Re: [PATCH 5.10] bpf: Fix mismatch memory accounting for devmap maps

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Fri, Sep 27, 2024 at 04:03:36PM +0800, Pu Lehui wrote:
> 
> 
> On 2024/9/27 15:56, Greg Kroah-Hartman wrote:
> > On Fri, Sep 20, 2024 at 10:39:50AM +0000, Pu Lehui wrote:
> > > From: Pu Lehui <pulehui@huawei.com>
> > > 
> > > Commit 70294d8bc31f ("bpf: Eliminate rlimit-based memory accounting for
> > > devmap maps") relies on the v5.11+ basic mechanism of memcg-based memory
> > > accounting [0]. The commit cannot be independently backported to the
> > > 5.10 stable branch, otherwise the related memory when creating devmap
> > > will be unrestricted and the associated bpf selftest map_ptr will fail.
> > > Let's roll back to rlimit-based memory accounting mode for devmap and
> > > re-adapt the commit 225da02acdc9 ("bpf: Fix DEVMAP_HASH overflow check
> > > on 32-bit arches") to the 5.10 stable branch.
> > > 
> > > Link: https://lore.kernel.org/bpf/20201201215900.3569844-1-guro@fb.com [0]
> > > Fixes: 225da02acdc9 ("bpf: Fix DEVMAP_HASH overflow check on 32-bit arches")
> > > Fixes: 70294d8bc31f ("bpf: Eliminate rlimit-based memory accounting for devmap maps")
> > 
> > Should we just revert these changes instead?
> 
> Yes, Greg. My patch is to revert these two commits and re-adapt commit
> 225da02acdc9 ("bpf: Fix DEVMAP_HASH overflow check on 32-bit arches").
> 
> Shall we need to split this patch into multiple patches?

Yes, please do so.