* [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-13 16:38 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-13 16:38 UTC (permalink / raw)
To: Catalin Marinas, Will Deacon, Dave Martin, linux-arm-kernel
Cc: Kristina Martsenko, Richard Henderson, Mark Rutland,
linux-kernel, Rich Felker, Dmitry V . Levin, Ricardo Salveti,
Anisse Astier
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
This fixes an UAPI regression introduced in commit 43d4da2c45b2
("arm64/sve: ptrace and ELF coredump support").
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index d78623acb649..03b6d6f029fc 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
+#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
/*
--
2.19.1
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-13 16:38 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-13 16:38 UTC (permalink / raw)
To: Catalin Marinas, Will Deacon, Dave Martin, linux-arm-kernel
Cc: Mark Rutland, Anisse Astier, Richard Henderson, Rich Felker,
linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
This fixes an UAPI regression introduced in commit 43d4da2c45b2
("arm64/sve: ptrace and ELF coredump support").
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index d78623acb649..03b6d6f029fc 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
+#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
/*
--
2.19.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-13 16:38 ` Anisse Astier
@ 2019-06-13 17:14 ` Dave Martin
-1 siblings, 0 replies; 22+ messages in thread
From: Dave Martin @ 2019-06-13 17:14 UTC (permalink / raw)
To: Anisse Astier
Cc: Catalin Marinas, Will Deacon, linux-arm-kernel, Mark Rutland,
Richard Henderson, Rich Felker, linux-kernel, Kristina Martsenko,
Dmitry V . Levin, Ricardo Salveti
On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> This fixes an UAPI regression introduced in commit 43d4da2c45b2
> ("arm64/sve: ptrace and ELF coredump support").
>
> Cc: stable@vger.kernel.org
Consider adding a Fixes: tag.
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index d78623acb649..03b6d6f029fc 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -65,8 +65,6 @@
>
> #ifndef __ASSEMBLY__
>
> -#include <linux/prctl.h>
> -
> /*
> * User structures for general purpose, floating point and debug registers.
> */
> @@ -113,10 +111,10 @@ struct user_sve_header {
>
> /*
> * Common SVE_PT_* flags:
> - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> + * These must be kept in sync with prctl interface in <linux/prctl.h>
Ack
> */
> -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
Makes sense, but...
Since sve_context.h was already introduced to solve a closely related
problem, I wonder whether we can provide shadow definitions there,
similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
derived. Although it's a slight abuse of that header, I think that
would be my preferred approach.
Otherwise, at least make the required relationship between ptrace.h and
prctl.h constants a bit more obvious, say,
#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
etc.
Cheers
---Dave
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-13 17:14 ` Dave Martin
0 siblings, 0 replies; 22+ messages in thread
From: Dave Martin @ 2019-06-13 17:14 UTC (permalink / raw)
To: Anisse Astier
Cc: Mark Rutland, Catalin Marinas, Richard Henderson, linux-kernel,
Will Deacon, Dmitry V . Levin, Ricardo Salveti,
Kristina Martsenko, Rich Felker, linux-arm-kernel
On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> This fixes an UAPI regression introduced in commit 43d4da2c45b2
> ("arm64/sve: ptrace and ELF coredump support").
>
> Cc: stable@vger.kernel.org
Consider adding a Fixes: tag.
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index d78623acb649..03b6d6f029fc 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -65,8 +65,6 @@
>
> #ifndef __ASSEMBLY__
>
> -#include <linux/prctl.h>
> -
> /*
> * User structures for general purpose, floating point and debug registers.
> */
> @@ -113,10 +111,10 @@ struct user_sve_header {
>
> /*
> * Common SVE_PT_* flags:
> - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> + * These must be kept in sync with prctl interface in <linux/prctl.h>
Ack
> */
> -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
Makes sense, but...
Since sve_context.h was already introduced to solve a closely related
problem, I wonder whether we can provide shadow definitions there,
similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
derived. Although it's a slight abuse of that header, I think that
would be my preferred approach.
Otherwise, at least make the required relationship between ptrace.h and
prctl.h constants a bit more obvious, say,
#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
etc.
Cheers
---Dave
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-13 17:14 ` Dave Martin
@ 2019-06-14 10:23 ` Will Deacon
-1 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-14 10:23 UTC (permalink / raw)
To: Dave Martin
Cc: Anisse Astier, Catalin Marinas, linux-arm-kernel, Mark Rutland,
Richard Henderson, Rich Felker, linux-kernel, Kristina Martsenko,
Dmitry V . Levin, Ricardo Salveti
On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > */
> > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
>
> Makes sense, but...
>
> Since sve_context.h was already introduced to solve a closely related
> problem, I wonder whether we can provide shadow definitions there,
> similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> derived. Although it's a slight abuse of that header, I think that
> would be my preferred approach.
Yes, that sounds better to me as well. Please send a v2!
Will
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-14 10:23 ` Will Deacon
0 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-14 10:23 UTC (permalink / raw)
To: Dave Martin
Cc: Mark Rutland, Anisse Astier, Catalin Marinas, Rich Felker,
linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Richard Henderson, linux-arm-kernel
On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > */
> > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
>
> Makes sense, but...
>
> Since sve_context.h was already introduced to solve a closely related
> problem, I wonder whether we can provide shadow definitions there,
> similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> derived. Although it's a slight abuse of that header, I think that
> would be my preferred approach.
Yes, that sounds better to me as well. Please send a v2!
Will
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-13 17:14 ` Dave Martin
@ 2019-06-14 11:22 ` Anisse Astier
-1 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-14 11:22 UTC (permalink / raw)
To: Dave Martin, Will Deacon
Cc: Catalin Marinas, Will Deacon, linux-arm-kernel, Mark Rutland,
Richard Henderson, Rich Felker, linux-kernel, Kristina Martsenko,
Dmitry V . Levin, Ricardo Salveti
Hi Dave,
Thanks for taking the time to review this patch,
On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > Otherwise this will create userspace build issues for any program
> > (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> > <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
> >
> > error: redefinition of 'struct prctl_mm_map'
> > struct prctl_mm_map {
> >
> > See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> > for a public example of people working around this issue.
> >
> > This fixes an UAPI regression introduced in commit 43d4da2c45b2
> > ("arm64/sve: ptrace and ELF coredump support").
> >
> > Cc: stable@vger.kernel.org
>
> Consider adding a Fixes: tag.
Will do in v2.
>
> > Signed-off-by: Anisse Astier <aastier@freebox.fr>
> > ---
> > arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> > index d78623acb649..03b6d6f029fc 100644
> > --- a/arch/arm64/include/uapi/asm/ptrace.h
> > +++ b/arch/arm64/include/uapi/asm/ptrace.h
> > @@ -65,8 +65,6 @@
> >
> > #ifndef __ASSEMBLY__
> >
> > -#include <linux/prctl.h>
> > -
> > /*
> > * User structures for general purpose, floating point and debug registers.
> > */
> > @@ -113,10 +111,10 @@ struct user_sve_header {
> >
> > /*
> > * Common SVE_PT_* flags:
> > - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> > + * These must be kept in sync with prctl interface in <linux/prctl.h>
>
> Ack
>
> > */
> > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
>
> Makes sense, but...
>
> Since sve_context.h was already introduced to solve a closely related
> problem, I wonder whether we can provide shadow definitions there,
> similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> derived. Although it's a slight abuse of that header, I think that
> would be my preferred approach.
Yes I saw this, and I considered doing something similar. But, those
defines are in uapi/linux/prctl.h, which does not include any asm/*.h
header. This would have then required adding a full infrastructure for
asm/prctl.h (that could then include sve_context.h for example), which
does not exist yet, instead of copying these two values.
Since this is part of the kernel-userspace ABI, I don't see this values
changing anytime soon, which is why I thought copying them shouldn't be
a big issue.
A simple solution would be to to include sve_context.h or a third
header, maybe linux/prctl_arm64_sve.h (with only these two/five
defines), in linux/prctl.h, and reuse it in uapi/asm/ptrace.h; but this
would break the self-contained nature of linux/prctl.h.
>
> Otherwise, at least make the required relationship between ptrace.h and
> prctl.h constants a bit more obvious, say,
>
> #define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
This one is much simpler and closer to what I had in mind with this
patch.
Will, what do you think of this second approach Dave proposed ?
Regards,
Anisse
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-14 11:22 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-14 11:22 UTC (permalink / raw)
To: Dave Martin, Will Deacon
Cc: Mark Rutland, Catalin Marinas, Richard Henderson, linux-kernel,
Will Deacon, Dmitry V . Levin, Ricardo Salveti,
Kristina Martsenko, Rich Felker, linux-arm-kernel
Hi Dave,
Thanks for taking the time to review this patch,
On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > Otherwise this will create userspace build issues for any program
> > (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> > <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
> >
> > error: redefinition of 'struct prctl_mm_map'
> > struct prctl_mm_map {
> >
> > See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> > for a public example of people working around this issue.
> >
> > This fixes an UAPI regression introduced in commit 43d4da2c45b2
> > ("arm64/sve: ptrace and ELF coredump support").
> >
> > Cc: stable@vger.kernel.org
>
> Consider adding a Fixes: tag.
Will do in v2.
>
> > Signed-off-by: Anisse Astier <aastier@freebox.fr>
> > ---
> > arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> > index d78623acb649..03b6d6f029fc 100644
> > --- a/arch/arm64/include/uapi/asm/ptrace.h
> > +++ b/arch/arm64/include/uapi/asm/ptrace.h
> > @@ -65,8 +65,6 @@
> >
> > #ifndef __ASSEMBLY__
> >
> > -#include <linux/prctl.h>
> > -
> > /*
> > * User structures for general purpose, floating point and debug registers.
> > */
> > @@ -113,10 +111,10 @@ struct user_sve_header {
> >
> > /*
> > * Common SVE_PT_* flags:
> > - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> > + * These must be kept in sync with prctl interface in <linux/prctl.h>
>
> Ack
>
> > */
> > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
>
> Makes sense, but...
>
> Since sve_context.h was already introduced to solve a closely related
> problem, I wonder whether we can provide shadow definitions there,
> similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> derived. Although it's a slight abuse of that header, I think that
> would be my preferred approach.
Yes I saw this, and I considered doing something similar. But, those
defines are in uapi/linux/prctl.h, which does not include any asm/*.h
header. This would have then required adding a full infrastructure for
asm/prctl.h (that could then include sve_context.h for example), which
does not exist yet, instead of copying these two values.
Since this is part of the kernel-userspace ABI, I don't see this values
changing anytime soon, which is why I thought copying them shouldn't be
a big issue.
A simple solution would be to to include sve_context.h or a third
header, maybe linux/prctl_arm64_sve.h (with only these two/five
defines), in linux/prctl.h, and reuse it in uapi/asm/ptrace.h; but this
would break the self-contained nature of linux/prctl.h.
>
> Otherwise, at least make the required relationship between ptrace.h and
> prctl.h constants a bit more obvious, say,
>
> #define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
This one is much simpler and closer to what I had in mind with this
patch.
Will, what do you think of this second approach Dave proposed ?
Regards,
Anisse
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-14 11:22 ` Anisse Astier
@ 2019-06-14 14:22 ` Will Deacon
-1 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-14 14:22 UTC (permalink / raw)
To: Anisse Astier
Cc: Dave Martin, Catalin Marinas, linux-arm-kernel, Mark Rutland,
Richard Henderson, Rich Felker, linux-kernel, Kristina Martsenko,
Dmitry V . Levin, Ricardo Salveti
Hi Anisse, Dave,
On Fri, Jun 14, 2019 at 01:22:22PM +0200, Anisse Astier wrote:
> On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> > On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
> >
> > Makes sense, but...
> >
> > Since sve_context.h was already introduced to solve a closely related
> > problem, I wonder whether we can provide shadow definitions there,
> > similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> > derived. Although it's a slight abuse of that header, I think that
> > would be my preferred approach.
>
> Yes I saw this, and I considered doing something similar. But, those
> defines are in uapi/linux/prctl.h, which does not include any asm/*.h
> header. This would have then required adding a full infrastructure for
> asm/prctl.h (that could then include sve_context.h for example), which
> does not exist yet, instead of copying these two values.
x86 appears to have an asm/prctl.h implementation, but it's not included
by anybody so I guess that doesn't really help us here.
> Since this is part of the kernel-userspace ABI, I don't see this values
> changing anytime soon, which is why I thought copying them shouldn't be
> a big issue.
Certainly not a big issue, just that the harder we make this to change
the better.
> A simple solution would be to to include sve_context.h or a third
> header, maybe linux/prctl_arm64_sve.h (with only these two/five
> defines), in linux/prctl.h, and reuse it in uapi/asm/ptrace.h; but this
> would break the self-contained nature of linux/prctl.h.
> >
> > Otherwise, at least make the required relationship between ptrace.h and
> > prctl.h constants a bit more obvious, say,
> >
> > #define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
>
> This one is much simpler and closer to what I had in mind with this
> patch.
>
> Will, what do you think of this second approach Dave proposed ?
Duplication is grotty, but it does the job so I'm ok with it. I don't have
any better ideas.
