[PATCH-for-5.1 v2 0/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

[PATCH-for-5.1 v2 0/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Philippe Mathieu-Daudé-3
Since v1: Fix build error reported by Peter:
https://www.mail-archive.com/qemu-devel@.../msg727251.html
Now following KVM implementation closely.

Philippe Mathieu-Daudé (1):
  accel/xen: Fix xen_enabled() behavior on target-agnostic objects

 include/sysemu/xen.h   | 18 ++++++++++++++----
 accel/stubs/xen-stub.c |  2 ++
 accel/xen/xen-all.c    |  7 +------
 3 files changed, 17 insertions(+), 10 deletions(-)

--
2.21.3


Reply | Threaded
Open this post in threaded view
|

[PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Philippe Mathieu-Daudé-3
CONFIG_XEN is generated by configure and stored in "config-target.h",
which is (obviously) only include for target-specific objects.
This is a problem for target-agnostic objects as CONFIG_XEN is never
defined and xen_enabled() is always inlined as 'false'.

Fix by following the KVM schema, defining CONFIG_XEN_IS_POSSIBLE
when we don't know to force the call of the non-inlined function,
returning the xen_allowed boolean.

Fixes: da278d58a092 ("accel: Move Xen accelerator code under accel/xen/")
Reported-by: Paul Durrant <[hidden email]>
Suggested-by: Peter Maydell <[hidden email]>
Signed-off-by: Philippe Mathieu-Daudé <[hidden email]>
---
 include/sysemu/xen.h   | 18 ++++++++++++++----
 accel/stubs/xen-stub.c |  2 ++
 accel/xen/xen-all.c    |  7 +------
 3 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/include/sysemu/xen.h b/include/sysemu/xen.h
index 1ca292715e..2c2c429ea8 100644
--- a/include/sysemu/xen.h
+++ b/include/sysemu/xen.h
@@ -8,9 +8,19 @@
 #ifndef SYSEMU_XEN_H
 #define SYSEMU_XEN_H
 
-#ifdef CONFIG_XEN
+#ifdef NEED_CPU_H
+# ifdef CONFIG_XEN
+#  define CONFIG_XEN_IS_POSSIBLE
+# endif
+#else
+# define CONFIG_XEN_IS_POSSIBLE
+#endif
 
-bool xen_enabled(void);
+#ifdef CONFIG_XEN_IS_POSSIBLE
+
+extern bool xen_allowed;
+
+#define xen_enabled()           (xen_allowed)
 
 #ifndef CONFIG_USER_ONLY
 void xen_hvm_modified_memory(ram_addr_t start, ram_addr_t length);
@@ -18,7 +28,7 @@ void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
                    struct MemoryRegion *mr, Error **errp);
 #endif
 
-#else /* !CONFIG_XEN */
+#else /* !CONFIG_XEN_IS_POSSIBLE */
 
 #define xen_enabled() 0
 #ifndef CONFIG_USER_ONLY
@@ -33,6 +43,6 @@ static inline void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
 }
 #endif
 
-#endif /* CONFIG_XEN */
+#endif /* CONFIG_XEN_IS_POSSIBLE */
 
 #endif
diff --git a/accel/stubs/xen-stub.c b/accel/stubs/xen-stub.c
index dcca4e678a..8ae658acff 100644
--- a/accel/stubs/xen-stub.c
+++ b/accel/stubs/xen-stub.c
@@ -9,6 +9,8 @@
 #include "hw/xen/xen.h"
 #include "qapi/qapi-commands-misc.h"
 
+bool xen_allowed;
+
 void xenstore_store_pv_console_info(int i, Chardev *chr)
 {
 }
diff --git a/accel/xen/xen-all.c b/accel/xen/xen-all.c
index 0c24d4b191..60b971d0a8 100644
--- a/accel/xen/xen-all.c
+++ b/accel/xen/xen-all.c
@@ -32,12 +32,7 @@
     do { } while (0)
 #endif
 
-static bool xen_allowed;
-
-bool xen_enabled(void)
-{
-    return xen_allowed;
-}
+bool xen_allowed;
 
