLinux/linux/issues/39
--
Kees Cook
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-06-21 21:44 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-06-20 22:56 [linux-next:master 2819/11071] fs/ceph/addr.c:794:12: warning: stack frame size (2096) exceeds limit (2048) in 'ceph_writepages_start' kernel test robot
2023-06-21 21:39 ` Kees Cook
2023-06-21 21:44 ` Kees Cook
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox