On Tue, Dec 6, 2022 at 4:14 PM Andrii Nakryiko <andrii.nakryiko@xxxxxxxxx> wrote: > > On Sun, Dec 4, 2022 at 7:01 PM Khem Raj <raj.khem@xxxxxxxxx> wrote: > > > > Clang warns on 32-bit ARM on this comparision > > > > libbpf.c:10497:18: error: result of comparison of constant 4294967296 with expression of type 'size_t' (aka 'unsigned int') is always false [-Werror,-Wtautological-constant-out-of-range-compare] > > if (ref_ctr_off >= (1ULL << PERF_UPROBE_REF_CTR_OFFSET_BITS)) > > ~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > > > Check for platform long int to be larger than 32-bits before enabling > > this check, it false on 32bit anyways. > > > > Cc: Alexei Starovoitov <ast@xxxxxxxxxx> > > Cc: Daniel Borkmann <daniel@xxxxxxxxxxxxx> > > Cc: Song Liu <song@xxxxxxxxxx> > > Cc: Yonghong Song <yhs@xxxxxx> > > Cc: Jiri Olsa <jolsa@xxxxxxxxxx> > > Cc: Paul Walmsley <paul.walmsley@xxxxxxxxxx> > > Cc: Palmer Dabbelt <palmer@xxxxxxxxxxx> > > Cc: Nathan Chancellor <nathan@xxxxxxxxxx> > > Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx> > > > > Signed-off-by: Khem Raj <raj.khem@xxxxxxxxx> > > --- > > tools/lib/bpf/libbpf.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c > > index 91b7106a4a73..65cb70cdc22b 100644 > > --- a/tools/lib/bpf/libbpf.c > > +++ b/tools/lib/bpf/libbpf.c > > @@ -9837,7 +9837,7 @@ static int perf_event_open_probe(bool uprobe, bool retprobe, const char *name, > > char errmsg[STRERR_BUFSIZE]; > > int type, pfd; > > > > - if (ref_ctr_off >= (1ULL << PERF_UPROBE_REF_CTR_OFFSET_BITS)) > > + if (BITS_PER_LONG > 32 && ref_ctr_off >= (1ULL << PERF_UPROBE_REF_CTR_OFFSET_BITS)) > > Where is BITS_PER_LONG defined? Is it a compiler-provided macro? A bit > hesitant about taking dependency on such a macro. Also quick googling > suggests user-space programs should use __BITS_PER_LONG? BITS_PER_LONG is in include/asm-generic/bitsperlong.h, I have sent v2 which implements your suggestion > > Can you try if just casting ref_ctr_off to __u64 would make this > warning go away? > > > return -EINVAL; > > > > memset(&attr, 0, attr_sz); > > -- > > 2.38.1 > >
