Messages in this thread | | | From | Christian Borntraeger <> | Subject | Re: [patch 27/85] [PATCH 07/44] [CVE-2009-0029] System call wrapper infrastructure | Date | Fri, 16 Jan 2009 09:44:59 +0100 |
| |
Am Donnerstag 15 Januar 2009 schrieb Jiri Slaby: > On 01/15/2009 10:59 PM, Greg KH wrote: > > + > > +#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS > > + > > +#define SYSCALL_DEFINE(name) static inline long SYSC_##name > > +#define SYSCALL_DEFINEx(x, name, ...) \ > > + asmlinkage long sys_##name(__SC_DECL##x(__VA_ARGS__)); \ > > + static inline long SYSC_##name(__SC_DECL##x(__VA_ARGS__)); \ > > Hmm, we don't support at least gcc 3.4.6 anymore :/. It seems to be the same as in:
I see two possible quick fixes: -funit-at-a-time will allow these inline constructs (tested with 3.4.6 on i386 with enabled syscall wrappers. x86_64 already sets this flag)
- dont use attribute __always_inline__ for these inlines but the standard inline. The question is, how to convince cpp to not replace inline with always_inline
| |