 xc_interface *xen_xc;
 xenforeignmemory_handle *xen_fmem;
--
2.21.3


Reply | Threaded
Open this post in threaded view
|

RE: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Paul Durrant-4
> -----Original Message-----
> From: Philippe Mathieu-Daudé <[hidden email]>
> Sent: 04 August 2020 08:50
> To: [hidden email]
> Cc: Peter Maydell <[hidden email]>; Anthony Perard <[hidden email]>; Paolo
> Bonzini <[hidden email]>; Stefano Stabellini <[hidden email]>; xen-
> [hidden email]; Paul Durrant <[hidden email]>; Philippe Mathieu-Daudé <[hidden email]>;
> Paul Durrant <[hidden email]>
> Subject: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects
>
> CONFIG_XEN is generated by configure and stored in "config-target.h",
> which is (obviously) only include for target-specific objects.
> This is a problem for target-agnostic objects as CONFIG_XEN is never
> defined and xen_enabled() is always inlined as 'false'.
>
> Fix by following the KVM schema, defining CONFIG_XEN_IS_POSSIBLE
> when we don't know to force the call of the non-inlined function,
> returning the xen_allowed boolean.
>
> Fixes: da278d58a092 ("accel: Move Xen accelerator code under accel/xen/")
> Reported-by: Paul Durrant <[hidden email]>
> Suggested-by: Peter Maydell <[hidden email]>
> Signed-off-by: Philippe Mathieu-Daudé <[hidden email]>
> ---
>  include/sysemu/xen.h   | 18 ++++++++++++++----
>  accel/stubs/xen-stub.c |  2 ++
>  accel/xen/xen-all.c    |  7 +------
>  3 files changed, 17 insertions(+), 10 deletions(-)
>
> diff --git a/include/sysemu/xen.h b/include/sysemu/xen.h
> index 1ca292715e..2c2c429ea8 100644
> --- a/include/sysemu/xen.h
> +++ b/include/sysemu/xen.h
> @@ -8,9 +8,19 @@
>  #ifndef SYSEMU_XEN_H
>  #define SYSEMU_XEN_H
>
> -#ifdef CONFIG_XEN
> +#ifdef NEED_CPU_H
> +# ifdef CONFIG_XEN
> +#  define CONFIG_XEN_IS_POSSIBLE
> +# endif
> +#else
> +# define CONFIG_XEN_IS_POSSIBLE
> +#endif
>
> -bool xen_enabled(void);
> +#ifdef CONFIG_XEN_IS_POSSIBLE
> +
> +extern bool xen_allowed;
> +
> +#define xen_enabled()           (xen_allowed)

Can this not move ahead of the #ifdef now (since xen_allowed is present in both xen-stub and xen-all)? I think this is what Peter was saying in his option '(2)'.

  Paul

>
>  #ifndef CONFIG_USER_ONLY
>  void xen_hvm_modified_memory(ram_addr_t start, ram_addr_t length);
> @@ -18,7 +28,7 @@ void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
>                     struct MemoryRegion *mr, Error **errp);
>  #endif
>
> -#else /* !CONFIG_XEN */
> +#else /* !CONFIG_XEN_IS_POSSIBLE */
>
>  #define xen_enabled() 0
>  #ifndef CONFIG_USER_ONLY
> @@ -33,6 +43,6 @@ static inline void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
>  }
>  #endif
>
> -#endif /* CONFIG_XEN */
> +#endif /* CONFIG_XEN_IS_POSSIBLE */
>
>  #endif
> diff --git a/accel/stubs/xen-stub.c b/accel/stubs/xen-stub.c
> index dcca4e678a..8ae658acff 100644
> --- a/accel/stubs/xen-stub.c
> +++ b/accel/stubs/xen-stub.c
> @@ -9,6 +9,8 @@
>  #include "hw/xen/xen.h"
>  #include "qapi/qapi-commands-misc.h"
>
> +bool xen_allowed;
> +
>  void xenstore_store_pv_console_info(int i, Chardev *chr)
>  {
>  }
> diff --git a/accel/xen/xen-all.c b/accel/xen/xen-all.c
> index 0c24d4b191..60b971d0a8 100644
> --- a/accel/xen/xen-all.c
> +++ b/accel/xen/xen-all.c
> @@ -32,12 +32,7 @@
>      do { } while (0)
>  #endif
>
> -static bool xen_allowed;
> -
> -bool xen_enabled(void)
> -{
> -    return xen_allowed;
> -}
> +bool xen_allowed;
>
>  xc_interface *xen_xc;
>  xenforeignmemory_handle *xen_fmem;
> --
> 2.21.3



Reply | Threaded
Open this post in threaded view
|

Re: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Philippe Mathieu-Daudé-3
Hi Paul,

On 8/4/20 9:57 AM, Paul Durrant wrote:

