Re: [PATCH 6.6 000/587] 6.6.103-rc1 review

[Brett Sheffield]

On 2025-08-26 16:58, Brett Sheffield wrote:
> On 2025-08-26 14:50, Greg Kroah-Hartman wrote:
> > On Tue, Aug 26, 2025 at 12:42:40PM +0000, Brett A C Sheffield wrote:
> > > Hi Greg,
> > > 
> > > On 2025-08-26 13:02, Greg Kroah-Hartman wrote:
> > > > This is the start of the stable review cycle for the 6.6.103 release.
> > > > There are 587 patches in this series, all will be posted as a response
> > > > to this one.  If anyone has any issues with these being applied, please
> > > > let me know.
> > > > 
> > > > Responses should be made by Thu, 28 Aug 2025 11:08:24 +0000.
> > > > Anything received after that time might be too late.
> > > 
> > > Quick query - should we be backporting a known regression, even if it is in
> > > mainline presently, or do we wait until the fix is applied to mainline and
> > > *then* backport both patches?
> > > 
> > > 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
> > > 
> > > introduces a regression which breaks IPv4 broadcast, which stops WOL working
> > > (breaking my CI system), among other things:
> > > 
> > > https://lore.kernel.org/regressions/20250822165231.4353-4-bacs@xxxxxxxxxxxxx
> > > 
> > > This regression has *already* been backported to:
> > > 
> > > - 6.16.3
> > > - 6.12.43
> > > 
> > > so I guess we wait for a fix for these.
> > > 
> > > However, it is not yet present in the other stable kernels.  The new stable
> > > release candidates today would spread the breakage to:
> > > 
> > > - 6.6.y
> > > - 6.1.y
> > > - 5.15.y
> > > - 5.10.y
> > > 
> > > Do we revert this patch in today's RCs for now, or keep it for full
> > > compatibility with mainline bugliness?
> > 
> > Is the fix in linux-next yet?  If it's there, I can queue it up
> > everywhere, which might be best.
> 
> Not yet, but I'll let you know as soon as it is.  I'd suggest dropping
> 9e30ecf23b1b from 6.6.y 6.1.y 5.15.y 5.10.y and 5.4.y until the fix is
> available.

Hi Greg,

The fix has hit Linus' tree now with commit:

5189446ba995 ("net: ipv4: fix regression in local-broadcast routes")

This needs to be applied to all stable trees.

Cheers,

Brett