On 2/15/22 3:54 PM, Segher Boessenkool wrote: > On Tue, Feb 15, 2022 at 01:03:09PM -0600, Peter Bergner wrote: >> The HTM tbegin. instruction can fail intermittently due to many reasons. > > Just a few really. But, if the transaction fails it will appear as if > the tbegin. had an error as well, is that what you are seeing?
Yeah, exactly. The transaction itself fails, which rolls us back to the tbegin. with an error return value so we fall into the else/error clause. > That is the way any HTM code should be written in the first place > (except for rollback-only transactions, but let's not go there -- > besides, it is normal for those to fail as well, and there needs to be a > fallback there as well :-) ) Agreed and I'm not sure why I didn't write it that way to begin with. Maybe I thought it was so simple that the likelihood of it failing was so small we'd never see it? Anyway, we do now, so... > The patch is fine. Okay for trunk and backports (after soak time ofc). > Thanks! Thanks, pushed. Peter
