[PATCH] ARM: add BUILD_BUG_ON to check if fixmap range spans multiple pmds

[Ard Biesheuvel]

On Tue, 26 Oct 2021 at 12:55, Russell King (Oracle)
<linux at armlinux.org.uk> wrote:
>
> On Tue, Oct 26, 2021 at 06:38:16PM +0800, Quanyang Wang wrote:
> > Hi Ard,
> >
> > On 10/26/21 6:12 PM, Ard Biesheuvel wrote:
> > > On Tue, 26 Oct 2021 at 11:53, Quanyang Wang <quanyang.wang at windriver.com> wrote:
> > > >
> > > > Hi,
> > > > Sorry for the inconvenience.
> > > >
> > > > On 10/26/21 4:59 PM, Russell King (Oracle) wrote:
> > > > > On Sun, Oct 24, 2021 at 11:44:31PM +0200, Linus Walleij wrote:
> > > > > > On Wed, Oct 20, 2021 at 7:50 AM <quanyang.wang at windriver.com> wrote:
> > > > > >
> > > > > > > From: Quanyang Wang <quanyang.wang at windriver.com>
> > > > > > >
> > > > > > > Not only the early fixmap range, but also the fixmap range should be
> > > > > > > checked if it spans multiple pmds. When enabling CONFIG_DEBUG_HIGHMEM,
> > > > > > > some systems which contain up to 16 CPUs will crash.
> > > > > > >
> > > > > > > Signed-off-by: Quanyang Wang <quanyang.wang at windriver.com>
> > > > > >
> > > > > > Looks reasonable to me.
> > > > > > Reviewed-by: Linus Walleij <linus.walleij at linaro.org>
> > > > > >
> > > > > > Please submit this patch into Russell's patch tracker.
> > > > >
> > > > > ... and has totally broken what looks like _all_ ARM kernel builds.
> > > > This patch is intended to trigger build error when it check the value of
> > > > __end_of_fixmap_region is equal or larger than 256.
> > >
> > > Why? The fixmap region is larger than one PMD, so why do we need to cap it?
> > In __kmap_local_pfn_prot, arch_kmap_local_set_pte(&init_mm, vaddr, kmap_pte
> > - idx, pteval) is used to set pteval.
> > But the ptep is calculated by "kmap_pte - idx", which means all ptes must be
> > placed next to each other and no gaps. But for ARM, the ptes for the range
> > "0xffe00000~0xfff00000" is not next to the ptes for the range
> > "0xffc80000~0xffdfffff".
> >
> > When the idx is larger than 256, virtual address is in 0xffdxxxxx, access
> > this address will crash since its pteval isn't set correctly.
>
> Thanks for the explanation.
>
> Sadly, this does seem to be correct. Even if the PTE tables are
> located next to each other in memory, they _still_ won't be a
> contiguous array of entries due to being interleaved with the Linux
> PTE table and the hardware PTE table.
>
> Since the address range 0xffe00000-0xfff00000 is already half of one
> PTE table containing 512 contiguous entries, we are limited to 256
> fixmap PTEs maximum. If we have more than that we will start trampling
> over memory below the PTE table _and_ we will start corrupting Linux
> PTE entries in the 0xfff00000-0xffffffff range.
>
> I suspect this hasn't been seen because of a general lack of ARM
> systems with more than 4 CPUs.
>

But doesn't that make it a kmap_local regression? Or do you think this
issue existed before that as well?