Re: [PATCH] x86/speculation/l1tf: fix overflow on l1tf_pfn_limit() on 32bit

[Michal Hocko]

On Mon 20-08-18 13:41:03, Vlastimil Babka wrote:
> On 08/20/2018 12:49 PM, Michal Hocko wrote:
> > On Mon 20-08-18 11:58:35, Vlastimil Babka wrote:
> >> On 32bit PAE kernels on 64bit hardware with enough physical bits,
> >> l1tf_pfn_limit() will overflow unsigned long. This in turn affects
> >> max_swapfile_size() and can lead to swapon returning -EINVAL. This has been
> >> observed in a 32bit guest with 42 bits physical address size, where
> >> max_swapfile_size() overflows exactly to 1 << 32, thus zero, and produces the
> >> following warning to dmesg:
> >>
> >> [    6.396845] Truncating oversized swap area, only using 0k out of 2047996k
> >>
> >> Fix this by using unsigned long long instead.
> >>
> >> Reported-by: Dominique Leuenberger <dimstar@xxxxxxx>
> >> Reported-by: Adrian Schroeter <adrian@xxxxxxx>
> >> Fixes: 17dbca119312 ("x86/speculation/l1tf: Add sysfs reporting for l1tf")
> >> Fixes: 377eeaa8e11f ("x86/speculation/l1tf: Limit swap file size to MAX_PA/2")
> >> Cc: stable@xxxxxxxxxxxxxxx
> >> Signed-off-by: Vlastimil Babka <vbabka@xxxxxxx>
> > 
> > Looks good to me. I would probably use phys_addr_t which would be more
> > descriptive but this is just minor thing.
> 
> Hmm phys_addr_t is still 32bit on !PAE so there the overflow could still
> happen. I guess max_swapfile_size() should skip the whole L1TF part for
> !PAE since there is no pte inverting done anyway.

Yeah, I misremembered that we are already doing that.

> Also the value is "number of pages" which is not the same as "physical
> address" so the phys_addr_t could be misleading anyway?

right

-- 
Michal Hocko
SUSE Labs