Skip to contentSkip to conversation

ptp: ptp_clockmatrix: initialize variables

PATCH In review [email protected] 2 participants 4 messages Oct 11, 2020 – Oct 13, 2020

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

PATCH 1/1 ptp: ptp_clockmatrix: initialize variables

From: Tom Rix <[email protected]>

Clang static analysis reports this representative problem

ptp_clockmatrix.c:1852:2: warning: 5th function call argument
is an uninitialized value

        snprintf(idtcm->version, sizeof(idtcm->version), "%u.%u.%u",
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

idtcm_display_version_info() calls several idtcm_read_*'s without checking a return status. Initialize the read variables so if a read fails, a garbage value is not displayed.

---
 drivers/ptp/ptp_clockmatrix.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/ptp/ptp_clockmatrix.c +6 -6
index e020faff7da5..47e5e62da5d2 100644
@@ -1832,12 +1832,12 @@ static int idtcm_enable_tod(struct idtcm_channel *channel)
 
 static void idtcm_display_version_info(struct idtcm *idtcm)
 {
-	u8 major;
-	u8 minor;
-	u8 hotfix;
-	u16 product_id;
-	u8 hw_rev_id;
-	u8 config_select;
+	u8 major = 0;
+	u8 minor = 0;
+	u8 hotfix = 0;
+	u16 product_id = 0;
+	u8 hw_rev_id = 0;
+	u8 config_select = 0;
 	char *fmt = "%d.%d.%d, Id: 0x%04x  HW Rev: %d  OTP Config Select: %d\n";
 
 	idtcm_read_major_release(idtcm, &major);
Signature · 1 line
2.18.1
maintainer
#1 lore raw
Quoted 11 lines · On Sun, Oct 11, 2020 at 01:09:55PM -0700, [email protected] wrote:

From: Tom Rix <[email protected]>

Clang static analysis reports this representative problem

ptp_clockmatrix.c:1852:2: warning: 5th function call argument
is an uninitialized value

        snprintf(idtcm->version, sizeof(idtcm->version), "%u.%u.%u",
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

idtcm_display_version_info() calls several idtcm_read_*'s without checking a return status.

So why not check the return status?

Your patch papers over the issue.

Thanks,
Richard

author
#2 lore raw
Quoted 5 lines · Richard Cochran
Quoted 4 lines · On Sun, Oct 11, 2020 at 01:09:55PM -0700, [email protected] wrote:

From: Tom Rix <[email protected]>

Clang static analysis reports this representative problem

Quoted 8 lines
Quoted 7 lines

ptp_clockmatrix.c:1852:2: warning: 5th function call argument
is an uninitialized value

        snprintf(idtcm->version, sizeof(idtcm->version), "%u.%u.%u",
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

idtcm_display_version_info() calls several idtcm_read_*'s without checking a return status.

So why not check the return status?

calling function is a print information only function that returns void.

I do think not adding error handling is worth it.

I could change the return and then the calls if if you like.

Tom

Quoted 6 lines

Your patch papers over the issue.

Thanks,
Richard

maintainer
#3 lore raw
Quoted 1 line · Tom Rix

calling function is a print information only function that returns void.

That is fine.

Quoted 3 lines

I do think not adding error handling is worth it.

I could change the return and then the calls if if you like.

How about printing the version string when readable, and otherwise print an appropriate informative error message?

Thanks,
Richard