linux/fs/seq_file.c, branch v3.19

fs, seq_file: fallback to vmalloc instead of oom kill processes

2014-12-13T20:42:49Z

Since commit 058504edd026 ("fs/seq_file: fallback to vmalloc allocation"), seq_buf_alloc() falls back to vmalloc() when the kmalloc() for contiguous memory fails. This was done to address order-4 slab allocations for reading /proc/stat on large machines and noticed because PAGE_ALLOC_COSTLY_ORDER < 4, so there is no infinite loop in the page allocator when allocating new slab for such high-order allocations. Contiguous memory isn't necessary for caller of seq_buf_alloc(), however. Other GFP_KERNEL high-order allocations that are <= PAGE_ALLOC_COSTLY_ORDER will simply loop forever in the page allocator and oom kill processes as a result. We don't want to kill processes so that we can allocate contiguous memory in situations when contiguous memory isn't necessary. This patch does the kmalloc() allocation with __GFP_NORETRY for high-order allocations. This still utilizes memory compaction and direct reclaim in the allocation path, the only difference is that it will fail immediately instead of oom kill processes when out of memory. [akpm@linux-foundation.org: add comment] Signed-off-by: David Rientjes Cc: Heiko Carstens Cc: Christoph Hellwig Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

seq_file: Rename seq_overflow() to seq_has_overflowed() and make public

2014-10-30T00:26:06Z

The return values of seq_printf/puts/putc are frequently misused. Start down a path to remove all the return value uses of these functions. Move the seq_overflow() to a global inlined function called seq_has_overflowed() that can be used by the users of seq_file() calls. Update the documentation to not show return types for seq_printf et al. Add a description of seq_has_overflowed(). Link: http://lkml.kernel.org/p/848ac7e3d1c31cddf638a8526fa3c59fa6fdeb8a.1412031505.git.joe@perches.com Cc: Al Viro Signed-off-by: Joe Perches [ Reworked the original patch from Joe ] Signed-off-by: Steven Rostedt

fs/seq_file: fallback to vmalloc allocation

2014-07-03T16:21:54Z

There are a couple of seq_files which use the single_open() interface. This interface requires that the whole output must fit into a single buffer. E.g. for /proc/stat allocation failures have been observed because an order-4 memory allocation failed due to memory fragmentation. In such situations reading /proc/stat is not possible anymore. Therefore change the seq_file code to fallback to vmalloc allocations which will usually result in a couple of order-0 allocations and hence also work if memory is fragmented. For reference a call trace where reading from /proc/stat failed: sadc: page allocation failure: order:4, mode:0x1040d0 CPU: 1 PID: 192063 Comm: sadc Not tainted 3.10.0-123.el7.s390x #1 [...] Call Trace: show_stack+0x6c/0xe8 warn_alloc_failed+0xd6/0x138 __alloc_pages_nodemask+0x9da/0xb68 __get_free_pages+0x2e/0x58 kmalloc_order_trace+0x44/0xc0 stat_open+0x5a/0xd8 proc_reg_open+0x8a/0x140 do_dentry_open+0x1bc/0x2c8 finish_open+0x46/0x60 do_last+0x382/0x10d0 path_openat+0xc8/0x4f8 do_filp_open+0x46/0xa8 do_sys_open+0x114/0x1f0 sysc_tracego+0x14/0x1a Signed-off-by: Heiko Carstens Tested-by: David Rientjes Cc: Ian Kent Cc: Hendrik Brueckner Cc: Thorsten Diehl Cc: Andrea Righi Cc: Christoph Hellwig Cc: Al Viro Cc: Stefan Bader Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

seq_file: always clear m->count when we free m->buf

2013-11-19T03:07:53Z

Once we'd freed m->buf, m->count should become zero - we have no valid contents reachable via m->buf. Reported-by: Charley (Hao Chuan) Chu Signed-off-by: Al Viro Signed-off-by: Linus Torvalds

seq_file: introduce seq_setwidth() and seq_pad()

