New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Not long after you get USB compatible ATAPI drivers.
https://www.theguardian.com/technology/2018/jan/05/intel-class-action-lawsuits-meltdown-spectre-bugs-computer
Seems like the 6 month delay, will cost a bit.
Yet, some rainy guy here tried to paint me as a boring conspiracy theorist...
And yes, sure, it's pure coincidence that that intel ceo sold his shares as soon as they knew about the cluster fuck. Who could doubt that.
Well, you do have some bizarre ideas:
Ha! Yes...we should have the government reviewing and certifying our technologies. They can tell us which ones are safe and which ones aren't, and they'll have the power to "brutally punish" people who publish what they feel is unsafe...my God, you must have loved the Clipper Chip.
So how do you imagine that working? You have a team of government scientists who relentlessly analyze code and then as soon as they find a flaw ("whenever a fuckup is detected"), the corporation is immediately put out of business (I mean, we're "brutally" punishing them, right?) This means all employees lose their jobs, impacting tens of thousands of families, local governments lose their tax bases, consumers and customers lose any warranty support for their products, whatever progress was made in those labs is lost...?
Well, don't fret...fortunately for you, the government is doing exactly as you describe. The judicial branch will preside over the coming lawsuits, and will dole out a billion dollar class-action lawsuit against Intel. Though I guess that probably isn't "brutal" enough for you...
Based on what we know at the moment, this is not like Volkswagon intentionally cheating environmental standards. Brutality may be appropriate there, but in this case, Intel just screwed up. I guess in your world, since every "detection" of a "flaw" would be punished "brutally", all innovation would grind to a halt as everyone would be terrified of making a mistake and we'd all live in a North Korea-like state.
Which corporate bosses are you talking about? The stockholders? 70.3% of Intel stock is owned by 2,545 institutions, representing millions of pensioners and mutual funds with small investors. When Intel pays the inevitable billion-dollar class action suit, they will indeed lose multimillions as Intel's balance sheet and stock declines. So...happy?
But in your world it is, right? You're up high on the hill saying (over and over here) that all beneath you is insecure, with a black-and-white smugness. Why, there's not even any point in patching x86-based OSes because the whole design is suboptimal, according to you. Why even discuss secure methods of coding in php because my God, php is interpreted so that's insecure, right?
It's a one note song. Great, dude, go play with your ultra secure systems. We're running Debian Linux VMs on OpenVZ. That's what the forum is about. Your "pffft, x86, how provincial" snobbery servers no purpose.
No, it's an economic choice. And it's hardly a rigid as you imply, as there have been plenty of cases where large ecosystems sacrificed it in favor of a different approach. Apple used to run on Motorola, now runs on Intel. Windows NT used to run on Alpha, MIPS, and PowerPC. HP-UX ran on PA-RISC, now on Itanium. Etc.
Security is not the only factor out here in the real world
I guess you missed news of AMD64.
It's worth noting that Intel "created an utterly poor spec" in 1978. I know you have a "one mistake and you're brutally crushed" world view, but the opposite side of your coin is that it's amazing it's endured for so long. It's like saying ipv4 was an "utterly poor spec". Sure, maybe it should have been 64-bit from the start...but it has proven to be remarkably durable.
I changed it, bro.
My Jitterbug phone is OK tho.
That's neat, but ARM is also affected?
The Jitterbug Flip is an ultra-simple-for-seniors flip phone...no idea what the embedded OS is but it's basically a toaster since it only does phone calls. I don't think you could even patch the OS without taking it apart. Heck, I'm not sure it even has an OS...well, probably does just to use common components but it's essentially a circuit.
But now that I google, I see that Jitterbug also makes smart phones for seniors and they appear to be Androidish...oh dear...
You know, I was talking with Marge while we were having our nails done and she has one of those new i-whatsie-pod-phone-things and she says a North Korean hacker can just come right in and steal her identity and everything. Can you imagine that? Thousand dollar phone and just like that. I'm sticking with my Flip...
Only some ARM chips. Specifically, those with speculative execution.
We have governments/systems that do punish murderers, rapists, thieves, fraudsters, and so on. Unlike what you try to make believe that's normal.
Funny btw. how you (rather clumsily) try to bend things, e.g. saying "what [people] feel is unsafe" - Wrong. The current cluster fuck is not something some people feel; it's reality.
No, of course not. What I imagine is what is commonplace, usual, and normal in many other areas in our life. One example is that companies who know about some serious danger must inform about that. If they don't they (should, and sometimes do) end up in court.
