[PATCH] lib/scatterlist.c: don't flush_kernel_dcache_page on slab page

[Aaro Koskinen]

Hi,

On Thu, Oct 31, 2013 at 03:27:45PM -0700, Andrew Morton wrote:
> On Tue, 29 Oct 2013 00:20:05 +0800 Ming Lei <ming.lei at canonical.com> wrote:
> 
> > Commit b1adaf65ba03([SCSI] block: add sg buffer copy helper functions)
> > introduces two sg buffer copy helpers, and calls flush_kernel_dcache_page()
> > on pages in SG list after these pages are written to.
> > 
> > Unfortunately, the commit may introduce a potential bug:
> > 
> > 	- Before sending some SCSI commands, kmalloc() buffer may be
> > 	passed to block layper, so flush_kernel_dcache_page() can
> > 	see a slab page finally
> > 
> > 	- According to cachetlb.txt, flush_kernel_dcache_page() is
> > 	only called on "a user page", which surely can't be a slab page.
> > 
> > 	- ARCH's implementation of flush_kernel_dcache_page() may
> > 	use page mapping information to do optimization so page_mapping()
> > 	will see the slab page, then VM_BUG_ON() is triggered.
> > 
> > Aaro Koskinen reported the bug on ARM/kirkwood when DEBUG_VM is enabled,
> > and this patch fixes the bug by adding test of '!PageSlab(miter->page)'
> > before calling flush_kernel_dcache_page().
> 
> We should work out which kernel(s) need this patch.  b1adaf65ba03 was
> merged in 2008, so presumably some more recent patch has exposed the
> problem, but I don't know what one that was.
> 
> Help me out here?

On ARM, this problem started to appear after
1bc39742aab09248169ef9d3727c9def3528b3f3 (ARM: 7755/1: handle user space
mapped pages in flush_kernel_dcache_page) and that was tagged for 3.2+
stable kernels.

A.