On 2020/07/01 19:10, Johannes Thumshirn wrote:
> Looks good,
> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@xxxxxxx>
>
> While we're at it the next block does a direct return manually unlocking
> 'ioc->pci_access_mutex' and rc is never set for any of the error paths
> in 'BRM_status_show'...
>
> Maybe we should add this one on top of your patch:
>
> diff --git a/drivers/scsi/mpt3sas/mpt3sas_ctl.c b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
> index 62e552838565..70d2d0987249 100644
> --- a/drivers/scsi/mpt3sas/mpt3sas_ctl.c
> +++ b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
> @@ -3149,20 +3149,20 @@ BRM_status_show(struct device *cdev, struct device_attribute *attr,
> }
> /* pci_access_mutex lock acquired by sysfs show path */
> mutex_lock(&ioc->pci_access_mutex);
> - if (ioc->pci_error_recovery || ioc->remove_host) {
> - mutex_unlock(&ioc->pci_access_mutex);
> - return 0;
> - }
> + if (ioc->pci_error_recovery || ioc->remove_host)
> + goto out;
>
> /* allocate upto GPIOVal 36 entries */
> sz = offsetof(Mpi2IOUnitPage3_t, GPIOVal) + (sizeof(u16) * 36);
> io_unit_pg3 = kzalloc(sz, GFP_KERNEL);
> if (!io_unit_pg3) {
> + rc = -ENOMEM;
> ioc_err(ioc, "%s: failed allocating memory for iounit_pg3: (%d) bytes\n",
> __func__, sz);
> goto out;
> }
>
> + rc = -EINVAL;
> if (mpt3sas_config_get_iounit_pg3(ioc, &mpi_reply, io_unit_pg3, sz) !=
> 0) {
> ioc_err(ioc, "%s: failed reading iounit_pg3\n",
>
Indeed... I did not look at the other early return path :)
Can you send something ?
--
Damien Le Moal
Western Digital Research
