[PATCH 1/4] sparc: Disable FPU in interrupt context

From: Alexander Krutwig Update #2270. --- c/src/lib/libbsp/sparc/shared/irq_asm.S | 30 ++ cpukit/score/cpu/sparc/cpu.c | 8 cpukit/score/cpu/sparc/rtems/score/cpu.h | 25 - 3 files changed, 62 insertions(+), 1 deletion(-

[PATCH 2/4] sparc: Avoid new window for FP save/restore

Update #2270. --- cpukit/score/cpu/sparc/cpu_asm.S | 102 ++- 1 file changed, 48 insertions(+), 54 deletions(-) diff --git a/cpukit/score/cpu/sparc/cpu_asm.S b/cpukit/score/cpu/sparc/cpu_asm.S index 92674de..ddb2a13 100644 --- a/cpukit/score/cpu/sparc/cpu_asm.S

[PATCH 4/4] sparc: Optimize non-deferred FP save/restore

Update #2270. --- cpukit/score/cpu/sparc/cpu.c | 16 cpukit/score/cpu/sparc/cpu_asm.S | 44 cpukit/score/cpu/sparc/rtems/score/cpu.h | 22 ++-- 3 files changed, 80 insertions(+), 2 deletions(-) diff --git a/cpukit/score

[PATCH 3/4] sparc: Remove superfluous FP enable

The FP context save/restore makes only sense in the context of FP threads. Update #2270. --- cpukit/score/cpu/sparc/cpu.c | 5 + cpukit/score/cpu/sparc/cpu_asm.S | 24 ++-- 2 files changed, 7 insertions(+), 22 deletions(-) diff --git a/cpukit/score/cpu/sparc/cpu.c b/

RPI2 Cache configuration Was: Re: GSoC 2015: Raspberry Pi 2 Support

On Thu, May 28, 2015 at 10:11 AM, Rohini Kulkarni wrote: > Hi All, > > I have to implement the cache coherency support for Cortex A7. But for A7 > MPCore, unlike for A9, I am not able to find any register description for > the Snoop Control Unit from the TRM. > I need help here on how to proceed.

Re: [PATCH 2/4] sparc: Avoid new window for FP save/restore

Straightforward leaf-optimization, except is there a reason to use jmp %o7 + 8 instead of ret? On Fri, May 29, 2015 at 8:54 AM, Sebastian Huber wrote: > Update #2270. > --- > cpukit/score/cpu/sparc/cpu_asm.S | 102 > ++- > 1 file changed, 48 insertions(+), 54

Re: [PATCH 1/4] sparc: Disable FPU in interrupt context

I am missing something. When an interrupt happens, it disables floating point. Where does fp get re-enabled afterward? On Fri, May 29, 2015 at 8:54 AM, Sebastian Huber wrote: > From: Alexander Krutwig > > Update #2270. > --- > c/src/lib/libbsp/sparc/shared/irq_asm.S | 30 ++

Re: [PATCH 4/4] sparc: Optimize non-deferred FP save/restore

On Fri, May 29, 2015 at 8:54 AM, Sebastian Huber wrote: > Update #2270. > --- > cpukit/score/cpu/sparc/cpu.c | 16 > cpukit/score/cpu/sparc/cpu_asm.S | 44 > > cpukit/score/cpu/sparc/rtems/score/cpu.h | 22 ++-- > 3 fi

Re: [rtems commit] tmtests/tm26: Avoid NULL pointer access

On 5/29/2015 8:37 AM, Sebastian Huber wrote: Module:rtems Branch:master Commit:5611839a7e2e371dd1f327c336c785095f634e55 Changeset: http://git.rtems.org/rtems/commit/?id=5611839a7e2e371dd1f327c336c785095f634e55 Author:Sebastian Huber Date: Fri May 29 14:18:32 2015 +0200

Re: [rtems commit] tmtests/tm26: Avoid NULL pointer access

- Joel Sherrill schrieb: > > > On 5/29/2015 8:37 AM, Sebastian Huber wrote: > > Module:rtems > > Branch:master > > Commit:5611839a7e2e371dd1f327c336c785095f634e55 > > Changeset: > > http://git.rtems.org/rtems/commit/?id=5611839a7e2e371dd1f327c336c785095f634e55 > > > > Author:

Re: [PATCH 2/4] sparc: Avoid new window for FP save/restore

- Gedare Bloom schrieb: > Straightforward leaf-optimization, except is there a reason to use jmp > %o7 + 8 instead of ret? Hm, the GCC generates this instruction, is this the same as "ret"? -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone

Re: [PATCH 1/4] sparc: Disable FPU in interrupt context

- Gedare Bloom schrieb: > I am missing something. When an interrupt happens, it disables > floating point. Where does fp get re-enabled afterward? It gets enabled later when the original PSR is restored. -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, G

Re: [rtems commit] tmtests/tm26: Avoid NULL pointer access

On May 29, 2015 10:45:27 AM CDT, Sebastian Huber wrote: > >- Joel Sherrill schrieb: >> >> >> On 5/29/2015 8:37 AM, Sebastian Huber wrote: >> > Module:rtems >> > Branch:master >> > Commit:5611839a7e2e371dd1f327c336c785095f634e55 >> > Changeset: >http://git.rtems.org/rtems/comm

Re: [PATCH 2/4] sparc: Avoid new window for FP save/restore

On Fri, May 29, 2015 at 10:46 AM, Sebastian Huber wrote: > > - Gedare Bloom schrieb: >> Straightforward leaf-optimization, except is there a reason to use jmp >> %o7 + 8 instead of ret? > > Hm, the GCC generates this instruction, is this the same as "ret"? > Yes I think so. Probably the manua

Re: [PATCH 2/4] sparc: Avoid new window for FP save/restore

On Fri, May 29, 2015 at 11:37 AM, Gedare Bloom wrote: > On Fri, May 29, 2015 at 10:46 AM, Sebastian Huber > wrote: >> >> - Gedare Bloom schrieb: >>> Straightforward leaf-optimization, except is there a reason to use jmp >>> %o7 + 8 instead of ret? >> >> Hm, the GCC generates this instruction

[PATCH] tmtests/tm26: Fix context switch to FP task

It is wrong to restore the floating point context here. The _Context_Switch() ends up in _Thread_Handler() which will call _Thread_Restore_fp(). In _Thread_Do_dispatch() the FP restore is after the context switch. --- testsuites/tmtests/tm26/task1.c | 3 --- 1 file changed, 3 deletions(-) diff