[U-Boot] [drivers/mtd/ubi/eba.c:1275]: (error) Uninitialized variable: aeb

[Heiko Schocher]

Hello Wolfgang,

Am 06.11.2014 16:10, schrieb Wolfgang Denk:
> Dear Heiko,
>
> In message<545B81D9.6070602 at denx.de>  you wrote:
>>>
>>> [drivers/mtd/ubi/eba.c:1275]: (error) Uninitialized variable: aeb
>>>
>>> can you please have a look?  Thanks!
>>
>> I see in drivers/mtd/ubi/eba.c:
>>
>> 1267:                ubi_rb_for_each_entry(rb, aeb,&av->root, u.rb)
>> 1268:                        scan_eba[i][aeb->lnum] = aeb->pnum;
>> 1269:
>> 1270:                av = ubi_find_av(ai_fastmap, idx2vol_id(ubi, i));
>> 1271:                if (!av)
>> 1272:                       continue;
>> 1273:
>> 1274:                ubi_rb_for_each_entry(rb, aeb,&av->root, u.rb)
>> 1275:                        fm_eba[i][aeb->lnum] = aeb->pnum;
>>
>> Why does cppcheck only report line 1275 not also line 1268 and also
>> 1351? I could not currently see, why this pops up for line 1274 ...
>
> we should also look at the ubi_rb_for_each_entry() macro:
>
>   912 #define ubi_rb_for_each_entry(rb, pos, root, member)                         \
>   913         for (rb = rb_first(root),                                            \
>   914              pos = (rb ? container_of(rb, typeof(*pos), member) : NULL);     \
>   915              rb;                                                             \
>   916              rb = rb_next(rb),                                               \
>   917              pos = (rb ? container_of(rb, typeof(*pos), member) : NULL))
>
> We can see there, that this will always set "pos", i. e. the 2nd
> argument, to a well-defined value.
>
> Looks like a false positive to me, then.

Yes.

bye,
Heiko
-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany