binfmt_elf: use vmalloc() for allocation of vma_filesz

We have observed page allocations failures of order 4 during core dump
while trying to allocate vma_filesz.  This results in a useless core
file of size 0.  To improve reliability use vmalloc().

Note that the vmalloc() allocation is bounded by sysctl_max_map_count,
which is 65,530 by default.  So with a 4k page size, and 8 bytes per
seg, this is a max of 128 pages or an order 7 allocation.  Other parts
of the core dump path, such as fill_files_note() are already using
vmalloc() for presumably similar reasons.

Link: http://lkml.kernel.org/r/1479745791-17611-1-git-send-email-jbaron@akamai.com
Signed-off-by: Jason Baron <jbaron@akamai.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index 2472af2798c735d056d26775a17906db707aa9a7..e6c1bd44380633046e3172662dd23db54f49326d 100644 (file)
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -2204,7 +2204,9 @@ static int elf_core_dump(struct coredump_params *cprm)
 
        dataoff = offset = roundup(offset, ELF_EXEC_PAGESIZE);
 
-       vma_filesz = kmalloc_array(segs - 1, sizeof(*vma_filesz), GFP_KERNEL);
+       if (segs - 1 > ULONG_MAX / sizeof(*vma_filesz))
+               goto end_coredump;
+       vma_filesz = vmalloc((segs - 1) * sizeof(*vma_filesz));
        if (!vma_filesz)
                goto end_coredump;
 
@@ -2311,7 +2313,7 @@ end_coredump:
 cleanup:
        free_note_info(&info);
        kfree(shdr4extnum);
-       kfree(vma_filesz);
+       vfree(vma_filesz);
        kfree(phdr4note);
        kfree(elf);
 out: