I think the original value was incorrect. glibc always assumed that the code for the vsyscall gettimeofday has the address offset 0. You can look at the commit log of the file sysdeps/unix/sysv/linux/x86_64/gettimeofday.c at the following url:
http://sourceware.org/git/?p=glibc.git;a=history;f=sysdeps/unix/sysv/linux/x86_64/gettimeofday.c;h=556e0d5dcbce857f6aaf95c755b705136ad595dd;hb=HEAD -- Nilay On Mon, March 25, 2013 10:52 pm, Steve Reinhardt wrote: > Does this patch break older executables? It seems to me that if it used > to > work with the old value, there may still be some programs around that > expect the older value... > > > On Mon, Mar 25, 2013 at 5:46 PM, Nilay Vaish <[email protected]> wrote: > >> On Mon, 25 Mar 2013, Michael Levenhagen wrote: >> >> Has it been determined that this change is correct? If yes is it going >> to >>> be committed? >>> >>> Mike >>> >>> diff -r b6a47b43f4df src/arch/x86/process.cc >>> --- a/src/arch/x86/process.cc Mon Mar 18 11:24:56 2013 +0100 >>> +++ b/src/arch/x86/process.cc Mon Mar 25 09:46:20 2013 -0600 >>> @@ -98,7 +98,7 @@ >>> vsyscallPage.base = 0xffffffffff600000ULL; >>> vsyscallPage.size = VMPageSize; >>> vsyscallPage.vtimeOffset = 0x400; >>> - vsyscallPage.**vgettimeofdayOffset = 0x410; >>> + vsyscallPage.**vgettimeofdayOffset = 0x0; >>> >>> _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
