Nothing in sigcontext.h seems to require anything from
linux/posix_types.h.  This include seems a MIPS relic originated from
an error in Linux 2.6.11-rc2 (in 2005).

The unneeded include was found debugging some vDSO self test build
failure (it's not the root cause though).

Link: https://lore.kernel.org/linux-mips/20240828030413.143930-2-xry111@xry111.site/
Link: https://lore.kernel.org/loongarch/0b540679ec8cfccec75aeb3463810924f6ff71e6.camel@xry111.site/
Signed-off-by: Xi Ruoyao <xry111@xry111.site>
---
 arch/loongarch/include/uapi/asm/sigcontext.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/arch/loongarch/include/uapi/asm/sigcontext.h b/arch/loongarch/include/uapi/asm/sigcontext.h
index 6c22f616b8f1..5cd121275bac 100644
--- a/arch/loongarch/include/uapi/asm/sigcontext.h
+++ b/arch/loongarch/include/uapi/asm/sigcontext.h
@@ -9,7 +9,6 @@
 #define _UAPI_ASM_SIGCONTEXT_H
 
 #include <linux/types.h>
-#include <linux/posix_types.h>
 
 /* FP context was used */
 #define SC_USED_FP		(1 << 0)
-- 
2.46.0

Hi, Ruoyao,

Just some questions:
1, Changing UAPI is not a good idea.
2. In another thread you said that "paper over" is not enough for some cases.
3. include/uapi/linux/types.h still include linux/posix_types.h, why
your "paper over" works?

Huacai

On Wed, Aug 28, 2024 at 11:14 AM Xi Ruoyao <xry111@xry111.site> wrote:
>
> Nothing in sigcontext.h seems to require anything from
> linux/posix_types.h.  This include seems a MIPS relic originated from
> an error in Linux 2.6.11-rc2 (in 2005).
>
> The unneeded include was found debugging some vDSO self test build
> failure (it's not the root cause though).
>
> Link: https://lore.kernel.org/linux-mips/20240828030413.143930-2-xry111@xry111.site/
> Link: https://lore.kernel.org/loongarch/0b540679ec8cfccec75aeb3463810924f6ff71e6.camel@xry111.site/
> Signed-off-by: Xi Ruoyao <xry111@xry111.site>
> ---
>  arch/loongarch/include/uapi/asm/sigcontext.h | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/arch/loongarch/include/uapi/asm/sigcontext.h b/arch/loongarch/include/uapi/asm/sigcontext.h
> index 6c22f616b8f1..5cd121275bac 100644
> --- a/arch/loongarch/include/uapi/asm/sigcontext.h
> +++ b/arch/loongarch/include/uapi/asm/sigcontext.h
> @@ -9,7 +9,6 @@
>  #define _UAPI_ASM_SIGCONTEXT_H
>
>  #include <linux/types.h>
> -#include <linux/posix_types.h>
>
>  /* FP context was used */
>  #define SC_USED_FP             (1 << 0)
> --
> 2.46.0
>

On Wed, 2024-08-28 at 12:11 +0800, Huacai Chen wrote:
> Hi, Ruoyao,
> 
> Just some questions:
> 1, Changing UAPI is not a good idea.

But removing unneeded includes from UAPI is fine.  For example, the
commit 44e0b165b6c078b84767da4ba06ffa27af562c96 has removed
linux/posix_types.h from termbits.h for all ports.

> 2. In another thread you said that "paper over" is not enough for some
> cases.
> 3. include/uapi/linux/types.h still include linux/posix_types.h, why
> your "paper over" works?

Well maybe it does not work (I've not seriously tested as it's just a
paper over).  But removing unneeded include is correct on its own
anyway.  And ...

> > The unneeded include was found debugging some vDSO self test build
> > failure (it's not the root cause though).

So it's just "found" debugging the issue.  Maybe I should change "it's
not the root cause" to "it's not really related to that issue"?
> 

-- 
Xi Ruoyao <xry111@xry111.site>
School of Aerospace Science and Technology, Xidian University

On Wed, Aug 28, 2024 at 12:27 PM Xi Ruoyao <xry111@xry111.site> wrote:
>
> On Wed, 2024-08-28 at 12:11 +0800, Huacai Chen wrote:
> > Hi, Ruoyao,
> >
> > Just some questions:
> > 1, Changing UAPI is not a good idea.
>
> But removing unneeded includes from UAPI is fine.  For example, the
> commit 44e0b165b6c078b84767da4ba06ffa27af562c96 has removed
> linux/posix_types.h from termbits.h for all ports.
Need more tests, not just kselftests.

Huacai

>
> > 2. In another thread you said that "paper over" is not enough for some
> > cases.
> > 3. include/uapi/linux/types.h still include linux/posix_types.h, why
> > your "paper over" works?
>
> Well maybe it does not work (I've not seriously tested as it's just a
> paper over).  But removing unneeded include is correct on its own
> anyway.  And ...
>
> > > The unneeded include was found debugging some vDSO self test build
> > > failure (it's not the root cause though).
>
> So it's just "found" debugging the issue.  Maybe I should change "it's
> not the root cause" to "it's not really related to that issue"?
> >
>
> --
> Xi Ruoyao <xry111@xry111.site>
> School of Aerospace Science and Technology, Xidian University

On Wed, 2024-08-28 at 12:43 +0800, Huacai Chen wrote:
> On Wed, Aug 28, 2024 at 12:27 PM Xi Ruoyao <xry111@xry111.site> wrote:
> > 
> > On Wed, 2024-08-28 at 12:11 +0800, Huacai Chen wrote:
> > > Hi, Ruoyao,
> > > 
> > > Just some questions:
> > > 1, Changing UAPI is not a good idea.
> > 
> > But removing unneeded includes from UAPI is fine.  For example, the
> > commit 44e0b165b6c078b84767da4ba06ffa27af562c96 has removed
> > linux/posix_types.h from termbits.h for all ports.
> Need more tests, not just kselftests.

I've built and tested Glibc with this change.  But if you want a full
system rebuild it won't happen in the next few months, and I can resend
the patch after the next time I rebuild the system.


-- 
Xi Ruoyao <xry111@xry111.site>
School of Aerospace Science and Technology, Xidian University