Thanks,
Will
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-14 14:22 ` Will Deacon
0 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-14 14:22 UTC (permalink / raw)
To: Anisse Astier
Cc: Mark Rutland, Catalin Marinas, Rich Felker, linux-kernel,
Kristina Martsenko, Dmitry V . Levin, Ricardo Salveti,
Richard Henderson, Dave Martin, linux-arm-kernel
Hi Anisse, Dave,
On Fri, Jun 14, 2019 at 01:22:22PM +0200, Anisse Astier wrote:
> On Thu, Jun 13, 2019 at 06:14:44PM +0100, Dave Martin wrote:
> > On Thu, Jun 13, 2019 at 06:38:01PM +0200, Anisse Astier wrote:
> > > -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> > > -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> > > +#define SVE_PT_VL_INHERIT (1 << 1) /* PR_SVE_VL_INHERIT */
> > > +#define SVE_PT_VL_ONEXEC (1 << 2) /* PR_SVE_SET_VL_ONEXEC */
> >
> > Makes sense, but...
> >
> > Since sve_context.h was already introduced to solve a closely related
> > problem, I wonder whether we can provide shadow definitions there,
> > similarly to way the arm64/include/uapi/asm/ptrace.h definitions are
> > derived. Although it's a slight abuse of that header, I think that
> > would be my preferred approach.
>
> Yes I saw this, and I considered doing something similar. But, those
> defines are in uapi/linux/prctl.h, which does not include any asm/*.h
> header. This would have then required adding a full infrastructure for
> asm/prctl.h (that could then include sve_context.h for example), which
> does not exist yet, instead of copying these two values.
x86 appears to have an asm/prctl.h implementation, but it's not included
by anybody so I guess that doesn't really help us here.
> Since this is part of the kernel-userspace ABI, I don't see this values
> changing anytime soon, which is why I thought copying them shouldn't be
> a big issue.
Certainly not a big issue, just that the harder we make this to change
the better.
> A simple solution would be to to include sve_context.h or a third
> header, maybe linux/prctl_arm64_sve.h (with only these two/five
> defines), in linux/prctl.h, and reuse it in uapi/asm/ptrace.h; but this
> would break the self-contained nature of linux/prctl.h.
> >
> > Otherwise, at least make the required relationship between ptrace.h and
> > prctl.h constants a bit more obvious, say,
> >
> > #define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_SET_VL_INHERIT */ >> 16)
>
> This one is much simpler and closer to what I had in mind with this
> patch.
>
> Will, what do you think of this second approach Dave proposed ?
Duplication is grotty, but it does the job so I'm ok with it. I don't have
any better ideas.
Thanks,
Will
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-13 16:38 ` Anisse Astier
@ 2019-06-14 16:46 ` Anisse Astier
-1 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-14 16:46 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Catalin Marinas, Kristina Martsenko, Richard Henderson,
Mark Rutland, linux-kernel, Rich Felker, Dmitry V . Levin,
Ricardo Salveti, Anisse Astier
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
Copying the defines is a bit imperfect here, but better than creating a
whole other header for just two defines that would never change, as part
of the kernel ABI.
Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
Changes since v1:
- made a bit more explicit that we copied defined symbols, in commit
and code.
- Use Fixes: tag in commit message
Thanks to Dave Martin and Will Deacon for the review.
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index d78623acb649..438759e7e8a7 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
+#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
/*
--
2.19.1
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-14 16:46 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-14 16:46 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Mark Rutland, Anisse Astier, Catalin Marinas, Rich Felker,
linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Richard Henderson
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
Copying the defines is a bit imperfect here, but better than creating a
whole other header for just two defines that would never change, as part
of the kernel ABI.
Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
Changes since v1:
- made a bit more explicit that we copied defined symbols, in commit
and code.
- Use Fixes: tag in commit message
Thanks to Dave Martin and Will Deacon for the review.
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index d78623acb649..438759e7e8a7 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
+#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
/*
--
2.19.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
[not found] ` <20190615221601.61FFC2184D@mail.kernel.org>
@ 2019-06-17 8:45 ` Anisse Astier
2019-06-17 13:22 ` Anisse Astier
0 siblings, 1 reply; 22+ messages in thread
From: Anisse Astier @ 2019-06-17 8:45 UTC (permalink / raw)
To: Sasha Levin; +Cc: Will Deacon, Catalin Marinas, stable, Dave Martin
On Sat, Jun 15, 2019 at 10:16:00PM +0000, Sasha Levin wrote:
> Hi,
>
> [This is an automated email]
>
> This commit has been processed because it contains a "Fixes:" tag,
> fixing commit: 43d4da2c45b2 arm64/sve: ptrace and ELF coredump support.
