Messages in this thread Patch in this message | | | Subject | Re: [PATCH] Call flush_dcache_page around PIO data transfers in libata-aff.c | From | Catalin Marinas <> | Date | Wed, 03 Mar 2010 11:18:25 +0000 |
| |
On Wed, 2010-03-03 at 08:24 +0000, Sebastian Andrzej Siewior wrote: > > drivers/ata/libata-sff.c | 6 ++++++ > > 1 files changed, 6 insertions(+), 0 deletions(-) > > > >diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c > >index 741065c..3d3c854 100644 > >--- a/drivers/ata/libata-sff.c > >+++ b/drivers/ata/libata-sff.c > >@@ -893,6 +896,9 @@ static void ata_pio_sector(struct ata_queued_cmd *qc) > > do_write); > > } > > > >+ if (!do_write) > >+ flush_dcache_page(page); > >+ > > qc->curbytes += qc->sect_size; > > qc->cursg_ofs += qc->sect_size; > > That part got merged and it leads to a BUG if CONFIG_DEBUG_VM is > enabled. My mips box and arm use page_mapping() in flush_dcache_page() > which has VM_BUG_ON(PageSlab(page)) in it. > > Any comments on that?
There is a long thread on USB HCD and cache coherency on ARM which I started pretty much at the same time with the libata-sff change. I need to do some testing but in the view of the discussions in the other thread, we may not need this patch at all:
http://thread.gmane.org/gmane.linux.usb.general/27072
Basically, the meaning of PG_arch_1 should be page clean rather than page dirty.
-- Catalin
| |