2021-10-19 10:53:35

by Jiapeng Chong

Subject: [PATCH] scsi: core: return -ENOMEM on ips_init_phase1() allocation failure

From: chongjiapeng <[email protected]>

Fixes the following smatch warning:

drivers/scsi/ips.c:6901 ips_init_phase1() warn: returning -1 instead of
-ENOMEM is sloppy.

Reported-by: Abaci Robot <[email protected]>
Fixes: dd00cc486ab1 ("some kmalloc/memset ->kzalloc (tree wide)")
Signed-off-by: chongjiapeng <[email protected]>
---
drivers/scsi/ips.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/ips.c b/drivers/scsi/ips.c
index 498bf04499ce..2ce97f407d18 100644
--- a/drivers/scsi/ips.c
+++ b/drivers/scsi/ips.c
@@ -6898,7 +6898,7 @@ ips_init_phase1(struct pci_dev *pci_dev, int *indexPtr)
if (ha == NULL) {
IPS_PRINTK(KERN_WARNING, pci_dev,
"Unable to allocate temporary ha struct\n");
- return -1;
+ return -ENOMEM;
}

ips_sh[index] = NULL;
--
2.19.1.6.gb485710b

2021-10-19 11:51:19

by James Bottomley

Subject: Re: [PATCH] scsi: core: return -ENOMEM on ips_init_phase1() allocation failure

On Tue, 2021-10-19 at 18:50 +0800, Jiapeng Chong wrote:
> From: chongjiapeng <[email protected]>
>
> Fixes the following smatch warning:
>
> drivers/scsi/ips.c:6901 ips_init_phase1() warn: returning -1 instead
> of -ENOMEM is sloppy.

Really, like megaraid, it's not. ips_init_phase1() returns either
Success or -1. If it returns -1, that gets translated to -ENODEV so,
again, we'd lose this nuanced return completely.

James