On Fri, Aug 09, 2013 at 02:26:40AM +0100, Ming Lei wrote: > In isp1760-hcd driver, flush_dcache_page() is introduced in commit > db8516f61b4(USB: isp1760: Flush the D-cache for the pipe-in transfer buffers) > to fix cache incoherency problem when PIO reading from USB mass storage > device to mapped page, so the flush should not need for slab pages which > won't be mapped to userspace. > > This patch fixes one kernel panic when CONFIG_DEBUG_VM is set. > > Cc: Catalin Marinas <[email protected]> > Tested-by: Christoffer Dall <[email protected]> > Signed-off-by: Ming Lei <[email protected]> > --- > drivers/usb/host/isp1760-hcd.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/drivers/usb/host/isp1760-hcd.c b/drivers/usb/host/isp1760-hcd.c > index 2facee5..c370e7d 100644 > --- a/drivers/usb/host/isp1760-hcd.c > +++ b/drivers/usb/host/isp1760-hcd.c > @@ -703,7 +703,9 @@ __acquires(priv->lock) > urb->status = 0; > } > > - if (usb_pipein(urb->pipe) && usb_pipetype(urb->pipe) != PIPE_CONTROL) { > + if (usb_pipein(urb->pipe) && > + usb_pipetype(urb->pipe) != PIPE_CONTROL && > + !PageSlab(virt_to_page(urb->transfer_buffer))) { > void *ptr; > for (ptr = urb->transfer_buffer; > ptr < urb->transfer_buffer + urb->transfer_buffer_length;
I think from the arm port perspective we can even revert the original patch. Rather than adding flush_dcache_page() in several drivers we moved to making the page cache pages dirty by default (and PG_arch_1 means 'clean') so that they are automatically flushed when first mapped into user space. -- Catalin -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
