I''ve run across a few seemingly needless changes to pre-existing Linux
files, that thus only bloat the diff when
generated against a vanilla tree. Examples below. Could we agree on removing
such deltas? Thanks, Jan
--- linux-2.6.16.13/include/asm-i386/spinlock.h 2006-08-02 15:29:34.000000000
+0200
+++ 2006-08-02/include/asm-i386/spinlock.h 2006-02-01 18:28:39.000000000 +0100
@@ -23,7 +23,7 @@
(*(volatile signed char *)(&(x)->slock) <= 0)
#define __raw_spin_lock_string \
- "\n1:\t" \
+ "\n1:\n" \
LOCK \
"decb %0\n\t" \
"jns 3f\n" \
@@ -35,7 +35,7 @@
"3:\n\t"
#define __raw_spin_lock_string_flags \
- "\n1:\t" \
+ "\n1:\n" \
LOCK \
"decb %0\n\t" \
"jns 4f\n\t" \
@@ -92,7 +92,7 @@ static inline int __raw_spin_trylock(raw
:"0" (0) : "memory");
#else
__asm__ __volatile__(
- "xchgb %b0,%1\n"
+ "xchgb %b0,%1"
:"=q" (oldval), "=m" (lock->slock)
:"0" (0) : "memory");
#endif
--- linux-2.6.16.13/include/asm-i386/system.h 2006-08-02 15:29:34.000000000
+0200
+++ 2006-08-02/include/asm-i386/system.h 2006-02-16 08:27:24.000000000 +0100
@@ -523,8 +523,8 @@ __asm__ __volatile__("6667:\nnop\nnop\nn
: \
: "i" (X86_FEATURE_XMM2) \
: "memory")
-#define smp_rmb() smp_alt_mb("lfence")
#define smp_mb() smp_alt_mb("mfence")
+#define smp_rmb() smp_alt_mb("lfence")
#define set_mb(var, value) do { \
unsigned long __set_mb_temp; \
__asm__ __volatile__("6667:movl %1, %0\n6668:\n" \
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
On 2 Aug 2006, at 16:31, Jan Beulich wrote:> I''ve run across a few seemingly needless changes to pre-existing Linux > files, that thus only bloat the diff when > generated against a vanilla tree. Examples below. Could we agree on > removing such deltas? Thanks, JanIf they are all as obviously trivial as the examples you give, we''d be happy to take patches to clean them up. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Here we go.>>> Keir Fraser <Keir.Fraser@cl.cam.ac.uk> 02.08.06 19:22 >>>On 2 Aug 2006, at 16:31, Jan Beulich wrote:> I''ve run across a few seemingly needless changes to pre-existing Linux > files, that thus only bloat the diff when > generated against a vanilla tree. Examples below. Could we agree on > removing such deltas? Thanks, JanIf they are all as obviously trivial as the examples you give, we''d be happy to take patches to clean them up. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 3/8/06 8:43 am, "Jan Beulich" <jbeulich@novell.com> wrote:> Here we go. > >>>> Keir Fraser <Keir.Fraser@cl.cam.ac.uk> 02.08.06 19:22 >>> > > On 2 Aug 2006, at 16:31, Jan Beulich wrote: > >> I''ve run across a few seemingly needless changes to pre-existing Linux >> files, that thus only bloat the diff when >> generated against a vanilla tree. Examples below. Could we agree on >> removing such deltas? Thanks, Jan > > If they are all as obviously trivial as the examples you give, we''d be > happy to take patches to clean them up.What tree are you diffing against? The two \n->\t changes in spinlock.h seem to make sense, but the additional \n in __raw_spin_trylock isn''t present in my vanilla 2.6.16.13 sources. And the line you change in system.h has no direct equivalent in the vanilla sources since they don''t have SMP alternatives. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
>What tree are you diffing against? The two \n->\t changes in spinlock.h seem >to make sense, but the additional \n in __raw_spin_trylock isn''t present in >my vanilla 2.6.16.13 sources. And the line you change in system.h has no >direct equivalent in the vanilla sources since they don''t have SMP >alternatives.I diff-ed a 2.6.16.13 tree with patches/linux-2.6.16.13/* applied against a fully overlaid one. If some of the differences don''t apply, then some cleanup of the pre- patches might be needed, too. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 7/8/06 4:48 pm, "Jan Beulich" <jbeulich@novell.com> wrote:>> What tree are you diffing against? The two \n->\t changes in spinlock.h seem >> to make sense, but the additional \n in __raw_spin_trylock isn''t present in >> my vanilla 2.6.16.13 sources. And the line you change in system.h has no >> direct equivalent in the vanilla sources since they don''t have SMP >> alternatives. > > I diff-ed a 2.6.16.13 tree with patches/linux-2.6.16.13/* applied against a > fully > overlaid one. If some of the differences don''t apply, then some cleanup of the > pre- > patches might be needed, too.There''ll be a sync with the smp alternatives support in upstream when we move from 2.6.16 in xen-unstable after 3.0.3, so this probably isn''t worth the effort right now. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel