Re: [tip:x86/mm] x86/kasan: Panic if there is not enough memory to boot

[Thomas Gleixner]

On Sun, 14 Jan 2018, Ingo Molnar wrote:

> 
> * tip-bot for Andrey Ryabinin <tipbot@xxxxxxxxx> wrote:
> 
> > Commit-ID:  74b61d1fb59bed3a6608161b05f2c0f5cb71acf6
> > Gitweb:     https://git.kernel.org/tip/74b61d1fb59bed3a6608161b05f2c0f5cb71acf6
> > Author:     Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx>
> > AuthorDate: Wed, 10 Jan 2018 18:36:02 +0300
> > Committer:  Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> > CommitDate: Sun, 14 Jan 2018 12:59:51 +0100
> > 
> > x86/kasan: Panic if there is not enough memory to boot
> > 
> > Currently KASAN doesn't panic in case it don't have enough memory
> > to boot. Instead, it crashes in some random place:
> > 
> >  kernel BUG at arch/x86/mm/physaddr.c:27!
> > 
> >  RIP: 0010:__phys_addr+0x268/0x276
> >  Call Trace:
> >   kasan_populate_shadow+0x3f2/0x497
> >   kasan_init+0x12e/0x2b2
> >   setup_arch+0x2825/0x2a2c
> >   start_kernel+0xc8/0x15f4
> >   x86_64_start_reservations+0x2a/0x2c
> >   x86_64_start_kernel+0x72/0x75
> >   secondary_startup_64+0xa5/0xb0
> > 
> > Use memblock_virt_alloc_try_nid() for allocations without failure
> > fallback. It will panic with an out of memory message.
> > 
> > Reported-by: kernel test robot <xiaolong.ye@xxxxxxxxx>
> > Signed-off-by: Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx>
> > Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> > Acked-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
> > Cc: kasan-dev@xxxxxxxxxxxxxxxx
> > Cc: Alexander Potapenko <glider@xxxxxxxxxx>
> > Cc: lkp@xxxxxx
> > Link: https://lkml.kernel.org/r/20180110153602.18919-1-aryabinin@xxxxxxxxxxxxx
> 
> I think this commit should be in tip:x86/pti for dependency reasons, to get 
> backported automatically. I suspect we want the same KASAN robustness in -stable.
> 
> If this patch is too high risk and should be in x86/mm, then it should still be 
> x86/pti based, to allow later merging into x86/pti.
> 
> The latest already-upstream commit of x86/pti is de791821c295, and this commit 
> applies cleanly to that base.

makes sense. I'll move it to x86/pti ...