On Fri 30-11-18 15:36:51, Kirill A. Shutemov wrote:
> On Fri, Nov 30, 2018 at 01:18:51PM +0100, Michal Hocko wrote:
> > On Fri 30-11-18 13:06:57, Jan Stancek wrote:
> > > LTP proc01 testcase has been observed to rarely trigger crashes
> > > on arm64:
> > > page_mapped+0x78/0xb4
> > > stable_page_flags+0x27c/0x338
> > > kpageflags_read+0xfc/0x164
> > > proc_reg_read+0x7c/0xb8
> > > __vfs_read+0x58/0x178
> > > vfs_read+0x90/0x14c
> > > SyS_read+0x60/0xc0
> > >
> > > Issue is that page_mapped() assumes that if compound page is not
> > > huge, then it must be THP. But if this is 'normal' compound page
> > > (COMPOUND_PAGE_DTOR), then following loop can keep running
> > > (for HPAGE_PMD_NR iterations) until it tries to read from memory
> > > that isn't mapped and triggers a panic:
> > > for (i = 0; i < hpage_nr_pages(page); i++) {
> > > if (atomic_read(&page[i]._mapcount) >= 0)
> > > return true;
> > > }
> > >
> > > I could replicate this on x86 (v4.20-rc4-98-g60b548237fed) only
> > > with a custom kernel module [1] which:
> > > - allocates compound page (PAGEC) of order 1
> > > - allocates 2 normal pages (COPY), which are initialized to 0xff
> > > (to satisfy _mapcount >= 0)
> > > - 2 PAGEC page structs are copied to address of first COPY page
> > > - second page of COPY is marked as not present
> > > - call to page_mapped(COPY) now triggers fault on access to 2nd
> > > COPY page at offset 0x30 (_mapcount)
> > >
> > > [1] https://github.com/jstancek/reproducers/blob/master/kernel/page_mapped_crash/repro.c
> > >
> > > Fix the loop to iterate for "1 << compound_order" pages.
> >
> > This is much less magic than the previous version. It is still not clear
> > to me how is mapping higher order pages to page tables other than THP
> > though. So a more detailed information about the source would bre really
> > welcome. Once we know that we can add a Fixes tag and also mark the
> > patch for stable because that sounds like a stable material.
>
> IIRC, sound subsystem can producuce custom mapped compound pages.
Do I assume correctly that consecutive ptes simply point to subpages?
> The bug dates back to e1534ae95004.
Thanks for the pointer. I thought this was a new and creative usage of
the pte->page mapping but it really looks like your commit just changed
the underlying semantic.
> > > Debugged-by: Laszlo Ersek <lersek@xxxxxxxxxx>
> > > Suggested-by: "Kirill A. Shutemov" <kirill@xxxxxxxxxxxxx>
> > > Signed-off-by: Jan Stancek <jstancek@xxxxxxxxxx>
> >
> > The patch looks sensible to me
> > Acked-by: Michal Hocko <mhocko@xxxxxxxx>
>
> Acked-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
>
> --
> Kirill A. Shutemov
--
Michal Hocko
SUSE Labs
