diff options
author | Joe Lawrence <joe.lawrence@stratus.com> | 2014-06-25 17:06:54 -0400 |
---|---|---|
committer | Christoph Hellwig <hch@lst.de> | 2014-07-25 17:16:58 -0400 |
commit | 9f21316fc2f297efd32b40a57083d5cecb4bda26 (patch) | |
tree | 1174b29f2f5d6907ede37373227683ff6876c66c /drivers/message/fusion/mptbase.c | |
parent | 32696198745a74b37bc27498a5ade169d2c04c5e (diff) | |
download | op-kernel-dev-9f21316fc2f297efd32b40a57083d5cecb4bda26.zip op-kernel-dev-9f21316fc2f297efd32b40a57083d5cecb4bda26.tar.gz |
mptfusion: tweak null pointer checks
Fixes the following smatch warnings:
drivers/message/fusion/mptbase.c:652 mptbase_reply() warn: variable
dereferenced before check 'reply' (see line 639)
[JL: No-brainer, the enclosing switch statement dereferences
reply, so we can't get here unless reply is valid.]
drivers/message/fusion/mptsas.c:1255 mptsas_taskmgmt_complete() error:
we previously assumed 'pScsiTmReply' could be null (see line 1227)
[HCH: Reading the code in mptsas_taskmgmt_complete it's pretty
obvious that it can't do anything useful if mr/pScsiTmReply are
NULL, so I suspect it would be best to just return at the
beginning of the function.
I'd love to understand if it actually could ever be zero, which I
doubt. Maybe the LSI people can shed some light on that?]
drivers/message/fusion/mptsas.c:3888 mptsas_not_responding_devices()
error: we previously assumed 'port_info->phy_info' could be null
(see line 3875)
[HCH: It's pretty obvious from reading mptsas_sas_io_unit_pg0 that
we never register a port_info with a NULL phy_info in the lists,
so all NULL checks on it could be deleted.]
drivers/message/fusion/mptscsih.c:1284 mptscsih_info() error:
we previously assumed 'h' could be null (see line 1274)
[HCH: shost_priv can't return NULL, so the if (h) should be
removed.]
drivers/message/fusion/mptscsih.c:1388 mptscsih_qcmd() error: we
previously assumed 'vdevice' could be null (see line 1373)
[HCH: vdevice can't ever be NULL here, it's allocated in
->slave_alloc and thus guaranteed to be around when
->queuecommand is called.]
Signed-off-by: Joe Lawrence <joe.lawrence@stratus.com>
Acked-by: Sreekanth Reddy <Sreekanth.Reddy@avagotech.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Diffstat (limited to 'drivers/message/fusion/mptbase.c')
-rw-r--r-- | drivers/message/fusion/mptbase.c | 10 |
1 files changed, 4 insertions, 6 deletions
diff --git a/drivers/message/fusion/mptbase.c b/drivers/message/fusion/mptbase.c index 9d4c782..a896d94 100644 --- a/drivers/message/fusion/mptbase.c +++ b/drivers/message/fusion/mptbase.c @@ -649,12 +649,10 @@ mptbase_reply(MPT_ADAPTER *ioc, MPT_FRAME_HDR *req, MPT_FRAME_HDR *reply) case MPI_FUNCTION_CONFIG: case MPI_FUNCTION_SAS_IO_UNIT_CONTROL: ioc->mptbase_cmds.status |= MPT_MGMT_STATUS_COMMAND_GOOD; - if (reply) { - ioc->mptbase_cmds.status |= MPT_MGMT_STATUS_RF_VALID; - memcpy(ioc->mptbase_cmds.reply, reply, - min(MPT_DEFAULT_FRAME_SIZE, - 4 * reply->u.reply.MsgLength)); - } + ioc->mptbase_cmds.status |= MPT_MGMT_STATUS_RF_VALID; + memcpy(ioc->mptbase_cmds.reply, reply, + min(MPT_DEFAULT_FRAME_SIZE, + 4 * reply->u.reply.MsgLength)); if (ioc->mptbase_cmds.status & MPT_MGMT_STATUS_PENDING) { ioc->mptbase_cmds.status &= ~MPT_MGMT_STATUS_PENDING; complete(&ioc->mptbase_cmds.done); |