4 ms·
For the most part he actually does a good job staying away from personal attacks. The problem is that his approach isn't going to yield great results. For the
by vlovich123 9y ago
For the most part he actually does a good job staying away from personal attacks. The problem is that his approach isn't going to yield great results. For there's 2 problematic parts to this e-mail.
The first is the personal attack when he calls Intel engineers morons. That 1 little unnecessary addon is uncalled for because it's quite reasonable that there's a lot of politics going on at Intel that's interfering with the engineering work. Could be that the next chip they're taping out can't be fixed properly without delaying & impacting revenue in a way that execs won't allow. Could be that there's legal reasons whereby Intel's "fix" is desirable from a management perspective to continue their denial of a problem with their chips to win or reduce the costs of the class actions. Or it could be any number of a hundred other reasons that the public isn't privy to.
The second, & this is pure speculation on my part that this didn't happen, is that he should have followed this in an offline discussion with the engineer. This public shaming doesn't do any good & just entrenches the frontline representatives of Intel to be antagonistic with Linus rather than to be his allies. I feel like if Linus wanted to be truly effective here in impacting Intel's direction (especially given their constant need to minimize the scale & scope of this problem) he could have followed up with the engineer off-list & gone "your explanation doesn't make any sense. can you clue me in on what's the motivation behind this approach?". It saves the engineer public embarrassment, it gets him an ally within Intel & it lets him have strategic information on what is motivating Intel engineers (consider - did his finger to nVidia in any way improve relations between Linux & nVidia?). Now if he did that & the answer still wasn't OK then he's free to use his bully pulpit as he wishes, but I suspect it was just instinctive & a build up of his frustration with Intel's handling of this whole matter.
- enriquto 9y agoThe first is the personal attack when he calls Intel engineers morons. That 1 little unnecessary addon is uncalled for because it's quite reasonable that there's a lot of politics going on (...) can't be fixed properly without delaying & impacting revenue in a way that execs won't allow. If that is really the case, then calling the engineers morons is actually very much called for!
- vlovich123 9y agoNote this is a hypothetical since it's quite possible it's not the case, in which case even by your logic the moron label is incorrect. So in your opinion engineers are unintelligent because they listen to the demands of people who can fire them or perhaps, god forbid, they're trying to balance technical solutions against revenue? That seems perfectly rational reasoning to me but that must be because I too am a moron.
- hartator 9y agoIt’s not hypotycal, it’s the part of hardware that made Intel CPU vulnerable to last week attack. Linus was right.
- vlovich123 9y agoI highly recommend reading the rest of the thread beyond just Linus' complaint. This isn't about Meltdown but Spectre which impacts all CPUs with speculative execution. I believe these patches apply to AMD too even though Intel engineers are posting them as they appear to be generic x86 code. Not only is David brilliant at defusing that situation, he provides quite a bit of context to clarify confusion for external readers while correcting Linus' potential misunderstanding of the patch too. The "moronic" patch Linus is complaining about is to fix Skylake that still is impacted by a potential exploit even with the though the other less ugly fixes fix the majority of chips. David even agrees with the unfortunateness of the whole situation & says he's OK dropping that last patch if Linus wants because the Skylake vulnerability is less clearly exploitable. Ingo provides a follow-up suggestion about a potentially cleaner & simpler way to fix just the Skylake issue which is great & the whole point of how RFCs are supposed to work. So not only does it look like Linus' may have overreacted, it didn't even move the conversation along in any way (angry responses rarely seem to).
- Izkata 9y ago> it didn't even move the conversation along in any way (angry responses rarely seem to). From your other comment [0] it sounds like it worked perfectly. The angry response first got the Intel engineer to defend their work and explain it better than they originally had, then further got others interested enough to improve it. [0] https://news.ycombinator.com/item?id=16216289 https://news.ycombinator.com/item?id=16216289
- mmjaa 9y ago>The problem is that his approach isn't going to yield great results. Its worked so far.
- nnq 9y ago> did his finger to nVidia in any way improve relations between Linux & nVidia Purely anecdotal evidence, but in my experience the few pieces of hardware that I can generally bet they'll be fine with Linux are those branded Intel or NVidia (one may argue about the "optimus" thing, but if you want CUDA on a Linux laptop, nvidia's drivers do their job fine in the end, and this is what most Linux users will care about I guess) Dunno the mechanics of it, but some things did work up just right about this :)