Re: preparations for 4.19.1

[Jan Beulich <jbeulich@xxxxxxxx>]

On 13.11.2024 16:20, Andrew Cooper wrote:
> On 12/11/2024 1:00 pm, Jan Beulich wrote:
>> the release is due by the end of the month. Please point out backports you 
>> find
>> missing from the respective staging branch, but which you consider relevant.
> 
> Looking over the XenServer patchqueue:

First, overall - that's quite a few. Most not even having a Fixes: tag. Or
in one case having even two of them alongside saying "no functional change".
Plus it would be really nice if someone else could take care of tool stack
backports again.

In more detail:

> These are all bugfixes, some that came from customer bugs:
> 
> e42e4d8c3e2c: tools/misc: xen-hvmcrash: Inject #DF instead of
> overwriting RIP

I'm on the edge here, but since you're asking for it, I'll include it.

> ba709d514aac: x86/viridian: Clarify some viridian logging strings

Not really a backporting candidate imo.

> d81dd3130351: x86/shutdown: change default reboot method preference

I'll include that, despite the different perspective I'm taking.

> 0d69635d27cb: tools/ocaml: Remove '-cc $(CC)' from OCAMLOPTFLAGS
> 8ffcf184affb: tools/ocaml: Fix OCaml libs rules
> 126293eae648: tools/ocaml: Drop the OCAMLOPTFLAG_G invocation
> 1965e9a93074: tools/ocaml: Fix the version embedded in META files

Provided these apply as-is, I'll blindly take these.

> e58a2858d588: x86/boot: Preserve the value clobbered by the load-base
> calculation

Based on its Fixes: tags I probably should have included this one already.

> This is a diagnostic improvement, also from a customer bug:
> 
> 2f413e22fa5e: x86/msr: add log messages to MSR state load error paths

Again - not really a backporting candidate imo.

> These are a SIGPIPE bugfix which happen to also have a perf
> improvement.  I cant remember if we discussed backporting them before. 
> (Juergen/Anthony?)
> 
> 42db2deb5e76: tools/libxs: Fix length check in xs_talkv()
> e2a93bed8b9e: tools/libxs: Rework xs_talkv() to take xsd_sockmsg within
> the iovec
> f050c03ce2ad: tools/libxs: Rationalise the definition of struct xs_handle
> 046efe529e82: tools/libxs: Track whether we're using a socket or file
> ebaeb0c64a6d: tools/libxs: Use writev()/sendmsg() instead of write()
> a17b6db9b007: tools/libxs: Stop playing with SIGPIPE

See the earlier reply to both Jürgen and you.

> These are from a livepatching snafu:
> 
> 3a28da8f4daf: xen/livepatch: remove useless check for duplicated sections
> 8c81423038f1: xen/livepatch: drop load_addr Elf section field
> 86d09d16dd74: xen/livepatch: simplify and unify logic in prepare_payload()
> fa49f4be413c: xen/livepatch: do Xen build-id check earlier
> aa5a06d5d6ed: x86/alternatives: do not BUG during apply
> 
> where the buildid check is much too late.

I certainly agree with taking the last two. The first three though are all
"no functional change", which generally I'd prefer to omit unless they're
strictly prereqs, or diverging from master is deemed to be a severe issue.

> And from looking at staging:
> 
> fa2d8318033e: x86/cpu-policy: Extend the guest max policy max leaf/subleaves
> 
> This fixes a real issue on older AMD systems.

Will include it.

Jan