Use 64-bit accesses for 64-bit floating-point general registers with
PTRACE_PEEKUSR, removing the truncation of their upper halves in the
FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context
access"), which inadvertently switched them to using 32-bit accesses.
The PTRACE_POKEUSR side is fine as it's never been broken and continues
using 64-bit accesses.
Cc: <[email protected]> # 3.19+
Fixes: bbd426f542cb ("MIPS: Simplify FP context access")
Signed-off-by: Maciej W. Rozycki <[email protected]>
---
Hi,
Here's another one, spotted in the course of GDB PR gdb/22286 regression
testing with the n64 ABI. Please apply.
Maciej
---
arch/mips/kernel/ptrace.c | 2 +-
arch/mips/kernel/ptrace32.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
linux-mips-ptrace-peekusr-fp64.diff
Index: linux/arch/mips/kernel/ptrace.c
===================================================================
--- linux.orig/arch/mips/kernel/ptrace.c 2018-05-15 17:44:25.000000000 +0100
+++ linux/arch/mips/kernel/ptrace.c 2018-05-16 11:22:00.714605000 +0100
@@ -1070,7 +1070,7 @@ long arch_ptrace(struct task_struct *chi
break;
}
#endif
- tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
+ tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
break;
case PC:
tmp = regs->cp0_epc;
Index: linux/arch/mips/kernel/ptrace32.c
===================================================================
--- linux.orig/arch/mips/kernel/ptrace32.c 2018-05-15 17:45:16.000000000 +0100
+++ linux/arch/mips/kernel/ptrace32.c 2018-05-16 11:22:16.313698000 +0100
@@ -109,7 +109,7 @@ long compat_arch_ptrace(struct task_stru
addr & 1);
break;
}
- tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
+ tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
break;
case PC:
tmp = regs->cp0_epc;
On Wed, May 16, 2018 at 04:39:58PM +0100, Maciej W. Rozycki wrote:
> Use 64-bit accesses for 64-bit floating-point general registers with
> PTRACE_PEEKUSR, removing the truncation of their upper halves in the
> FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context
> access"), which inadvertently switched them to using 32-bit accesses.
>
> The PTRACE_POKEUSR side is fine as it's never been broken and continues
> using 64-bit accesses.
>
> Cc: <[email protected]> # 3.19+
should that be 3.15+?
Cheers
James
On Thu, 24 May 2018, James Hogan wrote:
> > Use 64-bit accesses for 64-bit floating-point general registers with
> > PTRACE_PEEKUSR, removing the truncation of their upper halves in the
> > FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context
> > access"), which inadvertently switched them to using 32-bit accesses.
> >
> > The PTRACE_POKEUSR side is fine as it's never been broken and continues
> > using 64-bit accesses.
> >
> > Cc: <[email protected]> # 3.19+
>
> should that be 3.15+?
Indeed, I must have used the wrong result; thank you for catching this!
Maciej
Hi Maciej,
On Wed, May 16, 2018 at 04:39:58PM +0100, Maciej W. Rozycki wrote:
> Use 64-bit accesses for 64-bit floating-point general registers with
> PTRACE_PEEKUSR, removing the truncation of their upper halves in the
> FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context
> access"), which inadvertently switched them to using 32-bit accesses.
Good catch:
Reviewed-by: Paul Burton <[email protected]>
Thanks,
Paul
> The PTRACE_POKEUSR side is fine as it's never been broken and continues
> using 64-bit accesses.
>
> Cc: <[email protected]> # 3.19+
> Fixes: bbd426f542cb ("MIPS: Simplify FP context access")
> Signed-off-by: Maciej W. Rozycki <[email protected]>
> ---
> Hi,
>
> Here's another one, spotted in the course of GDB PR gdb/22286 regression
> testing with the n64 ABI. Please apply.
>
> Maciej
> ---
> arch/mips/kernel/ptrace.c | 2 +-
> arch/mips/kernel/ptrace32.c | 2 +-
> 2 files changed, 2 insertions(+), 2 deletions(-)
>
> linux-mips-ptrace-peekusr-fp64.diff
> Index: linux/arch/mips/kernel/ptrace.c
> ===================================================================
> --- linux.orig/arch/mips/kernel/ptrace.c 2018-05-15 17:44:25.000000000 +0100
> +++ linux/arch/mips/kernel/ptrace.c 2018-05-16 11:22:00.714605000 +0100
> @@ -1070,7 +1070,7 @@ long arch_ptrace(struct task_struct *chi
> break;
> }
> #endif
> - tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
> + tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
> break;
> case PC:
> tmp = regs->cp0_epc;
> Index: linux/arch/mips/kernel/ptrace32.c
> ===================================================================
> --- linux.orig/arch/mips/kernel/ptrace32.c 2018-05-15 17:45:16.000000000 +0100
> +++ linux/arch/mips/kernel/ptrace32.c 2018-05-16 11:22:16.313698000 +0100
> @@ -109,7 +109,7 @@ long compat_arch_ptrace(struct task_stru
> addr & 1);
> break;
> }
> - tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
> + tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
> break;
> case PC:
> tmp = regs->cp0_epc;
>
On Wed, May 16, 2018 at 04:39:58PM +0100, Maciej W. Rozycki wrote:
> Use 64-bit accesses for 64-bit floating-point general registers with
> PTRACE_PEEKUSR, removing the truncation of their upper halves in the
> FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context
> access"), which inadvertently switched them to using 32-bit accesses.
>
> The PTRACE_POKEUSR side is fine as it's never been broken and continues
> using 64-bit accesses.
>
> Cc: <[email protected]> # 3.19+
> Fixes: bbd426f542cb ("MIPS: Simplify FP context access")
> Signed-off-by: Maciej W. Rozycki <[email protected]>
> ---
> Hi,
>
> Here's another one, spotted in the course of GDB PR gdb/22286 regression
> testing with the n64 ABI. Please apply.
>
> Maciej
Thanks, applied to mips-fixes, hopefully for 4.17 (but if it misses
tomorrows linux-next it may have to wait 'til 4.18).
Cheers
James