[PATCH 4.4 32/91] mpt3sas: Dont overreach ioc->reply_post[] during initialization

From: Greg Kroah-Hartman
Date: Fri Aug 04 2017 - 19:56:21 EST


4.4-stable review patch. If anyone has any objections, please let me know.

------------------

From: Calvin Owens <calvinowens@xxxxxx>

commit 5ec8a1753bc29efa7e4b1391d691c9c719b30257 upstream.

In _base_make_ioc_operational(), we walk ioc->reply_queue_list and pull
a pointer out of successive elements of ioc->reply_post[] for each entry
in that list if RDPQ is enabled.

Since the code pulls the pointer for the next iteration at the bottom of
the loop, it triggers the a KASAN dump on the final iteration:

BUG: KASAN: slab-out-of-bounds in _base_make_ioc_operational+0x47b7/0x47e0 [mpt3sas] at addr ffff880754816ab0
Read of size 8 by task modprobe/305
<snip>
Call Trace:
[<ffffffff81dfc591>] dump_stack+0x4d/0x6c
[<ffffffff814c9689>] print_trailer+0xf9/0x150
[<ffffffff814ceda4>] object_err+0x34/0x40
[<ffffffff814d1231>] kasan_report_error+0x221/0x530
[<ffffffff814d1673>] __asan_report_load8_noabort+0x43/0x50
[<ffffffffa0043637>] _base_make_ioc_operational+0x47b7/0x47e0 [mpt3sas]
[<ffffffffa0049a51>] mpt3sas_base_attach+0x1991/0x2120 [mpt3sas]
[<ffffffffa0053c93>] _scsih_probe+0xeb3/0x16b0 [mpt3sas]
[<ffffffff81ebd047>] local_pci_probe+0xc7/0x170
[<ffffffff81ebf2cf>] pci_device_probe+0x20f/0x290
[<ffffffff820d50cd>] really_probe+0x17d/0x600
[<ffffffff820d56a3>] __driver_attach+0x153/0x190
[<ffffffff820cffac>] bus_for_each_dev+0x11c/0x1a0
[<ffffffff820d421d>] driver_attach+0x3d/0x50
[<ffffffff820d378a>] bus_add_driver+0x44a/0x5f0
[<ffffffff820d666c>] driver_register+0x18c/0x3b0
[<ffffffff81ebcb76>] __pci_register_driver+0x156/0x200
[<ffffffffa00c8135>] _mpt3sas_init+0x135/0x1000 [mpt3sas]
[<ffffffff81000423>] do_one_initcall+0x113/0x2b0
[<ffffffff813caa5a>] do_init_module+0x1d0/0x4d8
[<ffffffff81273909>] load_module+0x6729/0x8dc0
[<ffffffff81276123>] SYSC_init_module+0x183/0x1a0
[<ffffffff8127625e>] SyS_init_module+0xe/0x10
[<ffffffff828fe7d7>] entry_SYSCALL_64_fastpath+0x12/0x6a

Fix this by pulling the value at the beginning of the loop.

Signed-off-by: Calvin Owens <calvinowens@xxxxxx>
Reviewed-by: Johannes Thumshirn <jthumshirn@xxxxxxx>
Reviewed-by: Jens Axboe <axboe@xxxxxx>
Acked-by: Chaitra Basappa <chaitra.basappa@xxxxxxxxxxxx>
Signed-off-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx>
Signed-off-by: Amit Pundir <amit.pundir@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
drivers/scsi/mpt3sas/mpt3sas_base.c | 33 ++++++++++++++++-----------------
1 file changed, 16 insertions(+), 17 deletions(-)

--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -4981,15 +4981,14 @@ _base_make_ioc_ready(struct MPT3SAS_ADAP
static int
_base_make_ioc_operational(struct MPT3SAS_ADAPTER *ioc, int sleep_flag)
{
- int r, i;
+ int r, i, index;
unsigned long flags;
u32 reply_address;
u16 smid;
struct _tr_list *delayed_tr, *delayed_tr_next;
u8 hide_flag;
struct adapter_reply_queue *reply_q;
- long reply_post_free;
- u32 reply_post_free_sz, index = 0;
+ Mpi2ReplyDescriptorsUnion_t *reply_post_free_contig;

dinitprintk(ioc, pr_info(MPT3SAS_FMT "%s\n", ioc->name,
__func__));
@@ -5061,27 +5060,27 @@ _base_make_ioc_operational(struct MPT3SA
_base_assign_reply_queues(ioc);

/* initialize Reply Post Free Queue */
- reply_post_free_sz = ioc->reply_post_queue_depth *
- sizeof(Mpi2DefaultReplyDescriptor_t);
- reply_post_free = (long)ioc->reply_post[index].reply_post_free;
+ index = 0;
+ reply_post_free_contig = ioc->reply_post[0].reply_post_free;
list_for_each_entry(reply_q, &ioc->reply_queue_list, list) {
+ /*
+ * If RDPQ is enabled, switch to the next allocation.
+ * Otherwise advance within the contiguous region.
+ */
+ if (ioc->rdpq_array_enable) {
+ reply_q->reply_post_free =
+ ioc->reply_post[index++].reply_post_free;
+ } else {
+ reply_q->reply_post_free = reply_post_free_contig;
+ reply_post_free_contig += ioc->reply_post_queue_depth;
+ }
+
reply_q->reply_post_host_index = 0;
- reply_q->reply_post_free = (Mpi2ReplyDescriptorsUnion_t *)
- reply_post_free;
for (i = 0; i < ioc->reply_post_queue_depth; i++)
reply_q->reply_post_free[i].Words =
cpu_to_le64(ULLONG_MAX);
if (!_base_is_controller_msix_enabled(ioc))
goto skip_init_reply_post_free_queue;
- /*
- * If RDPQ is enabled, switch to the next allocation.
- * Otherwise advance within the contiguous region.
- */
- if (ioc->rdpq_array_enable)
- reply_post_free = (long)
- ioc->reply_post[++index].reply_post_free;
- else
- reply_post_free += reply_post_free_sz;
}
skip_init_reply_post_free_queue: