Re: Linux 6.19.4 - Oops, regression

[Genes Lists]

On Fri, 2026-02-27 at 16:01 -0500, Kris Karas (Bug Reporting) wrote:
> Genes Lists wrote:
> 
> 
> I second Greg's comment - "Odd" - in wondering why you need to patch 
> userspace and/or recompile it if the same userspace was working fine
> on 
> 6.19.3?
> 
> FWIW, nftables 1.1.6 (libnftnl 1.3.1, libmnl 1.0.5) is still working 
> fine for me under 6.19.4 using the .abort_skip_removal patches from 
> Pablo as forwarded by Greg.
> 
> Kris

Thanks Kris, I'll try help clarify. Things weren't crystal clear at
first.

There were actually 2 separate issues that happened to hit at the same
time.

 (1) the 6.19.4 netfilter crash 
     Now fixed or will be shortly in 6.19.5.

and 

 (2) a userspace issue with nft v1.1.6 and large sets leading to 
     spurious 'file exists' errors. 
     This has nothing at all to do with any kernel oops.

     This 2nd issue has been fixed in nftables by commit: 

  ---
    commit e83e32c8d1cd228d751fb92b756306c6eb6c0759
    Author: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx>
    Date:   Mon Jan 12 12:59:26 2026 +0100

    mnl: restore create element command with large batches
    
    The rework to reduce memory consumption has introduced a
    bug that result in spurious EEXIST with large batches.
    ---

This commit came after nftables v1.1.6 release.
Rebuilding nftables with this commit sorts out the 2nd (userspace)
issue.

gene

Attachment: signature.asc
Description: This is a digitally signed message part