>> -----Original Message-----
>> From: Philippe Mathieu-Daudé <[hidden email]>
>> Sent: 04 August 2020 08:50
>> To: [hidden email]
>> Cc: Peter Maydell <[hidden email]>; Anthony Perard <[hidden email]>; Paolo
>> Bonzini <[hidden email]>; Stefano Stabellini <[hidden email]>; xen-
>> [hidden email]; Paul Durrant <[hidden email]>; Philippe Mathieu-Daudé <[hidden email]>;
>> Paul Durrant <[hidden email]>
>> Subject: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects
>>
>> CONFIG_XEN is generated by configure and stored in "config-target.h",
>> which is (obviously) only include for target-specific objects.
>> This is a problem for target-agnostic objects as CONFIG_XEN is never
>> defined and xen_enabled() is always inlined as 'false'.
>>
>> Fix by following the KVM schema, defining CONFIG_XEN_IS_POSSIBLE
>> when we don't know to force the call of the non-inlined function,
>> returning the xen_allowed boolean.
>>
>> Fixes: da278d58a092 ("accel: Move Xen accelerator code under accel/xen/")
>> Reported-by: Paul Durrant <[hidden email]>
>> Suggested-by: Peter Maydell <[hidden email]>
>> Signed-off-by: Philippe Mathieu-Daudé <[hidden email]>
>> ---
>>  include/sysemu/xen.h   | 18 ++++++++++++++----
>>  accel/stubs/xen-stub.c |  2 ++
>>  accel/xen/xen-all.c    |  7 +------
>>  3 files changed, 17 insertions(+), 10 deletions(-)
>>
>> diff --git a/include/sysemu/xen.h b/include/sysemu/xen.h
>> index 1ca292715e..2c2c429ea8 100644
>> --- a/include/sysemu/xen.h
>> +++ b/include/sysemu/xen.h
>> @@ -8,9 +8,19 @@
>>  #ifndef SYSEMU_XEN_H
>>  #define SYSEMU_XEN_H
>>
>> -#ifdef CONFIG_XEN
>> +#ifdef NEED_CPU_H
>> +# ifdef CONFIG_XEN
>> +#  define CONFIG_XEN_IS_POSSIBLE
>> +# endif
>> +#else
>> +# define CONFIG_XEN_IS_POSSIBLE
>> +#endif
>>
>> -bool xen_enabled(void);
>> +#ifdef CONFIG_XEN_IS_POSSIBLE
>> +
>> +extern bool xen_allowed;
>> +
>> +#define xen_enabled()           (xen_allowed)
>
> Can this not move ahead of the #ifdef now (since xen_allowed is present in both xen-stub and xen-all)? I think this is what Peter was saying in his option '(2)'.

I think I respected Peter's option '(2)', following how KVM does, this
is the case with stub,

>
>   Paul
>
>>
>>  #ifndef CONFIG_USER_ONLY
>>  void xen_hvm_modified_memory(ram_addr_t start, ram_addr_t length);
>> @@ -18,7 +28,7 @@ void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
>>                     struct MemoryRegion *mr, Error **errp);
>>  #endif
>>
>> -#else /* !CONFIG_XEN */
>> +#else /* !CONFIG_XEN_IS_POSSIBLE */
>>
>>  #define xen_enabled() 0

    ^^^ here is the other case,

>>  #ifndef CONFIG_USER_ONLY
>> @@ -33,6 +43,6 @@ static inline void xen_ram_alloc(ram_addr_t ram_addr, ram_addr_t size,
>>  }
>>  #endif
>>
>> -#endif /* CONFIG_XEN */
>> +#endif /* CONFIG_XEN_IS_POSSIBLE */
>>
>>  #endif
>> diff --git a/accel/stubs/xen-stub.c b/accel/stubs/xen-stub.c
>> index dcca4e678a..8ae658acff 100644
>> --- a/accel/stubs/xen-stub.c
>> +++ b/accel/stubs/xen-stub.c
>> @@ -9,6 +9,8 @@
>>  #include "hw/xen/xen.h"
>>  #include "qapi/qapi-commands-misc.h"
>>
>> +bool xen_allowed;

here is the stub,

>> +
>>  void xenstore_store_pv_console_info(int i, Chardev *chr)
>>  {
>>  }
>> diff --git a/accel/xen/xen-all.c b/accel/xen/xen-all.c
>> index 0c24d4b191..60b971d0a8 100644
>> --- a/accel/xen/xen-all.c
>> +++ b/accel/xen/xen-all.c
>> @@ -32,12 +32,7 @@
>>      do { } while (0)
>>  #endif
>>
>> -static bool xen_allowed;
>> -
>> -bool xen_enabled(void)
>> -{
>> -    return xen_allowed;
>> -}
>> +bool xen_allowed;

here is the real variable.

>>
>>  xc_interface *xen_xc;
>>  xenforeignmemory_handle *xen_fmem;
>> --
>> 2.21.3
>
>


Reply | Threaded
Open this post in threaded view
|

RE: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Paul Durrant-4
> -----Original Message-----
> From: Philippe Mathieu-Daudé <[hidden email]>
> Sent: 04 August 2020 09:35
> To: [hidden email]; [hidden email]
> Cc: 'Peter Maydell' <[hidden email]>; 'Anthony Perard' <[hidden email]>; 'Paolo
> Bonzini' <[hidden email]>; 'Stefano Stabellini' <[hidden email]>; xen-
> [hidden email]; 'Paul Durrant' <[hidden email]>
> Subject: Re: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects
>
> Hi Paul,
>
> On 8/4/20 9:57 AM, Paul Durrant wrote:
> >> -----Original Message-----
> >> From: Philippe Mathieu-Daudé <[hidden email]>
> >> Sent: 04 August 2020 08:50
> >> To: [hidden email]
> >> Cc: Peter Maydell <[hidden email]>; Anthony Perard <[hidden email]>; Paolo
> >> Bonzini <[hidden email]>; Stefano Stabellini <[hidden email]>; xen-
> >> [hidden email]; Paul Durrant <[hidden email]>; Philippe Mathieu-Daudé
> <[hidden email]>;
> >> Paul Durrant <[hidden email]>
> >> Subject: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects
> >>
> >> CONFIG_XEN is generated by configure and stored in "config-target.h",
> >> which is (obviously) only include for target-specific objects.
> >> This is a problem for target-agnostic objects as CONFIG_XEN is never
> >> defined and xen_enabled() is always inlined as 'false'.
> >>
> >> Fix by following the KVM schema, defining CONFIG_XEN_IS_POSSIBLE
> >> when we don't know to force the call of the non-inlined function,
> >> returning the xen_allowed boolean.
> >>
> >> Fixes: da278d58a092 ("accel: Move Xen accelerator code under accel/xen/")
> >> Reported-by: Paul Durrant <[hidden email]>
> >> Suggested-by: Peter Maydell <[hidden email]>
> >> Signed-off-by: Philippe Mathieu-Daudé <[hidden email]>
> >> ---
> >>  include/sysemu/xen.h   | 18 ++++++++++++++----
> >>  accel/stubs/xen-stub.c |  2 ++
> >>  accel/xen/xen-all.c    |  7 +------
> >>  3 files changed, 17 insertions(+), 10 deletions(-)
> >>
> >> diff --git a/include/sysemu/xen.h b/include/sysemu/xen.h
> >> index 1ca292715e..2c2c429ea8 100644
> >> --- a/include/sysemu/xen.h
> >> +++ b/include/sysemu/xen.h
> >> @@ -8,9 +8,19 @@
> >>  #ifndef SYSEMU_XEN_H
> >>  #define SYSEMU_XEN_H
> >>
> >> -#ifdef CONFIG_XEN
> >> +#ifdef NEED_CPU_H
> >> +# ifdef CONFIG_XEN
> >> +#  define CONFIG_XEN_IS_POSSIBLE
> >> +# endif
> >> +#else
> >> +# define CONFIG_XEN_IS_POSSIBLE
> >> +#endif
> >>
> >> -bool xen_enabled(void);
> >> +#ifdef CONFIG_XEN_IS_POSSIBLE
> >> +
> >> +extern bool xen_allowed;
> >> +
> >> +#define xen_enabled()           (xen_allowed)
> >
> > Can this not move ahead of the #ifdef now (since xen_allowed is present in both xen-stub and xen-
> all)? I think this is what Peter was saying in his option '(2)'.
>
> I think I respected Peter's option '(2)', following how KVM does, this
> is the case with stub,

Ok, if it follows the KVM pattern then that's fine.

Reviewed-by: Paul Durrant <[hidden email]>


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH-for-5.1 v2 1/1] accel/xen: Fix xen_enabled() behavior on target-agnostic objects

Anthony PERARD-2
In reply to this post by Philippe Mathieu-Daudé-3
On Tue, Aug 04, 2020 at 09:49:30AM +0200, Philippe Mathieu-Daudé wrote:

> CONFIG_XEN is generated by configure and stored in "config-target.h",
> which is (obviously) only include for target-specific objects.
> This is a problem for target-agnostic objects as CONFIG_XEN is never
> defined and xen_enabled() is always inlined as 'false'.
>
> Fix by following the KVM schema, defining CONFIG_XEN_IS_POSSIBLE
> when we don't know to force the call of the non-inlined function,
> returning the xen_allowed boolean.
>
> Fixes: da278d58a092 ("accel: Move Xen accelerator code under accel/xen/")
> Reported-by: Paul Durrant <[hidden email]>
> Suggested-by: Peter Maydell <[hidden email]>
> Signed-off-by: Philippe Mathieu-Daudé <[hidden email]>

Reviewed-by: Anthony PERARD <[hidden email]>

This time, I've also check that build all targets works...

I'll prepare the pull request right away.

Thank you,

--
Anthony PERARD