[PATCH] Linux: Remove rseq support

[Carlos O'Donell]

On 7/16/20 10:48 AM, Mathieu Desnoyers wrote:
> ----- On Jul 15, 2020, at 11:11 AM, Florian Weimer fweimer@redhat.com wrote:
> 
>> The kernel ABI is not finalized, and there are now various proposals
>> to change the size of struct rseq, which would make the glibc ABI
>> dependent on the version of the kernels used for building glibc.
>> This is of course not acceptable.
>>
>> This reverts commit 48699da1c468543ade14777819bd1b4d652709de ("elf:
>> Support at least 32-byte alignment in static dlopen"), commit
>> 8f4632deb3545b2949cec5454afc3cb21a0024ea ("Linux: rseq registration
>> tests"), commit 6e29cb3f61ff5432c78a1c84b0d9b123a350ab36 ("Linux: Use
>> rseq in sched_getcpu if available"), and commit
>> 0c76fc3c2b346dc5401dc055d97d4279632b0fb3 ("Linux: Perform rseq
>> registration at C startup and thread creation"), resolving the conflicts
>> introduced by the ARC port and the TLS static surplus changes.
>>
>> Tested on x86_64-linux-gnu and built with build-many-glibcs.py.
> 
> I agree with this revert, let's take time to discuss and document the
> extensible rseq ABI scheme before we freeze it in the glibc ABI.
> 
> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Agreed, we are making progress technically, and extensibility has
come up as a relevant feature to consider. I'm interested to hash
this out in the next month and get rseq in Fedora Rawhide early so
we can push on the sandboxes to update their seccomp-bpf lists.
That way we get 5 months of live testing in a distro before glibc
publicly includes the ABI.

OK with revert for 2.32.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>

-- 
Cheers,
Carlos.