Skip to contentSkip to conversation

scsi: ufs: clean up ufshpb_check_hpb_reset_query()

PATCH In review Dan Carpenter 2 participants 4 messages May 31, 2022 – Jun 1, 2022

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi: ufs: clean up ufshpb_check_hpb_reset_query()

Smatch complains that the if (flag_res) is not required:

drivers/ufs/core/ufshpb.c:2306 ufshpb_check_hpb_reset_query()
warn: duplicate check 'flag_res' (previous on line 2301)

Re-write the "if (flag_res)" checking to be more clear.

---
 drivers/ufs/core/ufshpb.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)
drivers/ufs/core/ufshpb.c +6 -8
index fb122eaed28b..95b501b824df 100644
@@ -2299,17 +2299,15 @@ static bool ufshpb_check_hpb_reset_query(struct ufs_hba *hba)
 		}
 
 		if (!flag_res)
-			goto out;
+			return false;
 
 		usleep_range(1000, 1100);
 	}
-	if (flag_res) {
-		dev_err(hba->dev,
-			"%s fHpbReset was not cleared by the device\n",
-			__func__);
-	}
-out:
-	return flag_res;
+
+	dev_err(hba->dev,
+		"%s fHpbReset was not cleared by the device\n",
+		__func__);
+	return true;
 }
 
 /**
Signature · 1 line
2.35.1
reviewer
#1 lore raw
Quoted 8 lines

Smatch complains that the if (flag_res) is not required:

drivers/ufs/core/ufshpb.c:2306 ufshpb_check_hpb_reset_query()
warn: duplicate check 'flag_res' (previous on line 2301)

Re-write the "if (flag_res)" checking to be more clear.

In HPB Reset, the Host set this flag as '1' to inform the device that host reset its L2P data. The Device resets flag as '0' when the device inactivated all region information. 0h: HPB reset completed or not started yet.
1h: HPB reset in progress.

Would make sense to me to contain this logic within this function, Instead of returning just the flag value.

Thanks,
Avri

Quoted 35 lines
---
 drivers/ufs/core/ufshpb.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)
drivers/ufs/core/ufshpb.c +6 -8
index fb122eaed28b..95b501b824df 100644
@@ -2299,17 +2299,15 @@ static bool ufshpb_check_hpb_reset_query(struct
 ufs_hba *hba)
                }
 
                if (!flag_res)
-                       goto out;
+                       return false;
 
                usleep_range(1000, 1100);
        }
-       if (flag_res) {
-               dev_err(hba->dev,
-                       "%s fHpbReset was not cleared by the device\n",
-                       __func__);
-       }
-out:
-       return flag_res;
+
+       dev_err(hba->dev,
+               "%s fHpbReset was not cleared by the device\n",
+               __func__);
+       return true;
 }
 

/**
--
2.35.1

#2 lore raw
Quoted 13 lines · Avri Altman
Quoted 8 lines

Smatch complains that the if (flag_res) is not required:

drivers/ufs/core/ufshpb.c:2306 ufshpb_check_hpb_reset_query()
warn: duplicate check 'flag_res' (previous on line 2301)

Re-write the "if (flag_res)" checking to be more clear.

In HPB Reset, the Host set this flag as '1' to inform the device that host reset its L2P data. The Device resets flag as '0' when the device inactivated all region information. 0h: HPB reset completed or not started yet.
1h: HPB reset in progress.

Quoted 5 lines

Would make sense to me to contain this logic within this function, Instead of returning just the flag value.

Thanks,
Avri

I am not sure I understand.

To be honest, this function is not beautiful at all. With boolean functions, the name should tell you what the return means. Examples
are: if (!access_ok()), if (IS_ERR() etc. In this case the return is not clear from the name.

The second thing is that I really don't like returning true for failure and returning false for success. Returning zero and negatives is good but with true/false it should be true == success.

So, yes, I wasn't super happy with this function either. But I just did a minimal clean up to make the returns more clear. If you want to drop this patch and write a more extensive one then I would be really happy about that.

regards,
dan carpenter

reviewer
#3 lore raw
Quoted 22 lines · Avri Altman
Quoted 9 lines
Quoted 8 lines

Smatch complains that the if (flag_res) is not required:

drivers/ufs/core/ufshpb.c:2306 ufshpb_check_hpb_reset_query()
warn: duplicate check 'flag_res' (previous on line 2301)

Re-write the "if (flag_res)" checking to be more clear.

In HPB Reset, the Host set this flag as '1' to inform the device that host

reset its L2P data.

Quoted 1 line

The Device resets flag as '0' when the device inactivated all region

information.

Quoted 8 lines

0h: HPB reset completed or not started yet.
1h: HPB reset in progress.

Would make sense to me to contain this logic within this function, Instead of returning just the flag value.

Thanks,
Avri

Quoted 6 lines

I am not sure I understand.

To be honest, this function is not beautiful at all. With boolean functions, the name should tell you what the return means. Examples
are: if (!access_ok()), if (IS_ERR() etc. In this case the return is not clear from the name.

Right.

Quoted 4 lines

The second thing is that I really don't like returning true for failure and returning false for success. Returning zero and negatives is good but with true/false it should be true == success.

Exactly.

Quoted 5 lines

So, yes, I wasn't super happy with this function either. But I just did a minimal clean up to make the returns more clear. If you want to drop this patch and write a more extensive one then I would be really happy about that.

Yeah - I will.

Thanks,
Avri

Quoted 3 lines

regards,
dan carpenter