Re: [PATCH v2 1/2] soc: qcom: cmd-db: replace strncpy() with strscpy_pad()

[Krzysztof Kozlowski]

On 28/06/2022 01:06, Doug Anderson wrote:
> Hi,
> 
> On Thu, May 19, 2022 at 12:33 AM Krzysztof Kozlowski
> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>>
>> The use of strncpy() is considered deprecated for NUL-terminated
>> strings[1].  Replace strncpy() with strscpy_pad(), to keep existing
>> pad-behavior of strncpy.  This fixes W=1 warning:
>>
>>   drivers/soc/qcom/cmd-db.c: In function ‘cmd_db_get_header.part.0’:
>>   drivers/soc/qcom/cmd-db.c:151:9: warning: ‘strncpy’ specified bound 8 equals destination size [-Wstringop-truncation]
>>     151 |         strncpy(query, id, sizeof(query));
>>
>> [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings
>>
>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>> Reviewed-by: Stephen Boyd <sboyd@xxxxxxxxxx>
>>
>> ---
>>
>> Changes since v1:
>> 1. Split series per subsystem.
>> 2. Add tag.
>> ---
>>  drivers/soc/qcom/cmd-db.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/soc/qcom/cmd-db.c b/drivers/soc/qcom/cmd-db.c
>> index dd872017f345..c5137c25d819 100644
>> --- a/drivers/soc/qcom/cmd-db.c
>> +++ b/drivers/soc/qcom/cmd-db.c
>> @@ -148,7 +148,7 @@ static int cmd_db_get_header(const char *id, const struct entry_header **eh,
>>                 return ret;
>>
>>         /* Pad out query string to same length as in DB */
>> -       strncpy(query, id, sizeof(query));
>> +       strscpy_pad(query, id, sizeof(query));
> 
> Sorry to report that this breaks booting on
> sc7280-herobrine-herobrine-r1. 

Sorry to break booting and thanks for the report.

> I believe that the function was
> actually _relying_ on the "unsafe" behavior of strncpy(). Specifically
> I think:
> * The source string (id) was a '\0' terminated string.
> * The destination (query) was a fixed 8-byte string and doesn't need
> to be '\0' terminated.
> 
> So before your change we could actually support 8-byte strings. Now we
> can't because you'll only copy 7 bytes to the destination to leave
> room for the '\0' termination...
> 
> Looking at printouts, I see, for instance at least one ID that looks
> like "lnbclka2".

Ah, crap... I did not expect that.

> 
> Given the requirements of this function(), the old strncpy() is
> actually _exactly_ what we want. Is there any way to disable the
> warning? If not, I guess we could make "query" be 9 bytes bit, or
> "sizeof(ent->id) + 1" bytes big... Happy to post a patch, but it's
> basically a bikeshed for how you want it fixed (there are dozens of
> ways) and I'd rather you just tell me instead of me guessing.

If the source was not a string, I would propose memcpy(). Since it is a
string, I would prefer to have safe copy, so increased size. However I
see you sent now patches, so let me actually respond there.

Thanks for the analysis, much appreciated!


Best regards,
Krzysztof