Re: [RESEND PATCH v2] megaraid: fix out-of-bound array accesses

[Sumit Saxena <sumit.saxena@broadcom.com>]

On Thu, Dec 13, 2018 at 6:57 PM Qian Cai <cai@lca.pw> wrote:
>
> UBSAN reported those with MegaRAID SAS-3 3108,
>
> [   77.467308] UBSAN: Undefined behaviour in drivers/scsi/megaraid/megaraid_sas_fp.c:117:32
> [   77.475402] index 255 is out of range for type 'MR_LD_SPAN_MAP [1]'
> [   77.481677] CPU: 16 PID: 333 Comm: kworker/16:1 Not tainted 4.20.0-rc5+ #1
> [   77.488556] Hardware name: Huawei TaiShan 2280 /BC11SPCD, BIOS 1.50 06/01/2018
> [   77.495791] Workqueue: events work_for_cpu_fn
> [   77.500154] Call trace:
> [   77.502610]  dump_backtrace+0x0/0x2c8
> [   77.506279]  show_stack+0x24/0x30
> [   77.509604]  dump_stack+0x118/0x19c
> [   77.513098]  ubsan_epilogue+0x14/0x60
> [   77.516765]  __ubsan_handle_out_of_bounds+0xfc/0x13c
> [   77.521767]  mr_update_load_balance_params+0x150/0x158 [megaraid_sas]
> [   77.528230]  MR_ValidateMapInfo+0x2cc/0x10d0 [megaraid_sas]
> [   77.533825]  megasas_get_map_info+0x244/0x2f0 [megaraid_sas]
> [   77.539505]  megasas_init_adapter_fusion+0x9b0/0xf48 [megaraid_sas]
> [   77.545794]  megasas_init_fw+0x1ab4/0x3518 [megaraid_sas]
> [   77.551212]  megasas_probe_one+0x2c4/0xbe0 [megaraid_sas]
> [   77.556614]  local_pci_probe+0x7c/0xf0
> [   77.560365]  work_for_cpu_fn+0x34/0x50
> [   77.564118]  process_one_work+0x61c/0xf08
> [   77.568129]  worker_thread+0x534/0xa70
> [   77.571882]  kthread+0x1c8/0x1d0
> [   77.575114]  ret_from_fork+0x10/0x1c
>
> [   89.240332] UBSAN: Undefined behaviour in drivers/scsi/megaraid/megaraid_sas_fp.c:117:32
> [   89.248426] index 255 is out of range for type 'MR_LD_SPAN_MAP [1]'
> [   89.254700] CPU: 16 PID: 95 Comm: kworker/u130:0 Not tainted 4.20.0-rc5+ #1
> [   89.261665] Hardware name: Huawei TaiShan 2280 /BC11SPCD, BIOS 1.50 06/01/2018
> [   89.268903] Workqueue: events_unbound async_run_entry_fn
> [   89.274222] Call trace:
> [   89.276680]  dump_backtrace+0x0/0x2c8
> [   89.280348]  show_stack+0x24/0x30
> [   89.283671]  dump_stack+0x118/0x19c
> [   89.287167]  ubsan_epilogue+0x14/0x60
> [   89.290835]  __ubsan_handle_out_of_bounds+0xfc/0x13c
> [   89.295828]  MR_LdRaidGet+0x50/0x58 [megaraid_sas]
> [   89.300638]  megasas_build_io_fusion+0xbb8/0xd90 [megaraid_sas]
> [   89.306576]  megasas_build_and_issue_cmd_fusion+0x138/0x460 [megaraid_sas]
> [   89.313468]  megasas_queue_command+0x398/0x3d0 [megaraid_sas]
> [   89.319222]  scsi_dispatch_cmd+0x1dc/0x8a8
> [   89.323321]  scsi_request_fn+0x8e8/0xdd0
> [   89.327249]  __blk_run_queue+0xc4/0x158
> [   89.331090]  blk_execute_rq_nowait+0xf4/0x158
> [   89.335449]  blk_execute_rq+0xdc/0x158
> [   89.339202]  __scsi_execute+0x130/0x258
> [   89.343041]  scsi_probe_and_add_lun+0x2fc/0x1488
> [   89.347661]  __scsi_scan_target+0x1cc/0x8c8
> [   89.351848]  scsi_scan_channel.part.3+0x8c/0xc0
> [   89.356382]  scsi_scan_host_selected+0x130/0x1f0
> [   89.361002]  do_scsi_scan_host+0xd8/0xf0
> [   89.364927]  do_scan_async+0x9c/0x320
> [   89.368594]  async_run_entry_fn+0x138/0x420
> [   89.372780]  process_one_work+0x61c/0xf08
> [   89.376793]  worker_thread+0x13c/0xa70
> [   89.380546]  kthread+0x1c8/0x1d0
> [   89.383778]  ret_from_fork+0x10/0x1c
>
> This is because when populating Driver Map using firmware raid map, all
> non-existing VDs set their ldTgtIdToLd to 0xff, so it can be skipped
> later.
>
> From drivers/scsi/megaraid/megaraid_sas_base.c ,
> memset(instance->ld_ids, 0xff, MEGASAS_MAX_LD_IDS);
>
> From drivers/scsi/megaraid/megaraid_sas_fp.c ,
> /* For non existing VDs, iterate to next VD*/
> if (ld >= (MAX_LOGICAL_DRIVES_EXT - 1))
>         continue;
>
> However, there are a few places that failed to skip those non-existing
> VDs due to off-by-one errors. Then, those 0xff leaked into
> MR_LdRaidGet(0xff, map) and triggered the out-of-bound accesses.
>
> Fixes: 51087a8617fe (megaraid_sas : Extended VD support)
> Signed-off-by: Qian Cai <cai@lca.pw>