>
> The bot has tested the following trees: v5.1.9, v4.19.50.
>
> v5.1.9: Build failed! Errors:
> arch/arm64/kernel/ssbd.c:44:8: error: ‘PR_SPEC_ENABLE’ undeclared (first use in this function); did you mean ‘NR_PAGETABLE’?
> arch/arm64/kernel/ssbd.c:46:8: error: ‘PR_SPEC_DISABLE’ undeclared (first use in this function); did you mean ‘PFA_SPEC_IB_DISABLE’?
> arch/arm64/kernel/ssbd.c:47:8: error: ‘PR_SPEC_FORCE_DISABLE’ undeclared (first use in this function); did you mean ‘PFA_SPEC_IB_FORCE_DISABLE’?
> arch/arm64/kernel/ssbd.c:93:7: error: ‘PR_SPEC_STORE_BYPASS’ undeclared (first use in this function)
> arch/arm64/kernel/ssbd.c:106:10: error: ‘PR_SPEC_DISABLE’ undeclared (first use in this function); did you mean ‘PFA_SPEC_IB_DISABLE’?
> arch/arm64/kernel/ssbd.c:109:11: error: ‘PR_SPEC_PRCTL’ undeclared (first use in this function); did you mean ‘PF_SECURITY’?
> arch/arm64/kernel/ssbd.c:109:27: error: ‘PR_SPEC_FORCE_DISABLE’ undeclared (first use in this function); did you mean ‘PFA_SPEC_IB_FORCE_DISABLE’?
> arch/arm64/kernel/ssbd.c:112:26: error: ‘PR_SPEC_ENABLE’ undeclared (first use in this function); did you mean ‘NR_PAGETABLE’?
> arch/arm64/kernel/ssbd.c:116:10: error: ‘PR_SPEC_NOT_AFFECTED’ undeclared (first use in this function)
> arch/arm64/kernel/ssbd.c:123:7: error: ‘PR_SPEC_STORE_BYPASS’ undeclared (first use in this function)
>
> v4.19.50: Build OK!
>
> How should we proceed with this patch?
>
Looks like I was building without ARM64_SSBD. Although even enabling it
on Linus master does not trigger this build failure (I can reproduce on
5.1.y though).
I'll prepare a v3 anyway to make this linux/prctl.h dependency of ssbd.c
more explicit for 5.1.y backport.
Regards,
Anisse
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH v3 1/2] arm64: ssbd: explicitly depend on <linux/prctl.h>
2019-06-17 8:45 ` Anisse Astier
@ 2019-06-17 13:22 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-17 13:22 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Catalin Marinas, Kristina Martsenko, Richard Henderson,
Mark Rutland, linux-kernel, Rich Felker, Dmitry V . Levin,
Ricardo Salveti, Sasha Levin, Anisse Astier
Fix ssbd.c which depends implicitly on asm/ptrace.h including
linux/prctl.h (through for example linux/compat.h, then linux/time.h,
linux/seqlock.h, linux/spinlock.h and linux/irqflags.h), and uses
PR_SPEC* defines.
This is an issue since we'll remove the include in the next commit.
Fixes: 9cdc0108baa8 ("arm64: ssbd: Add prctl interface for per-thread mitigation")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
Contrary to what I said in the previous email, I can reproduce the build
error on Linus' master when ARM64_SSBD is enabled.
Changes since v2:
- fix build when ARM64_SSBD is enabled with additionnal patch
Changes since v1:
- made a bit more explicit that we copied defined symbols, in commit
and code.
- Use Fixes: tag in commit message
Thanks to Dave Martin and Will Deacon for the review, and Sasha Levin
for the auto-build bot.
---
arch/arm64/kernel/ssbd.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/arm64/kernel/ssbd.c b/arch/arm64/kernel/ssbd.c
index 885f13e58708..52cfc6148355 100644
--- a/arch/arm64/kernel/ssbd.c
+++ b/arch/arm64/kernel/ssbd.c
@@ -5,6 +5,7 @@
#include <linux/compat.h>
#include <linux/errno.h>
+#include <linux/prctl.h>
#include <linux/sched.h>
#include <linux/sched/task_stack.h>
#include <linux/thread_info.h>
--
2.19.1
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v3 1/2] arm64: ssbd: explicitly depend on <linux/prctl.h>
@ 2019-06-17 13:22 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-17 13:22 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Mark Rutland, Sasha Levin, Anisse Astier, Catalin Marinas,
Rich Felker, linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Richard Henderson
Fix ssbd.c which depends implicitly on asm/ptrace.h including
linux/prctl.h (through for example linux/compat.h, then linux/time.h,
linux/seqlock.h, linux/spinlock.h and linux/irqflags.h), and uses
PR_SPEC* defines.
This is an issue since we'll remove the include in the next commit.
Fixes: 9cdc0108baa8 ("arm64: ssbd: Add prctl interface for per-thread mitigation")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
Contrary to what I said in the previous email, I can reproduce the build
error on Linus' master when ARM64_SSBD is enabled.
Changes since v2:
- fix build when ARM64_SSBD is enabled with additionnal patch
Changes since v1:
- made a bit more explicit that we copied defined symbols, in commit
and code.
- Use Fixes: tag in commit message
Thanks to Dave Martin and Will Deacon for the review, and Sasha Levin
for the auto-build bot.
---
arch/arm64/kernel/ssbd.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/arm64/kernel/ssbd.c b/arch/arm64/kernel/ssbd.c
index 885f13e58708..52cfc6148355 100644
--- a/arch/arm64/kernel/ssbd.c
+++ b/arch/arm64/kernel/ssbd.c
@@ -5,6 +5,7 @@
#include <linux/compat.h>
#include <linux/errno.h>
+#include <linux/prctl.h>
#include <linux/sched.h>
#include <linux/sched/task_stack.h>
#include <linux/thread_info.h>
--
2.19.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v3 2/2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-17 13:22 ` Anisse Astier
@ 2019-06-17 13:22 ` Anisse Astier
-1 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-17 13:22 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Catalin Marinas, Kristina Martsenko, Richard Henderson,
Mark Rutland, linux-kernel, Rich Felker, Dmitry V . Levin,
Ricardo Salveti, Sasha Levin, Anisse Astier
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
Copying the defines is a bit imperfect here, but better than creating a
whole other header for just two defines that would never change, as part
of the kernel ABI.
This patch depends on patch "arm64: ssbd: explicitly depend on
<linux/prctl.h>" for kernel >= 4.18
Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index 97c53203150b..e932284993d4 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
+#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
/*
--
2.19.1
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH v3 2/2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-17 13:22 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-17 13:22 UTC (permalink / raw)
To: Will Deacon, Dave Martin, linux-arm-kernel
Cc: Mark Rutland, Sasha Levin, Anisse Astier, Catalin Marinas,
Rich Felker, linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Richard Henderson
Otherwise this will create userspace build issues for any program
(strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
<linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
error: redefinition of 'struct prctl_mm_map'
struct prctl_mm_map {
See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
for a public example of people working around this issue.
Copying the defines is a bit imperfect here, but better than creating a
whole other header for just two defines that would never change, as part
of the kernel ABI.
This patch depends on patch "arm64: ssbd: explicitly depend on
<linux/prctl.h>" for kernel >= 4.18
Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
Cc: stable@vger.kernel.org
Signed-off-by: Anisse Astier <aastier@freebox.fr>
---
arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
index 97c53203150b..e932284993d4 100644
--- a/arch/arm64/include/uapi/asm/ptrace.h
+++ b/arch/arm64/include/uapi/asm/ptrace.h
@@ -65,8 +65,6 @@
#ifndef __ASSEMBLY__
-#include <linux/prctl.h>
-
/*
* User structures for general purpose, floating point and debug registers.
*/
@@ -113,10 +111,10 @@ struct user_sve_header {
/*
* Common SVE_PT_* flags:
- * These must be kept in sync with prctl interface in <linux/ptrace.h>
+ * These must be kept in sync with prctl interface in <linux/prctl.h>
*/
-#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
-#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
+#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
+#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
/*
--
2.19.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-14 16:46 ` Anisse Astier
@ 2019-06-18 13:40 ` Dave Martin
-1 siblings, 0 replies; 22+ messages in thread
From: Dave Martin @ 2019-06-18 13:40 UTC (permalink / raw)
To: Anisse Astier
Cc: Will Deacon, linux-arm-kernel, Mark Rutland, Catalin Marinas,
Rich Felker, linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Richard Henderson
On Fri, Jun 14, 2019 at 06:46:00PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> Copying the defines is a bit imperfect here, but better than creating a
> whole other header for just two defines that would never change, as part
> of the kernel ABI.
>
> Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> Changes since v1:
> - made a bit more explicit that we copied defined symbols, in commit
> and code.
> - Use Fixes: tag in commit message
>
> Thanks to Dave Martin and Will Deacon for the review.
>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index d78623acb649..438759e7e8a7 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -65,8 +65,6 @@
>
> #ifndef __ASSEMBLY__
>
> -#include <linux/prctl.h>
> -
> /*
> * User structures for general purpose, floating point and debug registers.
> */
> @@ -113,10 +111,10 @@ struct user_sve_header {
>
> /*
> * Common SVE_PT_* flags:
> - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> + * These must be kept in sync with prctl interface in <linux/prctl.h>
> */
> -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> +#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
> +#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
FWIW,
Acked-by: Dave Martin <Dave.Martin@arm.com>
Ideally we would create a common <uapi/asm/prctl.h> that arm64 could
specialise as appropriate, but since x86 already has a similar header
shadowing this, it may be more trouble than it's worth to generalise it.
Cheers
---Dave
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-18 13:40 ` Dave Martin
0 siblings, 0 replies; 22+ messages in thread
From: Dave Martin @ 2019-06-18 13:40 UTC (permalink / raw)
To: Anisse Astier
Cc: Mark Rutland, Catalin Marinas, Rich Felker, linux-kernel,
Will Deacon, Dmitry V . Levin, Ricardo Salveti,
Kristina Martsenko, Richard Henderson, linux-arm-kernel
On Fri, Jun 14, 2019 at 06:46:00PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> Copying the defines is a bit imperfect here, but better than creating a
> whole other header for just two defines that would never change, as part
> of the kernel ABI.
>
> Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> Changes since v1:
> - made a bit more explicit that we copied defined symbols, in commit
> and code.
> - Use Fixes: tag in commit message
>
> Thanks to Dave Martin and Will Deacon for the review.
>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index d78623acb649..438759e7e8a7 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -65,8 +65,6 @@
>
> #ifndef __ASSEMBLY__
>
> -#include <linux/prctl.h>
> -
> /*
> * User structures for general purpose, floating point and debug registers.
> */
> @@ -113,10 +111,10 @@ struct user_sve_header {
>
> /*
> * Common SVE_PT_* flags:
> - * These must be kept in sync with prctl interface in <linux/ptrace.h>
> + * These must be kept in sync with prctl interface in <linux/prctl.h>
> */
> -#define SVE_PT_VL_INHERIT (PR_SVE_VL_INHERIT >> 16)
> -#define SVE_PT_VL_ONEXEC (PR_SVE_SET_VL_ONEXEC >> 16)
> +#define SVE_PT_VL_INHERIT ((1 << 17) /* PR_SVE_VL_INHERIT */ >> 16)
> +#define SVE_PT_VL_ONEXEC ((1 << 18) /* PR_SVE_SET_VL_ONEXEC */ >> 16)
FWIW,
Acked-by: Dave Martin <Dave.Martin@arm.com>
Ideally we would create a common <uapi/asm/prctl.h> that arm64 could
specialise as appropriate, but since x86 already has a similar header
shadowing this, it may be more trouble than it's worth to generalise it.
Cheers
---Dave
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v3 2/2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
2019-06-17 13:22 ` Anisse Astier
@ 2019-06-18 13:50 ` Will Deacon
-1 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-18 13:50 UTC (permalink / raw)
To: Anisse Astier
Cc: Dave Martin, linux-arm-kernel, Catalin Marinas,
Kristina Martsenko, Richard Henderson, Mark Rutland,
linux-kernel, Rich Felker, Dmitry V . Levin, Ricardo Salveti,
Sasha Levin
On Mon, Jun 17, 2019 at 03:22:22PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> Copying the defines is a bit imperfect here, but better than creating a
> whole other header for just two defines that would never change, as part
> of the kernel ABI.
>
> This patch depends on patch "arm64: ssbd: explicitly depend on
> <linux/prctl.h>" for kernel >= 4.18
>
> Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
I've pushed these two out to the arm64 fixes branch [1], with Dave's Ack
on this one. Note that I reworked the commit messages a bit to explain
better what's going on.
