Hi,

On Wed, May 13, 2026 at 07:35:10PM +0200, Solar Designer wrote:
> Hi,
> 
> Here's analysis by the Dirty Frag researcher:
> 
> On Thu, May 14, 2026 at 12:45:31AM +0900, Hyunwoo Kim wrote:
> > I'm attaching my current analysis, so if anyone could post it to 
> > oss-security on my behalf, I would greatly appreciate it.
> > 
> > Here is the analysis:
> > 
> > This vulnerability is a path that was accidentally activated _after_ the 
> > introduction of f4c50a4034e6 (2026-05-05), the patch for CVE-2026-43284 in 
> > the Dirty Frag chain.
> > 
> > In other words, the effective vulnerability window is from f4c50a4034e6 
> > (2026-05-05) to upstream -- approximately 9 days.
> > 
> > Exploitation requires the attacker to have permission to create user 
> > namespaces, unless chained with a separate vulnerability.
> > 
> > Another important point: the patch[1] they attached addresses the 
> > skb_try_coalesce path, but does not resolve the vulnerability because it 
> > misses other variant paths.
> > 
> > For now, a patch[2] that also covers the currently-analyzed 
> > __pskb_copy_fclone path has been submitted. Once the additional analysis is 
> > finalized, a v2 patch may be submitted.
> > 
> > At least for the time being, I recommend keeping the Dirty Frag mitigation 
> > in place:
> > ```
> > sh -c "printf 'install esp4 /bin/false\ninstall esp6 /bin/false\ninstall 
> > rxrpc /bin/false\n' > /etc/modprobe.d/dirtyfrag.conf; rmmod esp4 esp6 rxrpc 
> > 2>/dev/null; echo 3 > /proc/sys/vm/drop_caches; true"
> > ```
> > 
> > [1]: https://lore.kernel.org/all/[email protected]/
> > [2]: https://lore.kernel.org/all/agRfuVOeMI5pbHhY@v4bel/

FWIW, there was another variant posted via
https://lore.kernel.org/netdev/agVpIsaSherjHTYg@sultan-box/ for review
of the v2 patch. A v3 of the patch has in meanwhile been posted as
well as:

https://lore.kernel.org/netdev/agW4vC0r8QOUKtRT@v4bel/

Regards,
Salvatore

Reply via email to