[Crash-utility] Re: [PATCH v2] Fix a "Bus error" issue caused by 'crash --osrelease' or crash loading

Hi Lianbo, Thanks for the patch, I have run the test and no regressions found. So ack.

Thanks, Tao Liu On Fri, Aug 16, 2024 at 2:30 PM Lianbo Jiang <lijiang(a)redhat.com&gt; wrote: > > Sometimes, in a production environment, there are still some vmcores > that are incomplete, such as partial header or the data is corrupted. > When crash tool attempts to parse such vmcores, it may fail as below: > > $ ./crash --osrelease vmcore > Bus error (core dumped) > > or > > $ crash vmlinux vmcore > ... > Bus error (core dumped) > $ > > Gdb calltrace: > > $ gdb /home/lijiang/src/crash/crash /tmp/core.126301 > Core was generated by `./crash --osrelease /home/lijiang/src/39317/vmcore'. > Program terminated with signal SIGBUS, Bus error. > #0 __memcpy_evex_unaligned_erms () at ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:831 > 831 LOAD_ONE_SET((%rsi), PAGE_SIZE, %VMM(4), %VMM(5), %VMM(6), %VMM(7)) > (gdb) bt > #0 __memcpy_evex_unaligned_erms () at ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:831 > #1 0x0000000000651096 in read_dump_header (file=0x7ffc59ddff5f "/home/lijiang/src/39317/vmcore") at diskdump.c:820 > #2 0x0000000000651cf3 in is_diskdump (file=0x7ffc59ddff5f "/home/lijiang/src/39317/vmcore") at diskdump.c:1042 > #3 0x0000000000502ac9 in get_osrelease (dumpfile=0x7ffc59ddff5f "/home/lijiang/src/39317/vmcore") at main.c:1938 > #4 0x00000000004fb2e8 in main (argc=3, argv=0x7ffc59dde3a8) at main.c:271 > (gdb) frame 1 > #1 0x0000000000651096 in read_dump_header (file=0x7ffc59ddff5f "/home/lijiang/src/39317/vmcore") at diskdump.c:820 > 820 memcpy(dd->dumpable_bitmap, dd->bitmap + bitmap_len/2, > > This may happen on attempting access to a page of the buffer that lies > beyond the end of the mapped file(see the mmap() man page). > > Let's add a check to avoid such issues as much as possible, but still > not guarantee that it can work well in any extreme situation. > > Fixes: a3344239743b ("diskdump: use mmap/madvise to improve the start-up") > Reported-by: Buland Kumar Singh <bsingh(a)redhat.com&gt; > Signed-off-by: Lianbo Jiang <lijiang(a)redhat.com&gt; > --- > diskdump.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/diskdump.c b/diskdump.c > index 1f7118cacfc6..ce3cbb7b12dd 100644 > --- a/diskdump.c > +++ b/diskdump.c > @@ -805,6 +805,22 @@ restart: > goto err; > } > } else { > + struct stat sbuf; > + if (fstat(dd->dfd, &sbuf) != 0) { > + error(INFO, "Cannot fstat the dump file\n"); > + goto err; > + } > + > + /* > + * For memory regions mapped with the mmap(), attempts access to > + * a page of the buffer that lies beyond the end of the mapped file, > + * which may cause SIGBUS(see the mmap() man page). > + */ > + if (bitmap_len + offset > sbuf.st_size) { > + error(INFO, "Mmap: Beyond the end of mapped file, corrupted?\n"); > + goto err; > + } > + > dd->bitmap = mmap(NULL, bitmap_len, PROT_READ, > MAP_SHARED, dd->dfd, offset); > if (dd->bitmap == MAP_FAILED) > -- > 2.45.1