2013-11-15T00:32:20Z

There are several users who want to know bytes written by seq_*() for alignment purpose. Currently they are using %n format for knowing it because seq_*() returns 0 on success. This patch introduces seq_setwidth() and seq_pad() for allowing them to align without using %n format. Signed-off-by: Tetsuo Handa Signed-off-by: Kees Cook Cc: Joe Perches Cc: David Miller Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

seq_file: always update file->f_pos in seq_lseek()

2013-10-25T14:46:40Z

This issue was first pointed out by Jiaxing Wang several months ago, but no further comments: https://lkml.org/lkml/2013/6/29/41 As we know pread() does not change f_pos, so after pread(), file->f_pos and m->read_pos become different. And seq_lseek() does not update file->f_pos if offset equals to m->read_pos, so after pread() and seq_lseek()(lseek to m->read_pos), then a subsequent read may read from a wrong position, the following program produces the problem: char str1[32] = { 0 }; char str2[32] = { 0 }; int poffset = 10; int count = 20; /*open any seq file*/ int fd = open("/proc/modules", O_RDONLY); pread(fd, str1, count, poffset); printf("pread:%s\n", str1); /*seek to where m->read_pos is*/ lseek(fd, poffset+count, SEEK_SET); /*supposed to read from poffset+count, but this read from position 0*/ read(fd, str2, count); printf("read:%s\n", str2); out put: pread: ck_netbios_ns 12665 read: nf_conntrack_netbios /proc/modules: nf_conntrack_netbios_ns 12665 0 - Live 0xffffffffa038b000 nf_conntrack_broadcast 12589 1 nf_conntrack_netbios_ns, Live 0xffffffffa0386000 So we always update file->f_pos to offset in seq_lseek() to fix this issue. Signed-off-by: Jiaxing Wang Signed-off-by: Gu Zheng Signed-off-by: Al Viro

seq_file: add seq_list_*_percpu helpers

2013-07-08T09:36:41Z

When we convert the file_lock_list to a set of percpu lists, we'll need a way to iterate over them in order to output /proc/locks info. Add some seq_list_*_percpu helpers to handle that. Signed-off-by: Jeff Layton Acked-by: J. Bruce Fields Signed-off-by: Al Viro

new helper: single_open_size()

2013-04-09T18:13:29Z

Same as single_open(), but preallocates the buffer of given size. Doesn't make any sense for sizes up to PAGE_SIZE and doesn't make sense if output of show() exceeds PAGE_SIZE only rarely - seq_read() will take care of growing the buffer and redoing show(). If you _know_ that it will be large, it might make more sense to look into saner iterator, rather than go with single-shot one. If that's impossible, single_open_size() might be for you. Again, don't use that without a good reason; occasionally that's really the best way to go, but very often there are better solutions. Signed-off-by: Al Viro

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

2013-03-03T21:23:03Z

Pull more VFS bits from Al Viro: "Unfortunately, it looks like xattr series will have to wait until the next cycle ;-/ This pile contains 9p cleanups and fixes (races in v9fs_fid_add() etc), fixup for nommu breakage in shmem.c, several cleanups and a bit more file_inode() work" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: constify path_get/path_put and fs_struct.c stuff fix nommu breakage in shmem.c cache the value of file_inode() in struct file 9p: if v9fs_fid_lookup() gets to asking server, it'd better have hashed dentry 9p: make sure ->lookup() adds fid to the right dentry 9p: untangle ->lookup() a bit 9p: double iput() in ->lookup() if d_materialise_unique() fails 9p: v9fs_fid_add() can't fail now v9fs: get rid of v9fs_dentry 9p: turn fid->dlist into hlist 9p: don't bother with private lock in ->d_fsdata; dentry->d_lock will do just fine more file_inode() open-coded instances selinux: opened file can't have NULL or negative ->f_path.dentry (In the meantime, the hlist traversal macros have changed, so this required a semantic conflict fixup for the newly hlistified fid->dlist)

fs/seq_file.c:seq_lseek(): fix switch statement indenting

2013-02-28T03:10:12Z

[akpm@linux-foundation.org: checkpatch fixes] Cc: Cyrill Gorcunov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds