On Mon, 3 Feb 2025 at 17:23, Michal Simek <michal.si...@amd.com> wrote: > > > > On 2/3/25 16:13, Ilias Apalodimas wrote: > > Thanks Michal > > > > On Mon, 3 Feb 2025 at 17:03, Michal Simek <michal.si...@amd.com> wrote: > >> > >> LWIP wget.c/mbedtls_hardware_poll() is calling dm_rng_read() but dependency > >> is not recorded anywhere that's why select DM_RNG when LWIP stack i used. > >> > >> Signed-off-by: Michal Simek <michal.si...@amd.com> > >> --- > >> > >> net/Kconfig | 1 + > >> 1 file changed, 1 insertion(+) > >> > >> diff --git a/net/Kconfig b/net/Kconfig > >> index 24508026200d..6c684624fd6f 100644 > >> --- a/net/Kconfig > >> +++ b/net/Kconfig > >> @@ -249,6 +249,7 @@ config WGET > >> select PROT_TCP if NET > >> select PROT_TCP_LWIP if NET_LWIP > >> select PROT_DNS_LWIP if NET_LWIP > >> + select DM_RNG if NET_LWIP > > > > There's a WGET_HTTPS symbol, I think we should add that there > > You maybe right but there is only WGET symbol used. > obj-$(CONFIG_WGET) += wget.o
It only enables HTTPS if CONFIG_WGET_HTTPS is enabled. The whole machinery for adding https support (and calling that rng function) is protected by an #ifdef already. > > It means likely wget.c should be split to HTTP and HTTPS version or add some > ifdefs to that file. > I of course prefer to fix this issue which is visible with microblaze > configuration which one of my patch does. > What is missing is a check in wget_validate_uri() and forbid https:// if that symbol isn't enabled > Thanks, > Michal
