On 08/02/2012 13:14, "Dietmar Hahn" <dietmar.hahn@ts.fujitsu.com> wrote:> Hi, > > while locking at the vpmu stuff I found the following stuff: > > diff -r a55e0f9c7b4e xen/arch/x86/oprofile/nmi_int.c > --- a/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 12:28:47 2012 +0100 > +++ b/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 14:11:06 2012 +0100 > @@ -305,7 +305,7 @@ static int __init p4_init(char ** cpu_ty > } > > #ifndef CONFIG_SMP > - *cpu_type = "i386/p4", XENOPROF_CPU_TYPE_SIZE); > + *cpu_type = "i386/p4"; > model = &op_p4_spec; > return 1; > #else > > > Is CONFIG_SMP always switched on? It seems nobody tested without CONFIG_SMP.Correct. !CONFIG_SMP isn''t configurable or supported. -- Keir> Thanks. > > Dietmar. >
On 08/02/2012 14:05, "Andrew Cooper" <andrew.cooper3@citrix.com> wrote:>>> Is CONFIG_SMP always switched on? It seems nobody tested without CONFIG_SMP. >> Correct. !CONFIG_SMP isn''t configurable or supported. >> >> -- Keir > > A naive test: > > andrewcoop@andrewcoop:/local/xen-unstable.hg/xen$ rgrep "CONFIG_SMP" * | > wc -l > 166 > > Seems there are quite a few bits to remove. If noone has done so, when > I get round to gutting #ifdef __ia64__''s from the arch/x86 tree I shall > do the same with CONFIG_SMP.Yeah we never actually went and deleted the !CONFIG_SMP bits we inherited from Linux. You can feel free to clean it up afaic. -- Keir
Hi, while locking at the vpmu stuff I found the following stuff: diff -r a55e0f9c7b4e xen/arch/x86/oprofile/nmi_int.c --- a/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 12:28:47 2012 +0100 +++ b/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 14:11:06 2012 +0100 @@ -305,7 +305,7 @@ static int __init p4_init(char ** cpu_ty } #ifndef CONFIG_SMP - *cpu_type = "i386/p4", XENOPROF_CPU_TYPE_SIZE); + *cpu_type = "i386/p4"; model = &op_p4_spec; return 1; #else Is CONFIG_SMP always switched on? It seems nobody tested without CONFIG_SMP. Thanks. Dietmar. -- Company details: ts.fujitsu.com/imprint.html
On 08/02/12 05:57, Keir Fraser wrote:> On 08/02/2012 13:14, "Dietmar Hahn" <dietmar.hahn@ts.fujitsu.com> wrote: > >> Hi, >> >> while locking at the vpmu stuff I found the following stuff: >> >> diff -r a55e0f9c7b4e xen/arch/x86/oprofile/nmi_int.c >> --- a/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 12:28:47 2012 +0100 >> +++ b/xen/arch/x86/oprofile/nmi_int.c Wed Feb 08 14:11:06 2012 +0100 >> @@ -305,7 +305,7 @@ static int __init p4_init(char ** cpu_ty >> } >> >> #ifndef CONFIG_SMP >> - *cpu_type = "i386/p4", XENOPROF_CPU_TYPE_SIZE); >> + *cpu_type = "i386/p4"; >> model = &op_p4_spec; >> return 1; >> #else >> >> >> Is CONFIG_SMP always switched on? It seems nobody tested without CONFIG_SMP. > Correct. !CONFIG_SMP isn''t configurable or supported. > > -- KeirA naive test: andrewcoop@andrewcoop:/local/xen-unstable.hg/xen$ rgrep "CONFIG_SMP" * | wc -l 166 Seems there are quite a few bits to remove. If noone has done so, when I get round to gutting #ifdef __ia64__''s from the arch/x86 tree I shall do the same with CONFIG_SMP. ~Andrew>> Thanks. >> >> Dietmar. >> > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > lists.xensource.com/xen-devel-- Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer T: +44 (0)1223 225 900, citrix.com