[RFC PATCH v2 3/4] liveupdate: Replace compatibility string with a 2-bit pair feature header

[tarunsahu at google.com]

Pasha Tatashin <pasha.tatashin at soleen.com> writes:

> On 10-28 10:24, David Matlack wrote:
>> On Wed, Oct 28, 2026 at 10:17 AM Pratyush Yadav <pratyush at kernel.org> wrote:
>> >
>> > On Mon, Oct 26 2026, David Matlack wrote:
>> >
>> > > On 2026-10-25 04:38 PM, Pasha Tatashin wrote:
>> > >> On Fri, Oct 23, 2026 at 9:50 AM David Matlack <dmatlack at google.com> wrote:
>> > >
>> > [...]
>> > >> The reason I care about keeping the header compact is that we are
>> > >> defining a foundational kernel ABI and want to use this header across
>> > >> almost every serialized data structure so we can evolve them later
>> > >> without breaking compatibility. While some top-level structs are on
>> > >> sparse folios today, other serialized structs are tightly packed, and
>> > >> we are looking to add KHO slab support in the future where tightly
>> > >> packed allocations will be common.
>> > >
>> > > My intuition is we will have a small number of headers per preserved
>> > > file and per preserved FLB. One for the top-level struct and then
>> > > optionally a few more for sub-structs like arch-specific data. That is
>> > > why I am less worried 8 bytes per header, even if we fix the folio
>> > > problem.
>> > >
>> > > Maybe where our visions differ is when there are collections of
>> > > serialized structs, like KHO blocks. For collections I don't think the
>> > > elements will each have their own headers. Otherwise the elements could
>> > > be different sizes.
>> >
>> > kho_block today doesn't store entry_size in the serialized payload. It
>> > relies on its caller to do so. With extensible structures, this can get
>> > messy. Sure, the owner of the kho_block_set can track the size somewhere
>> > in its payload, but perhaps we should store the size in
>> > kho_block_header_ser to begin with?
>> 
>> +1000
>> 
>> KHO block should track the entry size automatically and not rely on
>> the caller. Otherwise every caller is going to have to store the entry
>> size in the struct that owns the KHO block, like I had to do in the
>> PCI v10 series :(
>
> I am fine with updating this as part of the versioning change so that we 
> do not have to do it twice.
>
> Tarun, could you pelase include entry size change into the next version as 
> well?

Yes, This will be important to track the extensible serialized structure
across kernel. For example struct luo_session_ser has larger size in N+1
kernel which gets deserialized on the beginning of the boot so incoming
block_set must use entry_size of the kernel N which we can update inside
kho_block_set_restore(). And this updated incoming->block_set be used
for iterator in luo_session_deserialize to properly walk through
session_ser entries.

I will send v4 soon including this.

~Tarun

>
> Thanks,
> Pasha
>
>> 
>> > Anyway, not a problem this series needs to solve, but something we
>> > should eventually do.