Hi guys,

I am now hitting a similar problem on armel.

I am cross-compiling "dnsmasq"

Glibc strace:

---snip---
socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP) = 5
fcntl64(5, F_GETFL)                     = 0x2 (flags O_RDWR)
fcntl64(5, F_SETFL, O_RDWR|O_NONBLOCK)  = 0
setsockopt(5, SOL_IP, IP_MTU_DISCOVER, [0], 4) = 0
setsockopt(5, SOL_IP, IP_PKTINFO, [1], 4) = 0
setsockopt(5, SOL_SOCKET, SO_BROADCAST, [1], 4) = 0
setsockopt(5, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0


Musl strace:

socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP) = 5
fcntl64(5, F_GETFL)                     = 0x2 (flags O_RDWR)
fcntl64(5, F_SETFL, O_RDWR|O_NONBLOCK|O_LARGEFILE) = 0
setsockopt(5, SOL_IP, IP_MTU_DISCOVER, [0], 4) = 0
setsockopt(5, SOL_IP, IP_PKTINFO, [1], 4) = 0
setsockopt(5, SOL_SOCKET, SO_BROADCAST, [1], 4) = 0
setsockopt(5, SOL_SOCKET, *0xf /* SO_??? */*, [1], 4) = -1 ENOPROTOOPT
(Protocol not available)

Regards
Paul

On Mon, Nov 5, 2012 at 7:31 PM, Rich Felker <dalias@aerifal.cx> wrote:

> On Mon, Nov 05, 2012 at 07:27:44PM +0200, Paul Schutte wrote:
> > Thanks for the reply.
> >
> > I suspected something like you mention and therefore I made sure that I
> > installed the correct kernel headers before I compiled musl or any of the
> > other software.
>
> Kernel headers are completely irrelevant; they're not used anywhere.
> nsz's analysis is right. I have a patch I'm about to commit. In case
> you'd like to try it directly without using git, I'm attaching it
> here.
>
> There may be more fixes still needed. MIPS has A LOT of gratuitous
> differences in constant definitions and I thought I'd caught all of
> them before, but apparently some slipped through. The fcntl part of
> the patch has been tested and seems correct. The socket parts should
> be right, but they're untested; I'm working on a test for them now.
>
> Rich
>