[PATCH 2/2] powerpc: POWER7 optimised copy_page using VMX and enhanced prefetch

Implement a POWER7 optimised copy_page using VMX and enhanced prefetch instructions. We use enhanced prefetch hints to prefetch both the load and store side. We copy a cacheline at a time and fall back to regular loads and stores if we are unable to use VMX (eg we are in an interrupt). The follow

[PATCH 1/2] powerpc: Rename copyuser_power7_vmx.c to vmx-helper.c

Subsequent patches will add more VMX library functions and it makes sense to keep all the c-code helper functions in the one file. Signed-off-by: Anton Blanchard --- Index: linux-build/arch/powerpc/lib/Makefile === --- linux-build.

Re: ppc/sata-fsl: orphan config value: CONFIG_MPC8315_DS

On 05/29/2012 05:07 PM, Anthony Foiani wrote: > Scott Wood writes: > >> CONFIG_MPC831x_RDB doesn't mean that you're running on such a board, >> only that the kernel supports those boards. It should be a runtime >> test. > > Point taken. > > If that SATA check is CPU/SOC-based, then it should b

[PATCH] powerpc: Clear RI and EE at the same time in system call exit

Hi Scott, > > syscall_exit_work: > > + mtmsrd r10,1 /* Restore RI */ > > That mtmsrd needs an #ifdef CONFIG_PPC_BOOK3S Thanks! Anton -- In system call exit we currently clear RI and EE separately. An mtmsrd is a slow operation and we can save cycles by doing it all in one go. Th

Re: ppc/sata-fsl: orphan config value: CONFIG_MPC8315_DS

Scott Wood writes: > CONFIG_MPC831x_RDB doesn't mean that you're running on such a board, > only that the kernel supports those boards. It should be a runtime > test. Point taken. If that SATA check is CPU/SOC-based, then it should be easy enough to test. The cpuinfo for my board is: # cat

Re: MSR loses DE bit in 3.3, BDI2000 cant handle breakpoints?

Dear Bob, In message <4fc511c1.4050...@mindchasers.com> you wrote: > > I debug using Freescale CodeWarrior and a USB tap, which also rely on > MSR[DE] being set. I develop from the mainline & have a patch set that > I just recently re-tweaked to support kernel debugging. > > If you want, I'll

Re: MSR loses DE bit in 3.3, BDI2000 cant handle breakpoints?

On 05/29/2012 08:00 AM, Joakim Tjernlund wrote: I cannot make simple break points using BDI2000 work in 3.3, abatro suggests that it depends on MSR[DE] is cleared by the kernel. With the emulator I can see that MSR[DE] is off quite often by just stopping at random times and looking at MSR so i

Re: MSR loses DE bit in 3.3, BDI2000 cant handle breakpoints?

Bob Cochran wrote on 2012/05/29 20:13:21: > > On 05/29/2012 08:00 AM, Joakim Tjernlund wrote: > > > > I cannot make simple break points using BDI2000 work in 3.3, abatro > > suggests that it > > depends on MSR[DE] is cleared by the kernel. With the emulator I can see > > that > > MSR[DE] is off

Re: [PATCH] powerpc: Clear RI and EE at the same time in system call exit

On 05/29/2012 06:20 AM, Anton Blanchard wrote: > > In system call exit we currently clear RI and EE separately. An > mtmsrd is a slow operation and we can save cycles by doing it all > in one go. > > This does complicate things a bit - we have to be careful to restore > RI if we branch out before

Re: ppc/sata-fsl: orphan config value: CONFIG_MPC8315_DS

On 05/26/2012 01:53 AM, Anthony Foiani wrote: > Li Yang-R58472 writes: > >> Thanks for bringing [CONFIG_MPC8315_DS] up again. Looks like we do >> have a problem here. > > My impression is that the simplest fix is Adrian's patch, which simply > keys off CONFIG_MPC831x_RDB. It's not very satisfy

RE: [linuxppc-release] [PATCH v5 1/5] powerpc/85xx: implement hardware timebase sync

Hi Kumar, There is no comment for these patches so far. Do you think these patches can be merged? We really want these patches to be merged in this merge window. Thanks. Best Regards, Chenhui > -Original Message- > From: Zhao Chenhui-B35336 > Sent: Friday, May 25, 2012 3:09 PM > To: W

MSR loses DE bit in 3.3, BDI2000 cant handle breakpoints?

I cannot make simple break points using BDI2000 work in 3.3, abatro suggests that it depends on MSR[DE] is cleared by the kernel. With the emulator I can see that MSR[DE] is off quite often by just stopping at random times and looking at MSR so it seems like the kernel is turning MSR[DE] off mos

[PATCH] powerpc: Clear RI and EE at the same time in system call exit

In system call exit we currently clear RI and EE separately. An mtmsrd is a slow operation and we can save cycles by doing it all in one go. This does complicate things a bit - we have to be careful to restore RI if we branch out before returning to userspace. On a POWER7 with virtual cputime di

RE: pread() and pwrite() system calls

On Tue, 2012-05-29 at 09:28 +0100, David Laight wrote: > > > A special pread/pwrite asm stub that just copies > > > r7 to r0 could be used. > > > > > > Would it be enough to do: > > > syscall_pread_pwrite: > > > mov 0,7 > > > sc > > > blr > > > and handle the -ve -> errno in C? > > > > Huh?

RE: pread() and pwrite() system calls

> > A special pread/pwrite asm stub that just copies > > r7 to r0 could be used. > > > > Would it be enough to do: > > syscall_pread_pwrite: > > mov 0,7 > > sc > > blr > > and handle the -ve -> errno in C? > > Huh? Won't fly, r0 is used for the system call number! I was copying that

[PATCH] powerpc: Use enhanced touch instructions in POWER7 copy_to_user/copy_from_user

Version 2.06 of the POWER ISA introduced enhanced touch instructions, allowing us to specify a number of attributes including the length of a stream. This patch adds a software stream for both loads and stores in the POWER7 copy_tofrom_user loop. Since the setup is quite complicated and we have t

Re: [PATCH v5 1/5] powerpc/85xx: implement hardware timebase sync

Hi Scott, Thanks for the valuable comment raised before and we have updated the patches accordingly. Please review the updated patch set and ACK if they are good to you. We hope it can be applied in this window. Leo On Fri, May 11, 2012 at 7:53 PM, Zhao Chenhui wrote: > Do hardware timebase s

[RFC PATCH 09/10] POWERPC: smp: remove call to ipi_call_lock()/ipi_call_unlock()

From: Yong Zhang 1) call_function.lock used in smp_call_function_many() is just to protect call_function.queue and &data->refs, cpu_online_mask is outside of the lock. And it's not necessary to protect cpu_online_mask, because data->cpumask is pre-calculate and even if a cpu is brougt up