Acked-by: Sumit Saxena <sumit.saxena@broadcom.com>
> ---
>  drivers/scsi/megaraid/megaraid_sas_fp.c     | 2 +-
>  drivers/scsi/megaraid/megaraid_sas_fusion.c | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/megaraid/megaraid_sas_fp.c b/drivers/scsi/megaraid/megaraid_sas_fp.c
> index 59ecbb3b53b5..a33628550425 100644
> --- a/drivers/scsi/megaraid/megaraid_sas_fp.c
> +++ b/drivers/scsi/megaraid/megaraid_sas_fp.c
> @@ -1266,7 +1266,7 @@ void mr_update_load_balance_params(struct MR_DRV_RAID_MAP_ALL *drv_map,
>
>         for (ldCount = 0; ldCount < MAX_LOGICAL_DRIVES_EXT; ldCount++) {
>                 ld = MR_TargetIdToLdGet(ldCount, drv_map);
> -               if (ld >= MAX_LOGICAL_DRIVES_EXT) {
> +               if (ld >= MAX_LOGICAL_DRIVES_EXT - 1) {
>                         lbInfo[ldCount].loadBalanceFlag = 0;
>                         continue;
>                 }
> diff --git a/drivers/scsi/megaraid/megaraid_sas_fusion.c b/drivers/scsi/megaraid/megaraid_sas_fusion.c
> index f74b5ea24f0f..49eaa87608f6 100644
> --- a/drivers/scsi/megaraid/megaraid_sas_fusion.c
> +++ b/drivers/scsi/megaraid/megaraid_sas_fusion.c
> @@ -2832,7 +2832,7 @@ static void megasas_build_ld_nonrw_fusion(struct megasas_instance *instance,
>                 device_id < instance->fw_supported_vd_count)) {
>
>                 ld = MR_TargetIdToLdGet(device_id, local_map_ptr);
> -               if (ld >= instance->fw_supported_vd_count)
> +               if (ld >= instance->fw_supported_vd_count - 1)
>                         fp_possible = 0;
>                 else {
>                         raid = MR_LdRaidGet(ld, local_map_ptr);
> --
> 2.17.2 (Apple Git-113)
>