Skip to contentSkip to conversation

Two fixes for UFS

PATCH Applied Bean Huo 2 participants 6 messages Sep 29, 2021 – Oct 12, 2021 SCSI

lore
Share & tools
lore

Bean Huo

author
#0 lore raw

v3 cover Two fixes for UFS

From: Bean Huo <[email protected]>

V1--V2:
Fix a typo in the commit message
V2--V3:
1. Change patch [2/3] subject
2. Add new patch [3/3]

Bean Huo (3):
scsi: ufs: ufshpb: Fix NULL pointer dereference
scsi: ufs: core: fix ufshcd_probe_hba() prototype to match the
definition
scsi: ufs: core: Remove return statement in void function

 drivers/scsi/ufs/ufshcd.c | 3 +--
 drivers/scsi/ufs/ufshpb.c | 3 ---
 2 files changed, 1 insertion(+), 5 deletions(-)
Signature · 1 line
2.25.1
author
#1 lore raw

PATCH v3 1/3 scsi: ufs: ufshpb: Fix NULL pointer dereference

From: Bean Huo <[email protected]>

Before ufshcd_scsi_add_wlus() is executed, call ufshcd_rpm_{get/put}_sync(), "Null pointer dereference" issue will be triggered. Because hba->sdev_ufs_device is initialized in ufshcd_scsi_add_wlus() later.

Unable to handle kernel NULL pointer dereference at virtual address

0000000000000348
Mem abort info:
ESR = 0x96000004
EC = 0x25: DABT (current EL), IL = 32 bits SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC = 0x04: level 0 translation fault
Data abort info:
ISV = 0, ISS = 0x00000004
CM = 0, WnR = 0
[0000000000000348] user address but active_mm is swapper

Internal error: Oops: 96000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 0 PID: 91 Comm: kworker/u16:1 Not tainted 5.15.0-rc1-beanhuo-linaro-1423
Hardware name: MicronRB (DT)
Workqueue: events_unbound async_run_entry_fn
pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : pm_runtime_drop_link+0x128/0x338
lr : ufshpb_get_dev_info+0x8c/0x148

sp : ffff800012573c10
x29: ffff800012573c10 x28: 0000000000000000 x27: 0000000000000003
x26: ffff000001d21298 x25: 000000005abcea60 x24: ffff800011d89000
x23: 0000000000000001 x22: ffff000001d21880 x21: ffff000001ec9300
x20: 0000000000000004 x19: 0000000000000198 x18: ffffffffffffffff
x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000041400
x14: 5eee00201100200a x13: 000000000000bb03 x12: 0000000000000000
x11: 0000000000000100 x10: 0200000000000000 x9 : bb0000021a162c01 x8 : 0302010021021003 x7 : 0000000000000000 x6 : ffff800012573af0 x5 : 0000000000000001 x4 : 0000000000000001 x3 : 0000000000000200 x2 : 0000000000000348 x1 : 0000000000000348 x0 : ffff80001095308c

Call trace:
 pm_runtime_drop_link+0x128/0x338
 ufshpb_get_dev_info+0x8c/0x148
 ufshcd_probe_hba+0xda0/0x11b8
 ufshcd_async_scan+0x34/0x330
 async_run_entry_fn+0x38/0x180
 process_one_work+0x1f4/0x498
 worker_thread+0x48/0x480
 kthread+0x140/0x158
 ret_from_fork+0x10/0x20
Code: 88027c01 35ffffa2 17fff6c4 f9800051 (885f7c40)
---[ end trace 2ba541335f595c95 ]

Since ufshpb_get_dev_info() is only called in asynchronous execution ufshcd_async_scan(), and before ufshpb_get_dev_info(), pm_runtime_get_sync() has been called:

...
/* Hold auto suspend until async scan completes */ pm_runtime_get_sync(dev); atomic_set(&hba->scsi_block_reqs_cnt, 0);
...
ufshcd_async_scan()

    ufshcd_probe_hba(hba, true);
        ufshcd_device_params_init(hba);
            ufshpb_get_dev_info();
 ...
    pm_runtime_put_sync(hba->dev);

Remove ufshcd_rpm_{get/put}_sync() from ufshpb_get_dev_info(), fix this issue.

Fixes 351b3a849ac7 ("scsi: ufs: ufshpb: Use proper power management API")
---
 drivers/scsi/ufs/ufshpb.c | 3 ---
 1 file changed, 3 deletions(-)
drivers/scsi/ufs/ufshpb.c +0 -3
index 9ea639bf6a59..33a38c06a296 100644
@@ -2877,11 +2877,8 @@ void ufshpb_get_dev_info(struct ufs_hba *hba, u8 *desc_buf)
 	if (version == HPB_SUPPORT_LEGACY_VERSION)
 		hpb_dev_info->is_legacy = true;
 
-	ufshcd_rpm_get_sync(hba);
 	ret = ufshcd_query_attr_retry(hba, UPIU_QUERY_OPCODE_READ_ATTR,
 		QUERY_ATTR_IDN_MAX_HPB_SINGLE_CMD, 0, 0, &max_hpb_single_cmd);
-	ufshcd_rpm_put_sync(hba);
-
 	if (ret)
 		dev_err(hba->dev, "%s: idn: read max size of single hpb cmd query request failed",
 			__func__);
Signature · 1 line
2.25.1
author
#2 lore raw

PATCH v3 2/3 scsi: ufs: core: fix ufshcd_probe_hba() prototype to match the definition

In reply to Bean Huo

From: Bean Huo <[email protected]>

Since commit 568dd9959611 ("scsi: ufs: Rename the second ufshcd_probe_hba() argument"), the second ufshcd_probe_hba() argument has been changed to init_dev_params.

Fixes 568dd9959611 ("scsi: ufs: Rename the second ufshcd_probe_hba() argument")
---
 drivers/scsi/ufs/ufshcd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/ufs/ufshcd.c +1 -1
index debef631c89a..081092418e2d 100644
@@ -225,7 +225,7 @@ static int ufshcd_eh_host_reset_handler(struct scsi_cmnd *cmd);
 static int ufshcd_clear_tm_cmd(struct ufs_hba *hba, int tag);
 static void ufshcd_hba_exit(struct ufs_hba *hba);
 static int ufshcd_clear_ua_wluns(struct ufs_hba *hba);
-static int ufshcd_probe_hba(struct ufs_hba *hba, bool async);
+static int ufshcd_probe_hba(struct ufs_hba *hba, bool init_dev_params);
 static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on);
 static int ufshcd_uic_hibern8_enter(struct ufs_hba *hba);
 static inline void ufshcd_add_delay_before_dme_cmd(struct ufs_hba *hba);
Signature · 1 line
2.25.1
author
#3 lore raw

PATCH v3 3/3 scsi: ufs: core: Remove return statement in void function

In reply to Bean Huo

From: Bean Huo <[email protected]>

return statement is not useful at the end of "void" function.

---
 drivers/scsi/ufs/ufshcd.c | 1 -
 1 file changed, 1 deletion(-)
drivers/scsi/ufs/ufshcd.c +0 -1
index 081092418e2d..2f0366c3cbf8 100644
@@ -5845,7 +5845,6 @@ static void ufshcd_exception_event_handler(struct work_struct *work)
 	ufs_debugfs_exception_event(hba, status);
 out:
 	ufshcd_scsi_unblock_requests(hba);
-	return;
 }
 
 /* Complete requests that have door-bell cleared */
Signature · 1 line
2.25.1
#4 lore raw
In reply to Bean Huo

Bean,

Quoted 5 lines

Bean Huo (3):
scsi: ufs: ufshpb: Fix NULL pointer dereference
scsi: ufs: core: fix ufshcd_probe_hba() prototype to match the
definition
scsi: ufs: core: Remove return statement in void function

Applied to 5.16/scsi-staging, thanks!

Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering
#5 lore raw
In reply to Bean Huo
Quoted 9 lines · Bean Huo

From: Bean Huo <[email protected]>

V1--V2:
Fix a typo in the commit message
V2--V3:
1. Change patch [2/3] subject
2. Add new patch [3/3]

[...]

Applied to 5.16/scsi-queue, thanks!

[1/3] scsi: ufs: ufshpb: Fix NULL pointer dereference

      https://git.kernel.org/mkp/scsi/c/1da3b0141e74
[2/3] scsi: ufs: core: fix ufshcd_probe_hba() prototype to match the definition
      https://git.kernel.org/mkp/scsi/c/68444d73d6a5
[3/3] scsi: ufs: core: Remove return statement in void function
      https://git.kernel.org/mkp/scsi/c/f44abcfc3f9f
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering