While converting SATA on Apalis iMX6 to use driver model I noticed it
v3 cover While converting SATA on Apalis iMX6 to use driver model I noticed it
crashing in case no SATA driver is actually there:
Apalis iMX6 # sata init
data abort
pc : [<8ff78e1a>] lr : [<8ff78e1b>]
reloc pc : [<1781be1a>] lr : [<1781be1b>]
sp : 8df53068 ip : 00000001 fp : 00000002
r10: 8df612e8 r9 : 8df5ceb0 r8 : 8ffc2b28
r7 : 8ff746ed r6 : 00000000 r5 : 68abf82a r4 : 00000000
r3 : 8ff73d39 r2 : c0000000 r1 : 00000000 r0 : 0000000c
Flags: nzcv IRQs off FIQs off Mode SVC_32
Code: 47700025 6803b570 6bdd4604 f0254805 (68abf82a)
Resetting CPU ...
resetting ...
Investigating this I identified 3 places where no proper checking is done.
This patch set adds such checks leading to a graceful error instead:
Apalis iMX6 # sata init
Cannot probe SATA device 0 (err=-19)
Changes in v3:
- Added Simon's reviewed-by.
Changes in v2:
- Update dm_test_uclass_devices_find() to test this behaviour as suggested by Simon.
- Dropped "[PATCH 2/3] dm: sata: add null pointer check for dev" as suggested by Simon.
- Added Simon's reviewed-by.
Marcel Ziswiler (2):
dm: device: fail uclass_find_first_device() if list_empty
cmd: sata: add null pointer check for dev
cmd/sata.c | 4 ++++ drivers/core/uclass.c | 2 +- test/dm/core.c | 4 ++++ 3 files changed, 9 insertions(+), 1 deletion(-)
Signature · 1 line
2.20.1
PATCH v3 1/2 dm: device: fail uclass_find_first_device() if list_empty
From: Marcel Ziswiler <[email protected]>
While uclass_find_device() fails with -ENODEV in case of list_empty strangely uclass_find_first_device() returns 0.
Fix uclass_find_first_device() to also fail with -ENODEV instead.
---
Changes in v3:
- Added Simon's reviewed-by.
Changes in v2:
- Update dm_test_uclass_devices_find() to test this behaviour as suggested by Simon.
drivers/core/uclass.c | 2 +- test/dm/core.c | 4 ++++ 2 files changed, 5 insertions(+), 1 deletion(-)
drivers/core/uclass.c +1 -1
@@ -225,7 +225,7 @@ int uclass_find_first_device(enum uclass_id id, struct udevice **devp)if (ret)return ret;if (list_empty (&uc->dev_head)) -return 0 ; +return -ENODEV ; *devp = list_first_entry(&uc->dev_head, struct udevice, uclass_node);
test/dm/core.c +4 -0
@@ -749,6 +749,10 @@ static int dm_test_uclass_devices_find(struct unit_test_state *uts)ut_assert (dev); } + ret =uclass_find_first_device (UCLASS_TEST_DUMMY , &dev); +ut_assert (ret == -ENODEV ); +ut_assert (!dev); +return 0 ; }DM_TEST (dm_test_uclass_devices_find,DM_TESTF_SCAN_PDATA );
Signature · 1 line
2.20.1
PATCH v3 1/2 dm: device: fail uclass_find_first_device() if list_empty
Quoted 9 lines · Marcel Ziswiler
From: Marcel Ziswiler <[email protected]>
While
uclass_find_device()fails with-ENODEVin case oflist_emptystrangelyuclass_find_first_device()returns 0.Fix
uclass_find_first_device()to also fail with-ENODEVinstead.Signed-off-by Marcel Ziswiler <[email protected]>Reviewed-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190210/6fa0533d/attachment.sig>
PATCH v3 2/2 cmd: sata: add null pointer check for dev
In reply to Marcel ZiswilerFrom: Marcel Ziswiler <[email protected]>
Calling sata_scan() with a null pointer probably won't make much sense.
---
Changes in v3: None
Changes in v2:
- Dropped "[PATCH 2/3] dm: sata: add null pointer check for dev" as suggested by Simon.
- Added Simon's reviewed-by.
cmd/sata.c | 4 ++++ 1 file changed, 4 insertions(+)
cmd/sata.c +4 -0
@@ -60,6 +60,10 @@ int sata_probe(int devnum)printf ("Cannot probe SATA device %d (err=%d)\n" , devnum, rc);return CMD_RET_FAILURE ; } +if (!dev) { +printf ("No SATA device found!\n" ); +return CMD_RET_FAILURE ; + } rc =sata_scan (dev);if (rc) {printf ("Cannot scan SATA device %d (err=%d)\n" , devnum, rc);
Signature · 1 line
2.20.1
PATCH v3 2/2 cmd: sata: add null pointer check for dev
Quoted 6 lines · Marcel Ziswiler
From: Marcel Ziswiler <[email protected]>
Calling
sata_scan()with a null pointer probably won't make much sense.Signed-off-by Marcel Ziswiler <[email protected]>Reviewed-by Simon Glass <[email protected]>
Applied to u-boot/master, thanks!
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190210/34c462fd/attachment.sig>