Messages in this thread | | | Subject | Re: [PATCH] scsi: target: configfs: Remove redundant assignment to ret | From | Bodo Stroesser <> | Date | Fri, 30 Apr 2021 16:29:11 +0200 |
| |
On 30.04.21 11:23, Jiapeng Chong wrote: > Variable ret is set to '-EINVAL' but this value is never read as it is > overwritten later on, hence it is a redundant assignment and can be > removed. > > Clean up the following clang-analyzer warning: > > drivers/target/target_core_configfs.c:2037:5: warning: Value stored to > 'ret' is never read [clang-analyzer-deadcode.DeadStores]. > > drivers/target/target_core_configfs.c:1973:5: warning: Value stored to > 'ret' is never read [clang-analyzer-deadcode.DeadStores]. > > drivers/target/target_core_configfs.c:1959:5: warning: Value stored to > 'ret' is never read [clang-analyzer-deadcode.DeadStores]. > > Reported-by: Abaci Robot <abaci@linux.alibaba.com> > Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com> > --- > drivers/target/target_core_configfs.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/drivers/target/target_core_configfs.c b/drivers/target/target_core_configfs.c > index 4b2e493..bda05c3 100644 > --- a/drivers/target/target_core_configfs.c > +++ b/drivers/target/target_core_configfs.c > @@ -1956,7 +1956,6 @@ static ssize_t target_pr_res_aptpl_metadata_store(struct config_item *item, > pr_err("APTPL metadata initiator_node=" > " exceeds PR_APTPL_MAX_IPORT_LEN: %d\n", > PR_APTPL_MAX_IPORT_LEN); > - ret = -EINVAL; > break; > } > break; > @@ -1970,7 +1969,6 @@ static ssize_t target_pr_res_aptpl_metadata_store(struct config_item *item, > pr_err("APTPL metadata initiator_isid" > "= exceeds PR_REG_ISID_LEN: %d\n", > PR_REG_ISID_LEN); > - ret = -EINVAL; > break; > } > break; > @@ -2034,7 +2032,6 @@ static ssize_t target_pr_res_aptpl_metadata_store(struct config_item *item, > pr_err("APTPL metadata target_node=" > " exceeds PR_APTPL_MAX_TPORT_LEN: %d\n", > PR_APTPL_MAX_TPORT_LEN); > - ret = -EINVAL; > break; > } > break; >
I don't think storing -EINVAL in ret is wrong. But maybe the "break" in the next line should better be "goto out"?
AFAICS, the "break" finally leads to calling core_scsi3_alloc_aptpl_registration () despite a too long string in i_port, isid or t_port. Is this behaviour intentional?
BTW: why is "initiator_sid" input handled as string? core_scsi3_alloc_aptpl_registration later calls get_unaligned_be64 for that string. I don't think this makes sense.
Finally, I think there is a possible memory leak. If the input written to res_aptpl_metadata for example contains "initiator_sid=xxxxx" more than once, a previously created string isid ("isid = match_strdup(args)") is not freed before it is overwritten in the loop. The same is true for all tokens handled with match_strdup.
| |