Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 26 Jun 2018 16:05:25 -0400
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: [PATCH] sys/socket.h: fix SO_PEERSEC value on MIPS

On Tue, Jun 26, 2018 at 08:21:38PM +0200, Matthias Schiffer wrote:
> On 06/25/2018 07:52 PM, Matthias Schiffer wrote:
> > On 06/24/2018 05:30 PM, Rich Felker wrote:
> >> On Sun, Jun 24, 2018 at 05:18:39PM +0200, Matthias Schiffer wrote:
> >>> Differing from all other archs supported by musl, MIPS defines SO_PEERSEC
> >>> to 30 instead of 31.
> >>>
> >>> Reported-by: Andrey Jr. Mlenikov <temnota.am@...il.com>
> >>> ---
> >>>  arch/mips/bits/socket.h    | 2 ++
> >>>  arch/mips64/bits/socket.h  | 2 ++
> >>>  arch/mipsn32/bits/socket.h | 2 ++
> >>>  include/sys/socket.h       | 3 +++
> >>>  4 files changed, 9 insertions(+)
> >>>
> >>> diff --git a/arch/mips/bits/socket.h b/arch/mips/bits/socket.h
> >>> index 191ebdb5a418..39c42b822ac3 100644
> >>> --- a/arch/mips/bits/socket.h
> >>> +++ b/arch/mips/bits/socket.h
> >>> @@ -48,5 +48,7 @@ struct cmsghdr {
> >>>  #define SO_SNDBUFFORCE  31
> >>>  #define SO_RCVBUFFORCE  33
> >>>  
> >>> +#define SO_PEERSEC      30
> >>> +
> >>>  #define SOCK_NONBLOCK     0200
> >>>  #define SOCK_CLOEXEC  02000000
> >>> diff --git a/arch/mips64/bits/socket.h b/arch/mips64/bits/socket.h
> >>> index cf801797beb0..39b764df4c62 100644
> >>> --- a/arch/mips64/bits/socket.h
> >>> +++ b/arch/mips64/bits/socket.h
> >>> @@ -64,5 +64,7 @@ struct cmsghdr {
> >>>  #define SO_SNDBUFFORCE  31
> >>>  #define SO_RCVBUFFORCE  33
> >>>  
> >>> +#define SO_PEERSEC      30
> >>> +
> >>>  #define SOCK_NONBLOCK     0200
> >>>  #define SOCK_CLOEXEC  02000000
> >>> diff --git a/arch/mipsn32/bits/socket.h b/arch/mipsn32/bits/socket.h
> >>> index 191ebdb5a418..39c42b822ac3 100644
> >>> --- a/arch/mipsn32/bits/socket.h
> >>> +++ b/arch/mipsn32/bits/socket.h
> >>> @@ -48,5 +48,7 @@ struct cmsghdr {
> >>>  #define SO_SNDBUFFORCE  31
> >>>  #define SO_RCVBUFFORCE  33
> >>>  
> >>> +#define SO_PEERSEC      30
> >>> +
> >>>  #define SOCK_NONBLOCK     0200
> >>>  #define SOCK_CLOEXEC  02000000
> >>> diff --git a/include/sys/socket.h b/include/sys/socket.h
> >>> index 507da5cc2a40..6b6cf2fc3988 100644
> >>> --- a/include/sys/socket.h
> >>> +++ b/include/sys/socket.h
> >>> @@ -203,7 +203,10 @@ struct linger {
> >>>  #define SO_TIMESTAMP            29
> >>>  #define SCM_TIMESTAMP           SO_TIMESTAMP
> >>>  
> >>> +#ifndef SO_PEERSEC
> >>>  #define SO_PEERSEC              31
> >>> +#endif
> >>> +
> >>>  #define SO_PASSSEC              34
> >>>  #define SO_TIMESTAMPNS          35
> >>>  #define SCM_TIMESTAMPNS         SO_TIMESTAMPNS
> >>> -- 
> >>> 2.18.0
> >>
> >> Should we perhaps do this under the SO_DEBUG conditional like all the
> >> other SO_* that vary by arch? Doing so would require adding
> >> definitions to the other weird archs (ibm ones) but might be cleaner
> >> in the top-level header -- I'm not sure.
> >>
> >> Rich
> >>
> > 
> > 
> > That would actually make sense. I did not research what SO_PEERSEC even is
> > before I sent the patch; only know I noticed that both the constants
> > defined under SO_DEBUG and the others below the #endif are SOL_SOCKET, so
> > it would be more consistent to move it under SO_DEBUG as well.
> 
> Should I send an updated patch, or are you going to fix it up yourself?

I'm doing it now. Thanks. I think we're going to find that there are
more SO_* that might vary by arch, especially if we ever add sparc.

What I'd kind of like to eventually do is get rid of the definitions
in sys/socket.h itself and always have them all in bits, with most
archs sharing the arch/generic version of the bits header. However
that doesn't work with the current bits/socket.h since the generic
definitions of msghdr and cmsghdr don't work on 64-bit. Perhaps we
should introduce a new bits/sockopt.h (but this imposes an extra file
open/parse for every #include <sys/socket.h>) or find some way to
avoid the gratuitous difference (maybe alltypes.h could somehow
provide padding definitions for a generic definition of [c]msghdr to
use).

Anyway I won't delay the immediate bugfix with what-ifs about future
direction.

Rich

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