diff mbox

[for-next,2/7] x86: move x86_64/compat/entry.S to pv/compat

Message ID 20170406171434.15484-3-wei.liu2@citrix.com (mailing list archive)
State New, archived
Headers show

Commit Message

Wei Liu April 6, 2017, 5:14 p.m. UTC
That file contains 32 bit PV guest entry points.

No functional change.

Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
 xen/arch/x86/pv/Makefile                    | 2 ++
 xen/arch/x86/{x86_64 => pv}/compat/Makefile | 0
 xen/arch/x86/{x86_64 => pv}/compat/entry.S  | 0
 xen/arch/x86/x86_64/Makefile                | 2 --
 4 files changed, 2 insertions(+), 2 deletions(-)
 rename xen/arch/x86/{x86_64 => pv}/compat/Makefile (100%)
 rename xen/arch/x86/{x86_64 => pv}/compat/entry.S (100%)

Comments

Jan Beulich April 21, 2017, 9:16 a.m. UTC | #1
>>> On 06.04.17 at 19:14, <wei.liu2@citrix.com> wrote:
> That file contains 32 bit PV guest entry points.
> 
> No functional change.
> 
> Signed-off-by: Wei Liu <wei.liu2@citrix.com>
> ---
>  xen/arch/x86/pv/Makefile                    | 2 ++
>  xen/arch/x86/{x86_64 => pv}/compat/Makefile | 0
>  xen/arch/x86/{x86_64 => pv}/compat/entry.S  | 0

While this one indeed looks fine to be moved as a whole, I'm not
convinced we want to retain this compat/ subdirectory level, as
we don't really have CONFIG_COMPAT anymore anyway (there
are only very few leftovers in x86-specific code, and it looks like
they've been wrongly re-introduced not so long ago). I.e. I'd
rather see this being merged into pv/entry.S.

Jan
Andrew Cooper April 21, 2017, 9:50 a.m. UTC | #2
On 21/04/17 10:16, Jan Beulich wrote:
>>>> On 06.04.17 at 19:14, <wei.liu2@citrix.com> wrote:
>> That file contains 32 bit PV guest entry points.
>>
>> No functional change.
>>
>> Signed-off-by: Wei Liu <wei.liu2@citrix.com>
>> ---
>>  xen/arch/x86/pv/Makefile                    | 2 ++
>>  xen/arch/x86/{x86_64 => pv}/compat/Makefile | 0
>>  xen/arch/x86/{x86_64 => pv}/compat/entry.S  | 0
> While this one indeed looks fine to be moved as a whole, I'm not
> convinced we want to retain this compat/ subdirectory level, as
> we don't really have CONFIG_COMPAT anymore anyway (there
> are only very few leftovers in x86-specific code, and it looks like
> they've been wrongly re-introduced not so long ago). I.e. I'd
> rather see this being merged into pv/entry.S.

We most certainly do have CONFIG_COMPAT (and it is used), but I also
want to get rid of the compat/ directories.

+1 for pv/entry.S (or alternatively to follow the Linux model and using
entry_64.S and entry_64_compat.S.  I guess it depends how much code we
end up having which is guest-type specific, once I have finished moving
all I reasonably can up into C.)

~Andrew
Jan Beulich April 21, 2017, 9:56 a.m. UTC | #3
>>> On 21.04.17 at 11:50, <andrew.cooper3@citrix.com> wrote:
> On 21/04/17 10:16, Jan Beulich wrote:
>>>>> On 06.04.17 at 19:14, <wei.liu2@citrix.com> wrote:
>>> That file contains 32 bit PV guest entry points.
>>>
>>> No functional change.
>>>
>>> Signed-off-by: Wei Liu <wei.liu2@citrix.com>
>>> ---
>>>  xen/arch/x86/pv/Makefile                    | 2 ++
>>>  xen/arch/x86/{x86_64 => pv}/compat/Makefile | 0
>>>  xen/arch/x86/{x86_64 => pv}/compat/entry.S  | 0
>> While this one indeed looks fine to be moved as a whole, I'm not
>> convinced we want to retain this compat/ subdirectory level, as
>> we don't really have CONFIG_COMPAT anymore anyway (there
>> are only very few leftovers in x86-specific code, and it looks like
>> they've been wrongly re-introduced not so long ago). I.e. I'd
>> rather see this being merged into pv/entry.S.
> 
> We most certainly do have CONFIG_COMPAT (and it is used),

Outside of x86-specific code, yes. In x86-specific code there's just
one pair of instances, in the PV-hypercall handling code. I think this
should be cleaned up.

Jan
diff mbox

Patch

diff --git a/xen/arch/x86/pv/Makefile b/xen/arch/x86/pv/Makefile
index 08b6a3af64..d8fc13f6fe 100644
--- a/xen/arch/x86/pv/Makefile
+++ b/xen/arch/x86/pv/Makefile
@@ -1,3 +1,5 @@ 
+subdir-y += compat
+
 obj-y += hypercall.o
 obj-bin-y += dom0_build.init.o
 obj-bin-y += entry.o
diff --git a/xen/arch/x86/x86_64/compat/Makefile b/xen/arch/x86/pv/compat/Makefile
similarity index 100%
rename from xen/arch/x86/x86_64/compat/Makefile
rename to xen/arch/x86/pv/compat/Makefile
diff --git a/xen/arch/x86/x86_64/compat/entry.S b/xen/arch/x86/pv/compat/entry.S
similarity index 100%
rename from xen/arch/x86/x86_64/compat/entry.S
rename to xen/arch/x86/pv/compat/entry.S
diff --git a/xen/arch/x86/x86_64/Makefile b/xen/arch/x86/x86_64/Makefile
index 1f0a4f7bb2..53e2f95e77 100644
--- a/xen/arch/x86/x86_64/Makefile
+++ b/xen/arch/x86/x86_64/Makefile
@@ -1,5 +1,3 @@ 
-subdir-y += compat
-
 obj-bin-y += gpr_switch.o
 obj-y += traps.o
 obj-$(CONFIG_KEXEC) += machine_kexec.o