x86-64: fix HYPERVISOR_update_descriptor()
Subsystems (2)
PATCH 1/1 x86-64: fix HYPERVISOR_update_descriptor()
--- arch/x86/include/asm/xen/hypercall.h | 2 ++ 1 file changed, 2 insertions(+)
arch/x86/include/asm/xen/hypercall.h +2 -0
@@ -296,6 +296,8 @@ HYPERVISOR_get_debugreg(int reg)static inline int HYPERVISOR_update_descriptor (u64 ma,u64 desc) { +if (sizeof (u64) ==sizeof (long )) +return _hypercall2 (int, update_descriptor, ma, desc);return _hypercall4 (int, update_descriptor, ma, ma>>32 , desc, desc>>32 ); }
Quoted 10 lines · Jan Beulich
Signed-off-by Jan Beulich <[email protected]>--- arch/x86/include/asm/xen/hypercall.h | 2 ++ 1 file changed, 2 insertions(+)
arch/x86/include/asm/xen/hypercall.h HYPERVISOR_get_debugreg(int reg) · line 301
Quoted 6 lines
arch/x86/include/asm/xen/hypercall.h +2
-0@@ -296,6 +296,8 @@ HYPERVISOR_get_debugreg(int reg)static inline int HYPERVISOR_update_descriptor (u64 ma,u64 desc) { +if (sizeof (u64) ==sizeof (long )) +return _hypercall2 (int, update_descriptor, ma, desc);return _hypercall4 (int, update_descriptor, ma, ma>>32 , desc, desc>>32 );
missing changelog and Impact line.
Ingo
Quoted 9 lines
Quoted 1 line
Quoted 1 line
Ingo Molnar <[email protected]> 12.03.09 11:54 >>>
Quoted 7 lines · Jan Beulich
Signed-off-by Jan Beulich <[email protected]>--- arch/x86/include/asm/xen/hypercall.h | 2 ++ 1 file changed, 2 insertions(+)
Quoted 11 lines
arch/x86/include/asm/xen/hypercall.h HYPERVISOR_get_debugreg(int reg) · line 301
Quoted 9 lines
arch/x86/include/asm/xen/hypercall.h +2
-0@@ -296,6 +296,8 @@ HYPERVISOR_get_debugreg(int reg)static inline int HYPERVISOR_update_descriptor (u64 ma,u64 desc) { +if (sizeof (u64) ==sizeof (long )) +return _hypercall2 (int, update_descriptor, ma, desc);return _hypercall4 (int, update_descriptor, ma, ma>>32 , desc, desc>>32 );missing changelog and Impact line.
I'm confused: What point is there to add a textual description that matches the subject? And where is the need for an impact line documented (clearly neither SubmitChecklist no SubmittingPatches have any occurrence of the word impact), i.e. what are the valid values to chose from?
Jan
Quoted 21 lines · Jan Beulich
Quoted 9 lines
Quoted 1 line
Quoted 1 line
Ingo Molnar <[email protected]> 12.03.09 11:54 >>>
Quoted 7 lines · Jan Beulich
Signed-off-by Jan Beulich <[email protected]>--- arch/x86/include/asm/xen/hypercall.h | 2 ++ 1 file changed, 2 insertions(+)Quoted 11 lines
arch/x86/include/asm/xen/hypercall.h HYPERVISOR_get_debugreg(int reg) · line 301
Quoted 9 lines
arch/x86/include/asm/xen/hypercall.h +2
-0@@ -296,6 +296,8 @@ HYPERVISOR_get_debugreg(int reg)static inline int HYPERVISOR_update_descriptor (u64 ma,u64 desc) { +if (sizeof (u64) ==sizeof (long )) +return _hypercall2 (int, update_descriptor, ma, desc);return _hypercall4 (int, update_descriptor, ma, ma>>32 , desc, desc>>32 );missing changelog and Impact line.
Quoted 2 lines
I'm confused: What point is there to add a textual description that matches the subject? [...]
For example, under what circumstances did you trigger the bug, how widely does it affect people, how did you test it. You are sending patches very close to the 2.6.29 release, and your commit log is non-existent.
Yes, i can figure out what the patch does, but that is not the point.
The point is for you to be forthcoming with such information and trying to be helpful to the maintenance process, by properly describing changes, by describing how you found the bug, how you tested the fix, how significant you find the fix, etc.
I.e. try to emit the information you have about this _already_, and generously so, instead of hiding it and forcing others to recover it. It might be a small work for me to recover it and put it into the changelog, but many of your past patches showed such a pattern and such overhead mounts up quickly.
Quoted 4 lines
[...] And where is the need for an impact line documented (clearly neither SubmitChecklist no SubmittingPatches have any occurrence of the word impact), i.e. what are the valid values to chose from?
See:
http://lkml.org/lkml/2008/10/28/67
Ingo
Quoted 13 lines
Quoted 1 line
Quoted 1 line
Ingo Molnar <[email protected]> 12.03.09 12:35 >>>
Quoted 2 lines · Jan Beulich
I'm confused: What point is there to add a textual description that matches the subject? [...]
For example, under what circumstances did you trigger the bug, how widely does it affect people, how did you test it. You are sending patches very close to the 2.6.29 release, and your commit log is non-existent.
Yes, i can figure out what the patch does, but that is not the point.
Quoted 8 lines
The point is for you to be forthcoming with such information and trying to be helpful to the maintenance process, by properly describing changes, by describing how you found the bug, how you tested the fix, how significant you find the fix, etc.
I.e. try to emit the information you have about this _already_, and generously so, instead of hiding it and forcing others to recover it.
Hmm, I'm really just following what I see from many others. And I have to admit that there are [tiny] patches that really don't need much explanation (and I often find quite the inverse - huge patches that have [almost] no description).
Quoted 3 lines
Quoted 3 lines
It might be a small work for me to recover it and
put it into the changelog, but many of your past patches showed such a pattern and such overhead mounts up quickly.
I'm sorry for that - I simply wasn't aware I'm causing you to do extra work. I usually try to be as verbose with patches as seems necessary to me - after all I have no other way to judge ho much is too little or too much.
Quoted 8 lines
Quoted 4 lines
[...] And where is the need for an impact line documented (clearly neither SubmitChecklist no SubmittingPatches have any occurrence of the word impact), i.e. what are the valid values to chose from?
See:
Thanks. Would certainly be helpful to put into Documentation/ if this is meant to be more than just a personal requirement of yours.
Jan
H. Peter Anvin
reviewerQuoted 6 lines · Jan Beulich
Quoted 3 lines
See:
Thanks. Would certainly be helpful to put into
Documentation/if this is meant to be more than just a personal requirement of yours.
It's something we started with in the x86 tree (it came out of a long discussion Ingo and I had about how to filter patches.) It's not like we can impose this workflow on others, but it has been *way* useful: a patch with a *meaningful* "Impact:" line typically needs less than a third of the human time than one without -- simply because it forces the submitter to think about it.
Similarly "Fix foo" is not a complete description in any way. Why was foo broken in the first place?
Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf.