Commit 0a987645672e ("um: Allow building and running on older
hosts") attempted to check for PTRACE_{GET,SET}REGSET under the premise
that these ptrace(2) parameters were directly linked with the presence
of the _xstate structure.
After Richard's commit 61e8d462457f ("um: Correctly check for
PTRACE_GETRESET/SETREGSET") which properly included linux/ptrace.h
instead of asm/ptrace.h, we could get into the original build failure
that I reported:
arch/x86/um/user-offsets.c: In function 'foo':
arch/x86/um/user-offsets.c:54: error: invalid application of 'sizeof' to
incomplete type 'struct _xstate'
On this particular host, we do have PTRACE_GETREGSET and
PTRACE_SETREGSET defined in linux/ptrace.h, but not the structure
_xstate that should be pulled from the following include chain: signal.h
-> bits/sigcontext.h.
Correctly fix this by checking for FP_XSTATE_MAGIC1 which is the correct
way to see if struct _xstate is available or not on the host.
Fixes: 61e8d462457f ("um: Correctly check for PTRACE_GETRESET/SETREGSET")
Fixes: 0a987645672e ("um: Allow building and running on older hosts")
Signed-off-by: Florian Fainelli <[email protected]>
---
arch/x86/um/user-offsets.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/x86/um/user-offsets.c b/arch/x86/um/user-offsets.c
index ae4cd58c0c7a..02250b2633b8 100644
--- a/arch/x86/um/user-offsets.c
+++ b/arch/x86/um/user-offsets.c
@@ -50,7 +50,7 @@ void foo(void)
DEFINE(HOST_GS, GS);
DEFINE(HOST_ORIG_AX, ORIG_EAX);
#else
-#if defined(PTRACE_GETREGSET) && defined(PTRACE_SETREGSET)
+#ifdef FP_XSTATE_MAGIC1
DEFINE(HOST_FP_SIZE, sizeof(struct _xstate) / sizeof(unsigned long));
#else
DEFINE(HOST_FP_SIZE, sizeof(struct _fpstate) / sizeof(unsigned long));
--
2.9.3
On Wed, Jul 19, 2017 at 1:43 AM, Florian Fainelli <[email protected]> wrote:
> Commit 0a987645672e ("um: Allow building and running on older
> hosts") attempted to check for PTRACE_{GET,SET}REGSET under the premise
> that these ptrace(2) parameters were directly linked with the presence
> of the _xstate structure.
>
> After Richard's commit 61e8d462457f ("um: Correctly check for
> PTRACE_GETRESET/SETREGSET") which properly included linux/ptrace.h
> instead of asm/ptrace.h, we could get into the original build failure
> that I reported:
>
> arch/x86/um/user-offsets.c: In function 'foo':
> arch/x86/um/user-offsets.c:54: error: invalid application of 'sizeof' to
> incomplete type 'struct _xstate'
>
> On this particular host, we do have PTRACE_GETREGSET and
> PTRACE_SETREGSET defined in linux/ptrace.h, but not the structure
> _xstate that should be pulled from the following include chain: signal.h
> -> bits/sigcontext.h.
>
> Correctly fix this by checking for FP_XSTATE_MAGIC1 which is the correct
> way to see if struct _xstate is available or not on the host.
>
> Fixes: 61e8d462457f ("um: Correctly check for PTRACE_GETRESET/SETREGSET")
> Fixes: 0a987645672e ("um: Allow building and running on older hosts")
> Signed-off-by: Florian Fainelli <[email protected]>
> ---
> arch/x86/um/user-offsets.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/um/user-offsets.c b/arch/x86/um/user-offsets.c
> index ae4cd58c0c7a..02250b2633b8 100644
> --- a/arch/x86/um/user-offsets.c
> +++ b/arch/x86/um/user-offsets.c
> @@ -50,7 +50,7 @@ void foo(void)
> DEFINE(HOST_GS, GS);
> DEFINE(HOST_ORIG_AX, ORIG_EAX);
> #else
> -#if defined(PTRACE_GETREGSET) && defined(PTRACE_SETREGSET)
> +#ifdef FP_XSTATE_MAGIC1
> DEFINE(HOST_FP_SIZE, sizeof(struct _xstate) / sizeof(unsigned long));
> #else
> DEFINE(HOST_FP_SIZE, sizeof(struct _fpstate) / sizeof(unsigned long));
> --
> 2.9.3
>
Applied, thanks for fixing!
--
Thanks,
//richard
On 07/24/2017 05:46 AM, Richard Weinberger wrote:
> On Wed, Jul 19, 2017 at 1:43 AM, Florian Fainelli <[email protected]> wrote:
>> Commit 0a987645672e ("um: Allow building and running on older
>> hosts") attempted to check for PTRACE_{GET,SET}REGSET under the premise
>> that these ptrace(2) parameters were directly linked with the presence
>> of the _xstate structure.
>>
>> After Richard's commit 61e8d462457f ("um: Correctly check for
>> PTRACE_GETRESET/SETREGSET") which properly included linux/ptrace.h
>> instead of asm/ptrace.h, we could get into the original build failure
>> that I reported:
>>
>> arch/x86/um/user-offsets.c: In function 'foo':
>> arch/x86/um/user-offsets.c:54: error: invalid application of 'sizeof' to
>> incomplete type 'struct _xstate'
>>
>> On this particular host, we do have PTRACE_GETREGSET and
>> PTRACE_SETREGSET defined in linux/ptrace.h, but not the structure
>> _xstate that should be pulled from the following include chain: signal.h
>> -> bits/sigcontext.h.
>>
>> Correctly fix this by checking for FP_XSTATE_MAGIC1 which is the correct
>> way to see if struct _xstate is available or not on the host.
>>
>> Fixes: 61e8d462457f ("um: Correctly check for PTRACE_GETRESET/SETREGSET")
>> Fixes: 0a987645672e ("um: Allow building and running on older hosts")
>> Signed-off-by: Florian Fainelli <[email protected]>
>> ---
>> arch/x86/um/user-offsets.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/x86/um/user-offsets.c b/arch/x86/um/user-offsets.c
>> index ae4cd58c0c7a..02250b2633b8 100644
>> --- a/arch/x86/um/user-offsets.c
>> +++ b/arch/x86/um/user-offsets.c
>> @@ -50,7 +50,7 @@ void foo(void)
>> DEFINE(HOST_GS, GS);
>> DEFINE(HOST_ORIG_AX, ORIG_EAX);
>> #else
>> -#if defined(PTRACE_GETREGSET) && defined(PTRACE_SETREGSET)
>> +#ifdef FP_XSTATE_MAGIC1
>> DEFINE(HOST_FP_SIZE, sizeof(struct _xstate) / sizeof(unsigned long));
>> #else
>> DEFINE(HOST_FP_SIZE, sizeof(struct _fpstate) / sizeof(unsigned long));
>> --
>> 2.9.3
>>
>
> Applied, thanks for fixing!
Can you submit this to Linus before v4.13 is released? Thank you!
--
Florian
Am Freitag, 18. August 2017, 23:40:36 CEST schrieb Florian Fainelli:
> On 07/24/2017 05:46 AM, Richard Weinberger wrote:
> > On Wed, Jul 19, 2017 at 1:43 AM, Florian Fainelli <[email protected]>
wrote:
> >> Commit 0a987645672e ("um: Allow building and running on older
> >> hosts") attempted to check for PTRACE_{GET,SET}REGSET under the premise
> >> that these ptrace(2) parameters were directly linked with the presence
> >> of the _xstate structure.
> >>
> >> After Richard's commit 61e8d462457f ("um: Correctly check for
> >> PTRACE_GETRESET/SETREGSET") which properly included linux/ptrace.h
> >> instead of asm/ptrace.h, we could get into the original build failure
> >> that I reported:
> >>
> >> arch/x86/um/user-offsets.c: In function 'foo':
> >> arch/x86/um/user-offsets.c:54: error: invalid application of 'sizeof' to
> >> incomplete type 'struct _xstate'
> >>
> >> On this particular host, we do have PTRACE_GETREGSET and
> >> PTRACE_SETREGSET defined in linux/ptrace.h, but not the structure
> >> _xstate that should be pulled from the following include chain: signal.h
> >> -> bits/sigcontext.h.
> >>
> >> Correctly fix this by checking for FP_XSTATE_MAGIC1 which is the correct
> >> way to see if struct _xstate is available or not on the host.
> >>
> >> Fixes: 61e8d462457f ("um: Correctly check for PTRACE_GETRESET/SETREGSET")
> >> Fixes: 0a987645672e ("um: Allow building and running on older hosts")
> >> Signed-off-by: Florian Fainelli <[email protected]>
> >> ---
> >>
> >> arch/x86/um/user-offsets.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/arch/x86/um/user-offsets.c b/arch/x86/um/user-offsets.c
> >> index ae4cd58c0c7a..02250b2633b8 100644
> >> --- a/arch/x86/um/user-offsets.c
> >> +++ b/arch/x86/um/user-offsets.c
> >> @@ -50,7 +50,7 @@ void foo(void)
> >>
> >> DEFINE(HOST_GS, GS);
> >> DEFINE(HOST_ORIG_AX, ORIG_EAX);
> >>
> >> #else
> >>
> >> -#if defined(PTRACE_GETREGSET) && defined(PTRACE_SETREGSET)
> >> +#ifdef FP_XSTATE_MAGIC1
> >>
> >> DEFINE(HOST_FP_SIZE, sizeof(struct _xstate) / sizeof(unsigned
> >> long));
> >>
> >> #else
> >>
> >> DEFINE(HOST_FP_SIZE, sizeof(struct _fpstate) / sizeof(unsigned
> >> long));
> >>
> >> --
> >> 2.9.3
> >
> > Applied, thanks for fixing!
>
> Can you submit this to Linus before v4.13 is released? Thank you!
Sure. The UML queue is just a bit slow, sorry for that.
Thanks,
//richard
--
sigma star gmbh - Eduard-Bodem-Gasse 6 - 6020 Innsbruck - Austria
ATU66964118 - FN 374287y