lkml.org 
[lkml]   [2022]   [Aug]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subject[ammarfaizi2-block:akpm/mm/mm-unstable 272/276] fs/hugetlbfs/inode.c:562:16: warning: variable 'm_index' is uninitialized when used here
tree:   https://github.com/ammarfaizi2/linux-block akpm/mm/mm-unstable
head: 73129b786c0eca5be67491a5e893e13b75133a62
commit: fcc0d3d00d74b012c36d52128687ab9d197d3f0d [272/276] hugetlb: handle truncate racing with page faults
config: arm64-randconfig-r014-20220823 (https://download.01.org/0day-ci/archive/20220825/202208251012.g1mOhhxy-lkp@intel.com/config)
compiler: clang version 16.0.0 (https://github.com/llvm/llvm-project d00e97df0fe8c67f694c4d027297f4382ce72b38)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# install arm64 cross compiling tool for clang build
# apt-get install binutils-aarch64-linux-gnu
# https://github.com/ammarfaizi2/linux-block/commit/fcc0d3d00d74b012c36d52128687ab9d197d3f0d
git remote add ammarfaizi2-block https://github.com/ammarfaizi2/linux-block
git fetch --no-tags ammarfaizi2-block akpm/mm/mm-unstable
git checkout fcc0d3d00d74b012c36d52128687ab9d197d3f0d
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=arm64 SHELL=/bin/bash fs/hugetlbfs/

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> fs/hugetlbfs/inode.c:562:16: warning: variable 'm_index' is uninitialized when used here [-Wuninitialized]
m_start, m_index, truncate_op);
^~~~~~~
fs/hugetlbfs/inode.c:540:26: note: initialize the variable 'm_index' to silence this warning
pgoff_t m_start, m_index;
^
= 0
1 warning generated.


vim +/m_index +562 fs/hugetlbfs/inode.c

502
503 /*
504 * remove_inode_hugepages handles two distinct cases: truncation and hole
505 * punch. There are subtle differences in operation for each case.
506 *
507 * truncation is indicated by end of range being LLONG_MAX
508 * In this case, we first scan the range and release found pages.
509 * After releasing pages, hugetlb_unreserve_pages cleans up region/reserve
510 * maps and global counts. Page faults can race with truncation.
511 * During faults, hugetlb_no_page() checks i_size before page allocation,
512 * and again after obtaining page table lock. It will 'back out'
513 * allocations in the truncated range.
514 * hole punch is indicated if end is not LLONG_MAX
515 * In the hole punch case we scan the range and release found pages.
516 * Only when releasing a page is the associated region/reserve map
517 * deleted. The region/reserve map for ranges without associated
518 * pages are not modified. Page faults can race with hole punch.
519 * This is indicated if we find a mapped page.
520 * Note: If the passed end of range value is beyond the end of file, but
521 * not LLONG_MAX this routine still performs a hole punch operation.
522 *
523 * Since page faults can race with this routine, care must be taken as both
524 * modify huge page reservation data. To somewhat synchronize these operations
525 * the hugetlb fault mutex is taken for EVERY index in the range to be hole
526 * punched or truncated. In this way, we KNOW either:
527 * - fault code has added a page beyond i_size, and we will remove here
528 * - fault code will see updated i_size and not add a page beyond
529 * The parameter 'lm__end' indicates the offset of the end of hole or file
530 * before truncation. For hole punch lm_end == lend.
531 */
532 static void remove_inode_hugepages(struct inode *inode, loff_t lstart,
533 loff_t lend, loff_t lm_end)
534 {
535 struct hstate *h = hstate_inode(inode);
536 struct address_space *mapping = &inode->i_data;
537 const pgoff_t start = lstart >> huge_page_shift(h);
538 const pgoff_t end = lend >> huge_page_shift(h);
539 pgoff_t m_end = lm_end >> huge_page_shift(h);
540 pgoff_t m_start, m_index;
541 struct folio_batch fbatch;
542 struct folio *folio;
543 pgoff_t next, index;
544 unsigned int i;
545 long freed = 0;
546 u32 hash;
547 bool truncate_op = (lend == LLONG_MAX);
548
549 folio_batch_init(&fbatch);
550 next = m_start = start;
551 while (filemap_get_folios(mapping, &next, end - 1, &fbatch)) {
552 for (i = 0; i < folio_batch_count(&fbatch); ++i) {
553 folio = fbatch.folios[i];
554
555 index = folio->index;
556 /*
557 * Take fault mutex for missing folios before index,
558 * while checking folios that might have been added
559 * due to a race with fault code.
560 */
561 freed += fault_lock_inode_indicies(h, inode, mapping,
> 562 m_start, m_index, truncate_op);
563
564 /*
565 * Remove folio that was part of folio_batch.
566 */
567 hash = hugetlb_fault_mutex_hash(mapping, index);
568 mutex_lock(&hugetlb_fault_mutex_table[hash]);
569 if (remove_inode_single_folio(h, inode, mapping, folio,
570 index, truncate_op))
571 freed++;
572 mutex_unlock(&hugetlb_fault_mutex_table[hash]);
573 }
574 folio_batch_release(&fbatch);
575 cond_resched();
576 }
577
578 /*
579 * Take fault mutex for missing folios at end of range while checking
580 * for folios that might have been added due to a race with fault code.
581 */
582 freed += fault_lock_inode_indicies(h, inode, mapping, m_start, m_end,
583 truncate_op);
584
585 if (truncate_op)
586 (void)hugetlb_unreserve_pages(inode, start, LONG_MAX, freed);
587 }
588

--
0-DAY CI Kernel Test Service
https://01.org/lkp

\
 
 \ /
  Last update: 2022-08-25 04:28    [W:0.032 / U:0.132 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site