[PATCH 3/5] arm: Flush TLB entries in setup_mm_for_reboot()
tony at atomide.com
Sat Dec 19 13:10:16 EST 2009
* Tony Lindgren <tony at atomide.com> [091219 09:53]:
> * Russell King - ARM Linux <linux at arm.linux.org.uk> [091219 01:04]:
> > On Fri, Dec 18, 2009 at 07:46:19PM -0800, Tony Lindgren wrote:
> > > diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
> > > index 8c7fbd1..a2802ea 100644
> > > --- a/arch/arm/mm/mmu.c
> > > +++ b/arch/arm/mm/mmu.c
> > > @@ -1068,4 +1068,7 @@ void setup_mm_for_reboot(char mode)
> > > pmd = __pmd(pmdval + (1 << (PGDIR_SHIFT - 1)));
> > > flush_pmd_entry(pmd);
> > > }
> > > +
> > > + local_flush_tlb_all();
> > > + flush_cache_all();
> > The cache flush is unnecessary. The caches should be off at this point.
> OK, right, cpu_proc_fin() has already disabled caches by then.
> Here's the updated version.
Hmm, without the flush_cache_all() rebooting with kexec still fails
with crc error eventually.
So it seems like there's a bug somewhere. Maybe something missing from
I don't see anything obvious in v7_flush_dcache_all that would explain
why it makes local_flush_tlb_all() behave. Like you said caches are
off at this point, and that should also disable the L2 cache.
> From f279fdc4c7dc824902d67511bfcb5f18512c8468 Mon Sep 17 00:00:00 2001
> From: Tony Lindgren <tony at atomide.com>
> Date: Fri, 18 Dec 2009 17:10:51 -0800
> Subject: [PATCH] arm: Flush TLB entries in setup_mm_for_reboot()
> We need to do that if we tinker with the MMU entries.
> This fixes the occasional bug with kexec where the new
> fails to uncompress with "crc error". Most likely at
> least kexec on v6 and v7 need this fix.
> Signed-off-by: Tony Lindgren <tony at atomide.com>
> diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
> index 8c7fbd1..a311080 100644
> --- a/arch/arm/mm/mmu.c
> +++ b/arch/arm/mm/mmu.c
> @@ -1068,4 +1068,6 @@ void setup_mm_for_reboot(char mode)
> pmd = __pmd(pmdval + (1 << (PGDIR_SHIFT - 1)));
> + local_flush_tlb_all();
More information about the linux-arm-kernel