Messages in this thread | | | From | Nick Desaulniers <> | Date | Wed, 19 Sep 2018 10:55:08 -0700 | Subject | Re: [ojeda-linux:compiler-attributes-v4 7/13] drivers/firewire/core-card.c:479 bm_work() warn: always true condition '(root_node->max_hops < (16 / 1 + (0))) => (0-15 < 16)' |
| |
On Tue, Sep 18, 2018 at 6:11 PM kbuild test robot <lkp@intel.com> wrote: > > tree: git://github.com/ojeda/linux compiler-attributes-v4 > head: 7c70bc9e399c67f5d1a84b845d2656f8808e0032 > commit: 431781b83c5e7ae3719498bf1e52ca99bcf6d20b [7/13] Compiler Attributes: remove unneeded sparse (__CHECKER__) tests > > smatch warnings: > drivers/firewire/core-card.c:479 bm_work() warn: always true condition '(root_node->max_hops < (16 / 1 + (0))) => (0-15 < 16)' > drivers/ide/ide-cd_verbose.c:271 ide_cd_log_error() warn: always true condition '(sense->sense_key < (64 / 4 + (0))) => (0-15 < 16)'
If I understand this correctly, it seems that the condition:
479 root_node->max_hops < ARRAY_SIZE(gap_count_table))
is always true. root_node->max_hops is declared as:
187 u8 max_hops:4; /* Max hops in this sub tree */
in drivers/firewire/core.h, while gap_count_table is declared as:
289 static const char gap_count_table[] = { 290 63, 5, 7, 8, 10, 13, 16, 18, 21, 24, 26, 29, 32, 35, 37, 40 291 };
Since max_hops is a 4b bit vector, its maximal unsigned representable value is (2^4) - 1 == 15, and 15 is always less that 16, so indeed this condition is always true.
One issue is that should gap_count_table grow, then this condition may become always false. Removing the condition might break the then clause should gap_count_table grow.
Maybe this can be converted to a BUILD_BUG_ON? -- Thanks, ~Nick Desaulniers
| |