Skip to contentSkip to conversation

scsi: ufs: Fix a possible NULL pointer issue

PATCH In review Can Guo 2 participants 3 messages Dec 14, 2020 – Dec 18, 2020 SCSI

lore
Share & tools
lore

Can Guo

author
#0 lore raw

PATCH v2 1/1 scsi: ufs: Fix a possible NULL pointer issue

During system resume/suspend, hba could be NULL. In this case, do not touch eh_sem.

Fixes 88a92d6ae4fe ("scsi: ufs: Serialize eh_work with system PM events and async scan")
---
 drivers/scsi/ufs/ufshcd.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/scsi/ufs/ufshcd.c +5 -4
index c1c401b..ef155a9 100644
@@ -8883,8 +8883,11 @@ int ufshcd_system_suspend(struct ufs_hba *hba)
 	int ret = 0;
 	ktime_t start = ktime_get();
 
+	if (!hba)
+		return 0;
+
 	down(&hba->eh_sem);
-	if (!hba || !hba->is_powered)
+	if (!hba->is_powered)
 		return 0;
 
 	if ((ufs_get_pm_lvl_to_dev_pwr_mode(hba->spm_lvl) ==
@@ -8932,10 +8935,8 @@ int ufshcd_system_resume(struct ufs_hba *hba)
 	int ret = 0;
 	ktime_t start = ktime_get();
 
-	if (!hba) {
-		up(&hba->eh_sem);
+	if (!hba)
 		return -EINVAL;
-	}
 
 	if (!hba->is_powered || pm_runtime_suspended(hba->dev))
 		/*
Signature · 1 line
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.
#1 lore raw
Quoted 13 lines · Can Guo
Fixes 88a92d6ae4fe ("scsi: ufs: Serialize eh_work with system PM

events and async scan")

---
 drivers/scsi/ufs/ufshcd.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/scsi/ufs/ufshcd.c +0 -0
index c1c401b..ef155a9 100644

drivers/scsi/ufs/ufshcd.c int ufshcd_system_suspend(struct ufs_hba *hba) · line 8891

Quoted 10 lines
drivers/scsi/ufs/ufshcd.c +4 -1
@@ -8883,8 +8883,11 @@ int ufshcd_system_suspend(struct ufs_hba *hba)
        int ret = 0;
        ktime_t start = ktime_get();
 
+       if (!hba)
+               return 0;
+
        down(&hba->eh_sem);
-       if (!hba || !hba->is_powered)
+       if (!hba->is_powered)
                return 0;

Can,

why not moving down(&hba->eh_sem) after "return 0;"?

author
#2 lore raw
Quoted 28 lines · Bean Huo
Quoted 13 lines · Can Guo
Fixes 88a92d6ae4fe ("scsi: ufs: Serialize eh_work with system PM

events and async scan")

---
 drivers/scsi/ufs/ufshcd.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/scsi/ufs/ufshcd.c +0 -0
index c1c401b..ef155a9 100644

drivers/scsi/ufs/ufshcd.c int ufshcd_system_suspend(struct ufs_hba *hba) · line 8891

Quoted 10 lines
drivers/scsi/ufs/ufshcd.c +4 -1
@@ -8883,8 +8883,11 @@ int ufshcd_system_suspend(struct ufs_hba *hba)
        int ret = 0;
        ktime_t start = ktime_get();
 
+       if (!hba)
+               return 0;
+
        down(&hba->eh_sem);
-       if (!hba || !hba->is_powered)
+       if (!hba->is_powered)
                return 0;

Can,

Quoted 1 line

why not moving down(&hba->eh_sem) after "return 0;"?

In your way, if hba is not powered, ufshcd_system_suspend() returns 0, which is a successful suspend. When ufshcd_system_resume() is called, if hba is not powered, it goes to out and does up(&hba->eh_sem), which shall cause unbalance to eh_sem.

Thanks,

Can Guo.