[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 11 Dec 2019 16:11:40 -0800
From: Andrii Nakryiko <andrii.nakryiko@...il.com>
To: Martin Lau <kafai@...com>
Cc: Andrii Nakryiko <andriin@...com>,
"bpf@...r.kernel.org" <bpf@...r.kernel.org>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Alexei Starovoitov <ast@...com>,
"daniel@...earbox.net" <daniel@...earbox.net>,
Kernel Team <Kernel-team@...com>
Subject: Re: [PATCH bpf-next] libbpf: fix printf compilation warnings on
ppc64le arch
On Wed, Dec 11, 2019 at 4:09 PM Martin Lau <kafai@...com> wrote:
>
> On Wed, Dec 11, 2019 at 11:26:34AM -0800, Andrii Nakryiko wrote:
> > On ppc64le __u64 and __s64 are defined as long int and unsigned long int,
> > respectively. This causes compiler to emit warning when %lld/%llu are used to
> > printf 64-bit numbers. Fix this by casting directly to unsigned long long
> > (through shorter typedef). In few cases casting error code to int explicitly
> > is cleaner, so that's what's done instead.
> >
> > Fixes: 1f8e2bcb2cd5 ("libbpf: Refactor relocation handling")
> > Fixes: abd29c931459 ("libbpf: allow specifying map definitions using BTF")
> > Signed-off-by: Andrii Nakryiko <andriin@...com>
> > ---
> > tools/lib/bpf/libbpf.c | 34 ++++++++++++++++++----------------
> > 1 file changed, 18 insertions(+), 16 deletions(-)
> >
> > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> > index 3f09772192f1..5ee54f9355a4 100644
> > --- a/tools/lib/bpf/libbpf.c
> > +++ b/tools/lib/bpf/libbpf.c
> > @@ -128,6 +128,8 @@ void libbpf_print(enum libbpf_print_level level, const char *format, ...)
> > # define LIBBPF_ELF_C_READ_MMAP ELF_C_READ
> > #endif
> >
> > +typedef unsigned long long __pu64;
> > +
> > static inline __u64 ptr_to_u64(const void *ptr)
> > {
> > return (__u64) (unsigned long) ptr;
> > @@ -1242,15 +1244,15 @@ static int bpf_object__init_user_btf_map(struct bpf_object *obj,
> > }
> > sz = btf__resolve_size(obj->btf, t->type);
> > if (sz < 0) {
> > - pr_warn("map '%s': can't determine key size for type [%u]: %lld.\n",
> > - map_name, t->type, sz);
> > + pr_warn("map '%s': can't determine key size for type [%u]: %d.\n",
> > + map_name, t->type, (int)sz);
> > return sz;
> > }
> > - pr_debug("map '%s': found key [%u], sz = %lld.\n",
> > - map_name, t->type, sz);
> > + pr_debug("map '%s': found key [%u], sz = %d.\n",
> > + map_name, t->type, (int)sz);
> > if (map->def.key_size && map->def.key_size != sz) {
> > - pr_warn("map '%s': conflicting key size %u != %lld.\n",
> > - map_name, map->def.key_size, sz);
> > + pr_warn("map '%s': conflicting key size %u != %d.\n",
> > + map_name, map->def.key_size, (int)sz);
> > return -EINVAL;
> > }
> > map->def.key_size = sz;
> > @@ -1285,15 +1287,15 @@ static int bpf_object__init_user_btf_map(struct bpf_object *obj,
> > }
> > sz = btf__resolve_size(obj->btf, t->type);
> > if (sz < 0) {
> > - pr_warn("map '%s': can't determine value size for type [%u]: %lld.\n",
> > - map_name, t->type, sz);
> > + pr_warn("map '%s': can't determine value size for type [%u]: %d.\n",
> > + map_name, t->type, (int)sz);
> > return sz;
> > }
> > - pr_debug("map '%s': found value [%u], sz = %lld.\n",
> > - map_name, t->type, sz);
> > + pr_debug("map '%s': found value [%u], sz = %d.\n",
> > + map_name, t->type, (int)sz);
> > if (map->def.value_size && map->def.value_size != sz) {
> > - pr_warn("map '%s': conflicting value size %u != %lld.\n",
> > - map_name, map->def.value_size, sz);
> > + pr_warn("map '%s': conflicting value size %u != %d.\n",
> > + map_name, map->def.value_size, (int)sz);
> It is not an error case (i.e. not sz < 0) here.
> Same for the above pr_debug().
You are right, not sure if it matters in practice, though. Highly
unlikely values will be bigger than 2GB, but even if they, they still
fit in 4 bytes, we'll just report them as negative values. I can do
similar __ps64 conversion, as for __pu64, though, it we are afraid
it's going to be a problem.
Powered by blists - more mailing lists