[musl] SA_RESTORER for rv64?

[musl] SA_RESTORER for rv64?

[enh]

[-- Attachment #1: Type: text/plain, Size: 282 bytes --]

arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
the initial commit, but i think that's just copy & paste from whichever
architecture the rv64 headers were based on? the linux kernel itself
doesn't have SA_RESTORER for rv64, unless i'm missing something?

[-- Attachment #2: Type: text/html, Size: 320 bytes --]

Re: [musl] SA_RESTORER for rv64?

[Rich Felker]

On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> the initial commit, but i think that's just copy & paste from whichever
> architecture the rv64 headers were based on? the linux kernel itself
> doesn't have SA_RESTORER for rv64, unless i'm missing something?

I suspect this is just a mistake. Have you seen any ill effects from
it? If riscv folks can confirm it's wrong, I'll remove it.

Rich

Re: [musl] SA_RESTORER for rv64?

[enh]

[-- Attachment #1: Type: text/plain, Size: 834 bytes --]

On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:

> On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER
> since
> > the initial commit, but i think that's just copy & paste from whichever
> > architecture the rv64 headers were based on? the linux kernel itself
> > doesn't have SA_RESTORER for rv64, unless i'm missing something?
>
> I suspect this is just a mistake. Have you seen any ill effects from
> it? If riscv folks can confirm it's wrong, I'll remove it.
>

i noticed this by inspection after someone had some code that (correctly)
didn't compile with bionic or glibc but did with musl. in addition to
rejecting the proposal to add a rv64 SA_RESTORER to bionic, i said i'd
reach out to musl about removing it there :-)


> Rich
>

[-- Attachment #2: Type: text/html, Size: 1371 bytes --]

Re: [musl] SA_RESTORER for rv64?

[Khem Raj]

On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
>
> On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > the initial commit, but i think that's just copy & paste from whichever
> > architecture the rv64 headers were based on? the linux kernel itself
> > doesn't have SA_RESTORER for rv64, unless i'm missing something?
>
> I suspect this is just a mistake. Have you seen any ill effects from
> it? If riscv folks can confirm it's wrong, I'll remove it.

Yeah I think it should be removed. Perhaps mips is in same boat.

>
> Rich

Re: [musl] SA_RESTORER for rv64?

[enh]

[-- Attachment #1: Type: text/plain, Size: 783 bytes --]

oops, never actually sent the patch. attached...

On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
>
> On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> >
> > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > the initial commit, but i think that's just copy & paste from whichever
> > > architecture the rv64 headers were based on? the linux kernel itself
> > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> >
> > I suspect this is just a mistake. Have you seen any ill effects from
> > it? If riscv folks can confirm it's wrong, I'll remove it.
>
> Yeah I think it should be removed. Perhaps mips is in same boat.
>
> >
> > Rich

[-- Attachment #2: 0001-risc-v-does-not-have-SA_RESTORER.patch --]
[-- Type: text/x-patch, Size: 795 bytes --]

From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
From: Elliott Hughes <enh@google.com>
Date: Fri, 3 Feb 2023 10:42:55 -0800
Subject: [PATCH] risc-v does not have SA_RESTORER.

The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
this out as obsolete. New architectures like risc-v do not define it.
---
 arch/riscv64/bits/signal.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
index 287367db..fd6157a3 100644
--- a/arch/riscv64/bits/signal.h
+++ b/arch/riscv64/bits/signal.h
@@ -76,7 +76,6 @@ typedef struct __ucontext
 #define SA_RESTART   0x10000000
 #define SA_NODEFER   0x40000000
 #define SA_RESETHAND 0x80000000
-#define SA_RESTORER  0x04000000
 
 #endif
 
-- 
2.39.1.519.gcb327c4b5f-goog

Re: [musl] SA_RESTORER for rv64?

[Rich Felker]

On Fri, Feb 03, 2023 at 10:44:56AM -0800, enh wrote:
> oops, never actually sent the patch. attached...
> 
> On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
> >
> > On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> > >
> > > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > > the initial commit, but i think that's just copy & paste from whichever
> > > > architecture the rv64 headers were based on? the linux kernel itself
> > > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> > >
> > > I suspect this is just a mistake. Have you seen any ill effects from
> > > it? If riscv folks can confirm it's wrong, I'll remove it.
> >
> > Yeah I think it should be removed. Perhaps mips is in same boat.
> >
> > >
> > > Rich

> From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
> From: Elliott Hughes <enh@google.com>
> Date: Fri, 3 Feb 2023 10:42:55 -0800
> Subject: [PATCH] risc-v does not have SA_RESTORER.
> 
> The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
> this out as obsolete. New architectures like risc-v do not define it.
> ---
>  arch/riscv64/bits/signal.h | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
> index 287367db..fd6157a3 100644
> --- a/arch/riscv64/bits/signal.h
> +++ b/arch/riscv64/bits/signal.h
> @@ -76,7 +76,6 @@ typedef struct __ucontext
>  #define SA_RESTART   0x10000000
>  #define SA_NODEFER   0x40000000
>  #define SA_RESETHAND 0x80000000
> -#define SA_RESTORER  0x04000000
>  
>  #endif
>  
> -- 
> 2.39.1.519.gcb327c4b5f-goog
> 

I don't think this patch works as-is, since musl unconditionally uses
SA_RESTORER. We probably need to make that conditional on its
presence, and it looks like there's also a wrong-struct-layout issue
on archs where it's absent...

Rich

Re: [musl] SA_RESTORER for rv64?

[enh]

On Sun, Feb 5, 2023 at 3:54 PM Rich Felker <dalias@libc.org> wrote:
>
> On Fri, Feb 03, 2023 at 10:44:56AM -0800, enh wrote:
> > oops, never actually sent the patch. attached...
> >
> > On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
> > >
> > > On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> > > >
> > > > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > > > the initial commit, but i think that's just copy & paste from whichever
> > > > > architecture the rv64 headers were based on? the linux kernel itself
> > > > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> > > >
> > > > I suspect this is just a mistake. Have you seen any ill effects from
> > > > it? If riscv folks can confirm it's wrong, I'll remove it.
> > >
> > > Yeah I think it should be removed. Perhaps mips is in same boat.
> > >
> > > >
> > > > Rich
>
> > From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
> > From: Elliott Hughes <enh@google.com>
> > Date: Fri, 3 Feb 2023 10:42:55 -0800
> > Subject: [PATCH] risc-v does not have SA_RESTORER.
> >
> > The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
> > this out as obsolete. New architectures like risc-v do not define it.
> > ---
> >  arch/riscv64/bits/signal.h | 1 -
> >  1 file changed, 1 deletion(-)
> >
> > diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
> > index 287367db..fd6157a3 100644
> > --- a/arch/riscv64/bits/signal.h
> > +++ b/arch/riscv64/bits/signal.h
> > @@ -76,7 +76,6 @@ typedef struct __ucontext
> >  #define SA_RESTART   0x10000000
> >  #define SA_NODEFER   0x40000000
> >  #define SA_RESETHAND 0x80000000
> > -#define SA_RESTORER  0x04000000
> >
> >  #endif
> >
> > --
> > 2.39.1.519.gcb327c4b5f-goog
> >
>
> I don't think this patch works as-is, since musl unconditionally uses
> SA_RESTORER. We probably need to make that conditional on its
> presence, and it looks like there's also a wrong-struct-layout issue
> on archs where it's absent...

yeah, bionic just uses the kernel uapi headers directly, and they look
like this:

struct sigaction {
  __sighandler_t sa_handler;
  unsigned long sa_flags;
#ifdef SA_RESTORER
  __sigrestore_t sa_restorer;
#endif
  sigset_t sa_mask;
};


> Rich

Re: [musl] SA_RESTORER for rv64?

[Rich Felker]

On Mon, Feb 06, 2023 at 08:51:13AM -0800, enh wrote:
> On Sun, Feb 5, 2023 at 3:54 PM Rich Felker <dalias@libc.org> wrote:
> >
> > On Fri, Feb 03, 2023 at 10:44:56AM -0800, enh wrote:
> > > oops, never actually sent the patch. attached...
> > >
> > > On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
> > > >
> > > > On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> > > > >
> > > > > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > > > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > > > > the initial commit, but i think that's just copy & paste from whichever
> > > > > > architecture the rv64 headers were based on? the linux kernel itself
> > > > > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> > > > >
> > > > > I suspect this is just a mistake. Have you seen any ill effects from
> > > > > it? If riscv folks can confirm it's wrong, I'll remove it.
> > > >
> > > > Yeah I think it should be removed. Perhaps mips is in same boat.
> > > >
> > > > >
> > > > > Rich
> >
> > > From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
> > > From: Elliott Hughes <enh@google.com>
> > > Date: Fri, 3 Feb 2023 10:42:55 -0800
> > > Subject: [PATCH] risc-v does not have SA_RESTORER.
> > >
> > > The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
> > > this out as obsolete. New architectures like risc-v do not define it.
> > > ---
> > >  arch/riscv64/bits/signal.h | 1 -
> > >  1 file changed, 1 deletion(-)
> > >
> > > diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
> > > index 287367db..fd6157a3 100644
> > > --- a/arch/riscv64/bits/signal.h
> > > +++ b/arch/riscv64/bits/signal.h
> > > @@ -76,7 +76,6 @@ typedef struct __ucontext
> > >  #define SA_RESTART   0x10000000
> > >  #define SA_NODEFER   0x40000000
> > >  #define SA_RESETHAND 0x80000000
> > > -#define SA_RESTORER  0x04000000
> > >
> > >  #endif
> > >
> > > --
> > > 2.39.1.519.gcb327c4b5f-goog
> > >
> >
> > I don't think this patch works as-is, since musl unconditionally uses
> > SA_RESTORER. We probably need to make that conditional on its
> > presence, and it looks like there's also a wrong-struct-layout issue
> > on archs where it's absent...
> 
> yeah, bionic just uses the kernel uapi headers directly, and they look
> like this:
> 
> struct sigaction {
>   __sighandler_t sa_handler;
>   unsigned long sa_flags;
> #ifdef SA_RESTORER
>   __sigrestore_t sa_restorer;
> #endif
>   sigset_t sa_mask;
> };

OK. It looks like we need to remove the wrong SA_RESTORER for archs
that aren't supposed to have it *and* add such an #ifdef. Right now,
we're passing bogus sa_mask on these archs... :(

Rich

Re: [musl] SA_RESTORER for rv64?

[Rich Felker]

[-- Attachment #1: Type: text/plain, Size: 2964 bytes --]

On Mon, Feb 06, 2023 at 12:49:53PM -0500, Rich Felker wrote:
> On Mon, Feb 06, 2023 at 08:51:13AM -0800, enh wrote:
> > On Sun, Feb 5, 2023 at 3:54 PM Rich Felker <dalias@libc.org> wrote:
> > >
> > > On Fri, Feb 03, 2023 at 10:44:56AM -0800, enh wrote:
> > > > oops, never actually sent the patch. attached...
> > > >
> > > > On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
> > > > >
> > > > > On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> > > > > >
> > > > > > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > > > > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > > > > > the initial commit, but i think that's just copy & paste from whichever
> > > > > > > architecture the rv64 headers were based on? the linux kernel itself
> > > > > > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> > > > > >
> > > > > > I suspect this is just a mistake. Have you seen any ill effects from
> > > > > > it? If riscv folks can confirm it's wrong, I'll remove it.
> > > > >
> > > > > Yeah I think it should be removed. Perhaps mips is in same boat.
> > > > >
> > > > > >
> > > > > > Rich
> > >
> > > > From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
> > > > From: Elliott Hughes <enh@google.com>
> > > > Date: Fri, 3 Feb 2023 10:42:55 -0800
> > > > Subject: [PATCH] risc-v does not have SA_RESTORER.
> > > >
> > > > The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
> > > > this out as obsolete. New architectures like risc-v do not define it.
> > > > ---
> > > >  arch/riscv64/bits/signal.h | 1 -
> > > >  1 file changed, 1 deletion(-)
> > > >
> > > > diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
> > > > index 287367db..fd6157a3 100644
> > > > --- a/arch/riscv64/bits/signal.h
> > > > +++ b/arch/riscv64/bits/signal.h
> > > > @@ -76,7 +76,6 @@ typedef struct __ucontext
> > > >  #define SA_RESTART   0x10000000
> > > >  #define SA_NODEFER   0x40000000
> > > >  #define SA_RESETHAND 0x80000000
> > > > -#define SA_RESTORER  0x04000000
> > > >
> > > >  #endif
> > > >
> > > > --
> > > > 2.39.1.519.gcb327c4b5f-goog
> > > >
> > >
> > > I don't think this patch works as-is, since musl unconditionally uses
> > > SA_RESTORER. We probably need to make that conditional on its
> > > presence, and it looks like there's also a wrong-struct-layout issue
> > > on archs where it's absent...
> > 
> > yeah, bionic just uses the kernel uapi headers directly, and they look
> > like this:
> > 
> > struct sigaction {
> >   __sighandler_t sa_handler;
> >   unsigned long sa_flags;
> > #ifdef SA_RESTORER
> >   __sigrestore_t sa_restorer;
> > #endif
> >   sigset_t sa_mask;
> > };
> 
> OK. It looks like we need to remove the wrong SA_RESTORER for archs
> that aren't supposed to have it *and* add such an #ifdef. Right now,
> we're passing bogus sa_mask on these archs... :(

How does the attached look?

Rich

[-- Attachment #2: SA_RESTORER.diff --]
[-- Type: text/plain, Size: 4369 bytes --]

diff --git a/arch/microblaze/bits/signal.h b/arch/microblaze/bits/signal.h
index 490f83bf..f25b7c6a 100644
--- a/arch/microblaze/bits/signal.h
+++ b/arch/microblaze/bits/signal.h
@@ -46,7 +46,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #endif
 
diff --git a/arch/mips/bits/signal.h b/arch/mips/bits/signal.h
index 1b69e762..a3b3857a 100644
--- a/arch/mips/bits/signal.h
+++ b/arch/mips/bits/signal.h
@@ -66,7 +66,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #undef SIG_BLOCK
 #undef SIG_UNBLOCK
diff --git a/arch/mips/ksigaction.h b/arch/mips/ksigaction.h
index 63fdfab0..2141a0a2 100644
--- a/arch/mips/ksigaction.h
+++ b/arch/mips/ksigaction.h
@@ -4,10 +4,6 @@ struct k_sigaction {
 	unsigned flags;
 	void (*handler)(int);
 	unsigned long mask[4];
-	/* The following field is past the end of the structure the
-	 * kernel will read or write, and exists only to avoid having
-	 * mips-specific preprocessor conditionals in sigaction.c. */
-	void (*restorer)();
 };
 
 hidden void __restore(), __restore_rt();
diff --git a/arch/mips64/bits/signal.h b/arch/mips64/bits/signal.h
index 4f91c9fc..ffec7fd0 100644
--- a/arch/mips64/bits/signal.h
+++ b/arch/mips64/bits/signal.h
@@ -85,7 +85,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #undef SIG_BLOCK
 #undef SIG_UNBLOCK
diff --git a/arch/mips64/ksigaction.h b/arch/mips64/ksigaction.h
index c16e4731..c0b73ae9 100644
--- a/arch/mips64/ksigaction.h
+++ b/arch/mips64/ksigaction.h
@@ -4,7 +4,6 @@ struct k_sigaction {
 	unsigned flags;
 	void (*handler)(int);
 	unsigned long mask[2];
-	void (*restorer)();
 };
 
 hidden void __restore(), __restore_rt();
diff --git a/arch/mipsn32/bits/signal.h b/arch/mipsn32/bits/signal.h
index 4f91c9fc..ffec7fd0 100644
--- a/arch/mipsn32/bits/signal.h
+++ b/arch/mipsn32/bits/signal.h
@@ -85,7 +85,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #undef SIG_BLOCK
 #undef SIG_UNBLOCK
diff --git a/arch/mipsn32/ksigaction.h b/arch/mipsn32/ksigaction.h
index b565f1fc..2141a0a2 100644
--- a/arch/mipsn32/ksigaction.h
+++ b/arch/mipsn32/ksigaction.h
@@ -4,7 +4,6 @@ struct k_sigaction {
 	unsigned flags;
 	void (*handler)(int);
 	unsigned long mask[4];
-	void (*restorer)();
 };
 
 hidden void __restore(), __restore_rt();
diff --git a/arch/or1k/bits/signal.h b/arch/or1k/bits/signal.h
index be576d1d..c45be676 100644
--- a/arch/or1k/bits/signal.h
+++ b/arch/or1k/bits/signal.h
@@ -43,7 +43,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #endif
 
diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
index 287367db..fd6157a3 100644
--- a/arch/riscv64/bits/signal.h
+++ b/arch/riscv64/bits/signal.h
@@ -76,7 +76,6 @@ typedef struct __ucontext
 #define SA_RESTART   0x10000000
 #define SA_NODEFER   0x40000000
 #define SA_RESETHAND 0x80000000
-#define SA_RESTORER  0x04000000
 
 #endif
 
diff --git a/src/internal/ksigaction.h b/src/internal/ksigaction.h
index 8ebd5938..f0b6a837 100644
--- a/src/internal/ksigaction.h
+++ b/src/internal/ksigaction.h
@@ -6,7 +6,9 @@
 struct k_sigaction {
 	void (*handler)(int);
 	unsigned long flags;
+#ifdef SA_RESTORER
 	void (*restorer)(void);
+#endif
 	unsigned mask[2];
 };
 
diff --git a/src/signal/sigaction.c b/src/signal/sigaction.c
index 2203471b..e45308fa 100644
--- a/src/signal/sigaction.c
+++ b/src/signal/sigaction.c
@@ -44,8 +44,11 @@ int __libc_sigaction(int sig, const struct sigaction *restrict sa, struct sigact
 			}
 		}
 		ksa.handler = sa->sa_handler;
-		ksa.flags = sa->sa_flags | SA_RESTORER;
+		ksa.flags = sa->sa_flags;
+#ifdef SA_RESTORER
+		ksa.flags |= SA_RESTORER;
 		ksa.restorer = (sa->sa_flags & SA_SIGINFO) ? __restore_rt : __restore;
+#endif
 		memcpy(&ksa.mask, &sa->sa_mask, _NSIG/8);
 	}
 	int r = __syscall(SYS_rt_sigaction, sig, sa?&ksa:0, old?&ksa_old:0, _NSIG/8);

Re: [musl] SA_RESTORER for rv64?

[enh]

yes, that looks like what we have in bionic for riscv64 (and what we
had for mips before it was removed). thanks!

On Wed, Feb 8, 2023 at 1:45 PM Rich Felker <dalias@libc.org> wrote:
>
> On Mon, Feb 06, 2023 at 12:49:53PM -0500, Rich Felker wrote:
> > On Mon, Feb 06, 2023 at 08:51:13AM -0800, enh wrote:
> > > On Sun, Feb 5, 2023 at 3:54 PM Rich Felker <dalias@libc.org> wrote:
> > > >
> > > > On Fri, Feb 03, 2023 at 10:44:56AM -0800, enh wrote:
> > > > > oops, never actually sent the patch. attached...
> > > > >
> > > > > On Thu, Nov 10, 2022 at 9:31 AM Khem Raj <raj.khem@gmail.com> wrote:
> > > > > >
> > > > > > On Thu, Nov 10, 2022 at 9:19 AM Rich Felker <dalias@libc.org> wrote:
> > > > > > >
> > > > > > > On Thu, Nov 10, 2022 at 07:44:23AM -0800, enh wrote:
> > > > > > > > arch/riscv64/bits/signal.h has contained a definition for SA_RESTORER since
> > > > > > > > the initial commit, but i think that's just copy & paste from whichever
> > > > > > > > architecture the rv64 headers were based on? the linux kernel itself
> > > > > > > > doesn't have SA_RESTORER for rv64, unless i'm missing something?
> > > > > > >
> > > > > > > I suspect this is just a mistake. Have you seen any ill effects from
> > > > > > > it? If riscv folks can confirm it's wrong, I'll remove it.
> > > > > >
> > > > > > Yeah I think it should be removed. Perhaps mips is in same boat.
> > > > > >
> > > > > > >
> > > > > > > Rich
> > > >
> > > > > From 6413de6d9f785c98e5bc0cf40be947f1169d2fd7 Mon Sep 17 00:00:00 2001
> > > > > From: Elliott Hughes <enh@google.com>
> > > > > Date: Fri, 3 Feb 2023 10:42:55 -0800
> > > > > Subject: [PATCH] risc-v does not have SA_RESTORER.
> > > > >
> > > > > The kernel's include/uapi/asm-generic/signal-defs.h explicitly calls
> > > > > this out as obsolete. New architectures like risc-v do not define it.
> > > > > ---
> > > > >  arch/riscv64/bits/signal.h | 1 -
> > > > >  1 file changed, 1 deletion(-)
> > > > >
> > > > > diff --git a/arch/riscv64/bits/signal.h b/arch/riscv64/bits/signal.h
> > > > > index 287367db..fd6157a3 100644
> > > > > --- a/arch/riscv64/bits/signal.h
> > > > > +++ b/arch/riscv64/bits/signal.h
> > > > > @@ -76,7 +76,6 @@ typedef struct __ucontext
> > > > >  #define SA_RESTART   0x10000000
> > > > >  #define SA_NODEFER   0x40000000
> > > > >  #define SA_RESETHAND 0x80000000
> > > > > -#define SA_RESTORER  0x04000000
> > > > >
> > > > >  #endif
> > > > >
> > > > > --
> > > > > 2.39.1.519.gcb327c4b5f-goog
> > > > >
> > > >
> > > > I don't think this patch works as-is, since musl unconditionally uses
> > > > SA_RESTORER. We probably need to make that conditional on its
> > > > presence, and it looks like there's also a wrong-struct-layout issue
> > > > on archs where it's absent...
> > >
> > > yeah, bionic just uses the kernel uapi headers directly, and they look
> > > like this:
> > >
> > > struct sigaction {
> > >   __sighandler_t sa_handler;
> > >   unsigned long sa_flags;
> > > #ifdef SA_RESTORER
> > >   __sigrestore_t sa_restorer;
> > > #endif
> > >   sigset_t sa_mask;
> > > };
> >
> > OK. It looks like we need to remove the wrong SA_RESTORER for archs
> > that aren't supposed to have it *and* add such an #ifdef. Right now,
> > we're passing bogus sa_mask on these archs... :(
>
> How does the attached look?
>
> Rich