Re: [systemd-devel] [PATCH] execute: Fix seccomp support on x32

2012-08-04 Thread microcai
2012/8/4 Lennart Poettering lenn...@poettering.net:
 On Tue, 24.07.12 22:45, Jonathan Callen (a...@gentoo.org) wrote:

 In the x32 ABI, syscall numbers start at 0x4000.  Mask that bit on
 x32 for lookups in the syscall_names array and syscall_filter and ensure
 that syscall.h is parsed correctly.

 Hmpf, can't say I find this patch particularly beautiful?

 Can we solve this differently? For example, I'd be open to replace the
 direct seccomp code in systemd by some code based on libseccomp (now
 that libseccomp actually fixed its static global state issues). That way
 we should get this portability for free?

looks like some  source-based distro will blame you again :)



 Lennart

 --
 Lennart Poettering - Red Hat, Inc.
 ___
 systemd-devel mailing list
 systemd-devel@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/systemd-devel
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/systemd-devel


Re: [systemd-devel] [PATCH] execute: Fix seccomp support on x32

2012-08-04 Thread Bryan Kadzban
microcai wrote:
 2012/8/4 Lennart Poettering lenn...@poettering.net:
 On Tue, 24.07.12 22:45, Jonathan Callen (a...@gentoo.org) wrote:
 
 In the x32 ABI, syscall numbers start at 0x4000.  Mask that
 bit on x32 for lookups in the syscall_names array and
 syscall_filter and ensure that syscall.h is parsed correctly.
 Hmpf, can't say I find this patch particularly beautiful?
 
 Can we solve this differently? For example, I'd be open to replace
 the direct seccomp code in systemd by some code based on libseccomp
 (now that libseccomp actually fixed its static global state
 issues). That way we should get this portability for free?
 
 looks like some  source-based distro will blame you again :)

libseccomp uses pkg-config, so it's much less of a problem IMO;
SECCOMP_CFLAGS=  SECCOMP_LIBS=  will allow configure to at least
finish properly if the library is not present.  (Whether the system
builds at all in this case is of course an issue that anyone sending
those params to ./configure will have to deal with.)
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/systemd-devel