kexec: bail out upon SIGKILL when allocating memory.

syzbot found that a thread can stall for minutes inside kexec_load() after
that thread was killed by SIGKILL [1].  It turned out that the reproducer
was trying to allocate 2408MB of memory using kimage_alloc_page() from
kimage_load_normal_segment().  Let's check for SIGKILL before doing memory
allocation.

[1] https://syzkaller.appspot.com/bug?id=a0e3436829698d5824231251fad9d8e998f94f5e

Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Tetsuo Handa <[email protected]>
Reported-by: syzbot <[email protected]>
Cc: Eric Biederman <[email protected]>
Reviewed-by: Andrew Morton <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index d5870723b8ada61a757541e382f29886bdf7f214..15d70a90b50dcaee7e4345c45fe95d825a34cd8d 100644 (file)
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -300,6 +300,8 @@ static struct page *kimage_alloc_pages(gfp_t gfp_mask, unsigned int order)
 {
        struct page *pages;
 
+       if (fatal_signal_pending(current))
+               return NULL;
        pages = alloc_pages(gfp_mask & ~__GFP_ZERO, order);
        if (pages) {
                unsigned int count, i;