Cheers,
Will
[1] https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=for-next/fixes
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v3 2/2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h>
@ 2019-06-18 13:50 ` Will Deacon
0 siblings, 0 replies; 22+ messages in thread
From: Will Deacon @ 2019-06-18 13:50 UTC (permalink / raw)
To: Anisse Astier
Cc: Mark Rutland, Sasha Levin, Catalin Marinas, Richard Henderson,
linux-kernel, Kristina Martsenko, Dmitry V . Levin,
Ricardo Salveti, Rich Felker, Dave Martin, linux-arm-kernel
On Mon, Jun 17, 2019 at 03:22:22PM +0200, Anisse Astier wrote:
> Otherwise this will create userspace build issues for any program
> (strace, qemu) that includes both <sys/prctl.h> (with musl libc) and
> <linux/ptrace.h> (which then includes <asm/ptrace.h>), like this:
>
> error: redefinition of 'struct prctl_mm_map'
> struct prctl_mm_map {
>
> See https://github.com/foundriesio/meta-lmp/commit/6d4a106e191b5d79c41b9ac78fd321316d3013c0
> for a public example of people working around this issue.
>
> Copying the defines is a bit imperfect here, but better than creating a
> whole other header for just two defines that would never change, as part
> of the kernel ABI.
>
> This patch depends on patch "arm64: ssbd: explicitly depend on
> <linux/prctl.h>" for kernel >= 4.18
>
> Fixes: 43d4da2c45b2 ("arm64/sve: ptrace and ELF coredump support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Anisse Astier <aastier@freebox.fr>
> ---
> arch/arm64/include/uapi/asm/ptrace.h | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
I've pushed these two out to the arm64 fixes branch [1], with Dave's Ack
on this one. Note that I reworked the commit messages a bit to explain
better what's going on.
Cheers,
Will
[1] https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=for-next/fixes
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH v3 1/2] arm64: ssbd: explicitly depend on <linux/prctl.h>
[not found] ` <20190618173036.5DC0D2084D@mail.kernel.org>
@ 2019-06-19 7:52 ` Anisse Astier
0 siblings, 0 replies; 22+ messages in thread
From: Anisse Astier @ 2019-06-19 7:52 UTC (permalink / raw)
To: Sasha Levin; +Cc: Will Deacon, Catalin Marinas, stable
Hi Sasha,
On Tue, Jun 18, 2019 at 05:30:35PM +0000, Sasha Levin wrote:
> Hi,
>
> [This is an automated email]
>
> This commit has been processed because it contains a "Fixes:" tag,
> fixing commit: 9cdc0108baa8 arm64: ssbd: Add prctl interface for per-thread mitigation.
>
> The bot has tested the following trees: v5.1.10, v4.19.51, v4.14.126, v4.9.181.
>
> v5.1.10: Build OK!
> v4.19.51: Build OK!
> v4.14.126: Failed to apply! Possible dependencies:
> Unable to calculate
>
> v4.9.181: Failed to apply! Possible dependencies:
> Unable to calculate
I just looked at the 4.14 and 4.9 branches, and their backport of ssbd
already includes this patch, so that's OK.
Regards,
Anisse
^ permalink raw reply [flat|nested] 22+ messages in thread
end of thread, other threads:[~2019-06-19 7:53 UTC | newest]
Thread overview: 22+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-06-13 16:38 [PATCH] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h> Anisse Astier
2019-06-13 16:38 ` Anisse Astier
2019-06-13 17:14 ` Dave Martin
2019-06-13 17:14 ` Dave Martin
2019-06-14 10:23 ` Will Deacon
2019-06-14 10:23 ` Will Deacon
2019-06-14 11:22 ` Anisse Astier
2019-06-14 11:22 ` Anisse Astier
2019-06-14 14:22 ` Will Deacon
2019-06-14 14:22 ` Will Deacon
2019-06-14 16:46 ` [PATCH v2] " Anisse Astier
2019-06-14 16:46 ` Anisse Astier
[not found] ` <20190615221601.61FFC2184D@mail.kernel.org>
2019-06-17 8:45 ` Anisse Astier
2019-06-17 13:22 ` [PATCH v3 1/2] arm64: ssbd: explicitly depend on <linux/prctl.h> Anisse Astier
2019-06-17 13:22 ` Anisse Astier
2019-06-17 13:22 ` [PATCH v3 2/2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h> Anisse Astier
2019-06-17 13:22 ` Anisse Astier
2019-06-18 13:50 ` Will Deacon
2019-06-18 13:50 ` Will Deacon
[not found] ` <20190618173036.5DC0D2084D@mail.kernel.org>
2019-06-19 7:52 ` [PATCH v3 1/2] arm64: ssbd: explicitly depend on <linux/prctl.h> Anisse Astier
2019-06-18 13:40 ` [PATCH v2] arm64/sve: <uapi/asm/ptrace.h> should not depend on <uapi/linux/prctl.h> Dave Martin
2019-06-18 13:40 ` Dave Martin
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.