Messages in this thread Patch in this message | | | Date | Thu, 18 Jul 2019 14:35:45 -0700 | From | Kees Cook <> | Subject | Re: BUG: KASAN: global-out-of-bounds in ata_exec_internal_sg+0x50f/0xc70 |
| |
On Tue, Jul 16, 2019 at 11:28:29AM -0700, Nick Desaulniers wrote: > On Wed, Jul 10, 2019 at 10:44 AM Jeffrin Thalakkottoor > <jeffrin@rajagiritech.edu.in> wrote: > > > > hello all , > > > > i encountered a KASAN bug related . here are some related information... > > > > > > -------------------x-----------------------------x------------------ > > [ 30.037312] BUG: KASAN: global-out-of-bounds in > > ata_exec_internal_sg+0x50f/0xc70 > > [ 30.037447] Read of size 16 at addr ffffffff91f41f80 by task scsi_eh_1/149 > > > > > > [ 30.039935] The buggy address belongs to the variable: > > [ 30.040059] cdb.48319+0x0/0x40 > > (gdb) l *ata_exec_internal_sg+0x50f > > 0xffffffff81c7b59f is in ata_exec_internal_sg (./include/linux/string.h:359). > > So looks like ata_exec_internal_sg() is panic'ing when... > > > 354 if (q_size < size) > > 355 __read_overflow2(); > > 356 } > > 357 if (p_size < size || q_size < size) > > 358 fortify_panic(__func__); > > 359 return __builtin_memcpy(p, q, size);
^^^ here, so within memcpy(), but after the "easy" sanity checks.
The only place where I see ata_exec_internal_sg() calling memcpy() is here:
/* prepare & issue qc */ qc->tf = *tf; if (cdb) memcpy(qc->cdb, cdb, ATAPI_CDB_LEN);
the "16" is consistent with the report:
include/linux/ata.h: ATAPI_CDB_LEN = 16,
which matches the claim about the cdb variable from KASAN. And it's a read, so "cdb" is wrong. Do you have a longer back trace? What called ata_exec_internal_sg()?
ata_exec_internal() is the only caller of ata_exec_internal_sg(). Nearly all callers of ata_exec_internal() pass a NULL cdb. Those that don't are:
atapi_eh_tur() u8 cdb[ATAPI_CDB_LEN] = ... atapi_eh_request_sense() u8 cdb[ATAPI_CDB_LEN] = ...
These two are on the static and correctly sized.
eject_tray() static const char cdb[ATAPI_CDB_LEN] = ... zpodd_get_mech_type() static const char cdb[] = ...
These are both in rodata, and only the first is correctly sized. I assume the following will fix it:
diff --git a/drivers/ata/libata-zpodd.c b/drivers/ata/libata-zpodd.c index 173e6f2dd9af..eefda51f97d3 100644 --- a/drivers/ata/libata-zpodd.c +++ b/drivers/ata/libata-zpodd.c @@ -56,7 +56,7 @@ static enum odd_mech_type zpodd_get_mech_type(struct ata_device *dev) unsigned int ret; struct rm_feature_desc *desc; struct ata_taskfile tf; - static const char cdb[] = { GPCMD_GET_CONFIGURATION, + static const char cdb[ATAPI_CDB_LEN] = { GPCMD_GET_CONFIGURATION, 2, /* only 1 feature descriptor requested */ 0, 3, /* 3, removable medium feature */ 0, 0, 0,/* reserved */
-- Kees Cook
| |