RFC: linux-aarch64 and LSE support
Andrew Dinn
adinn at redhat.com
Wed Sep 14 09:25:48 UTC 2022
Hi Andrew/Kim,
On 09/09/2022 22:50, Kim Barrett wrote:
>> On Sep 9, 2022, at 5:24 AM, Andrew Haley <aph-open at littlepinkcloud.com> wrote:
. . .
>> So, if we revert to using the GCC intrinsics, we'd hurt performance on some
>> systems, and we'd lose some control. On the other hand it'd be cleaner. Much
>> cleaner. :-)
>
> Exactly.
>
> So what do folks think?
>
> Obviously, I can add support for the bitops using the existing structure (even
> though I'd rather not) and we can revisit this somewhat messy situation again
> later. Or we can clean up the code now.
I am swayed by Kim's arguments that we should clean this up now. This is
technical debt and the changes Kim wants to make imply an interest rate
hike (extra development and maintenance).
The potential for a performance hit on some (mostly legacy?) HW is
unfortunate but I am not convinced it would matter much for most apps.
Also, this change need not be backported which means there will still be
a legacy LTS version for deployed applications where the change does
make a noticeable difference to performance. A-and ... as Kim suggests,
it may still be possible to configure a way around problematic hardware
by passing the appropriate switches to gcc.
If that is all the argument against then I definitely prefer simplifying
this now.
regards,
Andrew Dinn
-----------
More information about the hotspot-dev
mailing list