Linux Virtualization - Apr 2018 - [PATCH v3] kvmalloc: always use vmalloc if CONFIG_DEBUG_SG

On Tue, Apr 24, 2018 at 08:29:14AM -0400, Mikulas Patocka
wrote:
> 
> 
> On Mon, 23 Apr 2018, Matthew Wilcox wrote:
> 
> > On Mon, Apr 23, 2018 at 08:06:16PM -0400, Mikulas Patocka wrote:
> > > Some bugs (such as buffer overflows) are better detected
> > > with kmalloc code, so we must test the kmalloc path too.
> > 
> > Well now, this brings up another item for the collective TODO list --
> > implement redzone checks for vmalloc.  Unless this is something
already
> > taken care of by kasan or similar.
> 
> The kmalloc overflow testing is also not ideal - it rounds the size up to 
> the next slab size and detects buffer overflows only at this boundary.
> 
> Some times ago, I made a "kmalloc guard" patch that places a
magic number
> immediatelly after the requested size - so that it can detect overflows at 
> byte boundary 
> ( https://www.redhat.com/archives/dm-devel/2014-September/msg00018.html )
> 
> That patch found a bug in crypto code:
> ( http://lkml.iu.edu/hypermail/linux/kernel/1409.1/02325.html )
Is it still worth doing this, now we have kasan?

On Tue, 24 Apr 2018, Matthew Wilcox wrote:
> On Tue, Apr 24, 2018 at 08:29:14AM -0400, Mikulas Patocka wrote:
> > 
> > 
> > On Mon, 23 Apr 2018, Matthew Wilcox wrote:
> > 
> > > On Mon, Apr 23, 2018 at 08:06:16PM -0400, Mikulas Patocka wrote:
> > > > Some bugs (such as buffer overflows) are better detected
> > > > with kmalloc code, so we must test the kmalloc path too.
> > > 
> > > Well now, this brings up another item for the collective TODO
list --
> > > implement redzone checks for vmalloc.  Unless this is something
already
> > > taken care of by kasan or similar.
> > 
> > The kmalloc overflow testing is also not ideal - it rounds the size up
to
> > the next slab size and detects buffer overflows only at this boundary.
> > 
> > Some times ago, I made a "kmalloc guard" patch that places a
magic number
> > immediatelly after the requested size - so that it can detect
overflows at
> > byte boundary 
> > (
https://www.redhat.com/archives/dm-devel/2014-September/msg00018.html )
> > 
> > That patch found a bug in crypto code:
> > ( http://lkml.iu.edu/hypermail/linux/kernel/1409.1/02325.html )
> 
> Is it still worth doing this, now we have kasan?
The kmalloc guard has much lower overhead than kasan.

(BTW. when I tried kasan, it oopsed with persistent memory)

Mikulas

Hello, Mikulas.

On Tue, Apr 24, 2018 at 02:41:47PM -0400, Mikulas Patocka
wrote:
> 
> 
> On Tue, 24 Apr 2018, Matthew Wilcox wrote:
> 
> > On Tue, Apr 24, 2018 at 08:29:14AM -0400, Mikulas Patocka wrote:
> > > 
> > > 
> > > On Mon, 23 Apr 2018, Matthew Wilcox wrote:
> > > 
> > > > On Mon, Apr 23, 2018 at 08:06:16PM -0400, Mikulas Patocka
wrote:
> > > > > Some bugs (such as buffer overflows) are better
detected
> > > > > with kmalloc code, so we must test the kmalloc path
too.
> > > > 
> > > > Well now, this brings up another item for the collective
TODO list --
> > > > implement redzone checks for vmalloc.  Unless this is
something already
> > > > taken care of by kasan or similar.
> > > 
> > > The kmalloc overflow testing is also not ideal - it rounds the
size up to
> > > the next slab size and detects buffer overflows only at this
boundary.
> > > 
> > > Some times ago, I made a "kmalloc guard" patch that
places a magic number
> > > immediatelly after the requested size - so that it can detect
overflows at
> > > byte boundary 
> > > (
https://www.redhat.com/archives/dm-devel/2014-September/msg00018.html )
> > > 
> > > That patch found a bug in crypto code:
> > > ( http://lkml.iu.edu/hypermail/linux/kernel/1409.1/02325.html )
> > 
> > Is it still worth doing this, now we have kasan?
> 
> The kmalloc guard has much lower overhead than kasan.
I skimm at your code and it requires rebuilding the kernel.
I think that if rebuilding is required as the same with the KASAN,
using the KASAN is better since it has far better coverage for
detection the bug.

However, I think that if the redzone can be setup tightly
without rebuild, it would be worth implementing. I have an idea to
implement it only for the SLUB. Could I try it? (I'm asking this
because I'm inspired from the above patch.) :)
Or do you wanna try it?

Thanks.