Actually it's very simple and very normal. In fact, most legal systems are based to a large degree on punishment. Evidently a fine of e.g. 1000$ that is a severe punishment for an average citizen is but a joke for e.g. an intel ceo. Maybe I should have written "painful" rather than "brutal".
BULLSHIT! This is like "intel knew and anyway launched a new product - with that very problem - and their ceo sold his shares".
_[your picture is little to do with my argument and lots to do with your absurdly biased perception]
_
Like often you evidently did not even understand my argument. What exactly is your problem to understand the difference between "bosses" (like ceo) and "stockholders"?
But as you ask: No, I'm not happy. While that might help somewhat by having lots of angry stockholders hunt the intel bosses I'm not happy because those stockholders are not the culprits.
And once more you are arguing more against your absurd perception (i.a. of me) than against me.
If I can get to a certain destination, say a friends house, either by using a reasonably safe route or by using an utterly unsafe route and I suggest to avoid the utterly unsafe route I'm a smug elitist? Strange world that you live in.
As you mentioned php as an example: What I did was to - multiple times! - mention a concrete, open source, and available almost 1:1 alternative to php ("hack"). Evil, smug me!
Allow me to replace that with an at least true and honest statement: "You stupid, clueless, evil spirited fuck!"
I did - here, in this thread, and today, write that I will (not only stick with x86 systems but even) not apply the patches because imo the gain in security isn't worth the performance loss.
Like so often you don't argue against me but against your absurd perception, you do not even properly read (or bluntly ignore) what evil me has actually said, and are, Pardon me, way to clueless anyway.
Clueless and stupid, reliably. For a start, all them funny architectures are hardly a fart next to x86 in terms of significance. Moreover: apple is pretty much closed and tightly controlled system. As long as the users get their apple experience, most won't care about the cpu. And windows was and is 99+% x86 and that did not change.
More importantly, though, you simply missed the point. If intel left compatibility to itself (its older stuff) almost nobody would buy that processor. The test is simple: Does it flawlessly and withourt hiccup run windows, linux, photoshop, firefox, microsoft office? To have "yes" to be the answer the processor must be compatible, period.
bla bla bla. The reality is that x86 is almost exclusively defined by intel and amd has to stay within that frame.
Technically, the exact same assertion is completely true of Intel chips, too...
Not disagreeing, but x86-64 is basically AMD’s baby. Not saying they weren’t restricted by x86 compatibility, and the current issues probably work on x86 32-bit. But AMD kinda defined the 64-bit spec.
A9. I know A7 doesn't, but I don't know about A8 offhand.
Precisely.
Sure. But an x86-64 isn't an entirely different processor; it's, to put it bluntly, but an extended x86. And of course it has to also carry to full load of historic plunders with it.
Btw, it should be noted that I was fair to intel by saying that they, too, had and have to carry all that shit with them. It's just that they are to a very large degree the source, the creator of those plunders.
In a somewhat painful way the current cluster fuck demonstrates well what I say. Every OS on x86 (32 or 64 bits, no matter) of course is acting in, and hast to adapt to and to work within and with that space largely defined by intel. So, unlike what many might think, yes, even minute details like how spec exec is implemented are defining compatibility.
Think a moment about that! Because it means that, just assumed intel wanted to get rid of old shit and did change such micro-details, pretty all OSs would go havoc.
Pardon me but amd's freedom was comparable to that of a ford car paintshop ("Any colour as long as it's black").
Not really, 64-bit spec is not something radically different, and speculative execution can be 32 bit too.
https://developer.arm.com/support/security-update
Thats quite a few. + Apple ARM is also affected by meltdown+spectre. No idea which Apple ARM's though.
About the whole "it's all Intel's fault even though AMD made the 64-bit thing" stance...
I'd just like to throw a reminder out there that if Intel had had their way, today instead of using amd64 with its flawed speculative execution, we'd be using an x86-incompatible 64-bit ISA where all of the out-of-order execution is controlled by the compiler, making any potential bugs like the ones we're facing a matter of software only.
Uhm, Itanium wasn't meant to be intels 64-bit x86 (but some kind of high end product for the "professional market"). Itaniums "controlled by the compiler out-of-order execution" is something completely different from what we discuss here.
The flawed spec exec mechanism you (justifiably) complain about wasn't amd's doing. It was designed by intel many years before amd64 and any x86-64, no matter whether from intel or from amd, had to carry that along (for compatibility).
@bsdguy- your sig only handles for 15 to 30% loss of "speed". Please correct for 25-53%.
ok guys this thread is out of control.
Temporary ban to @bsdguy for the fucking long reads? Do you have that much time?
Smartass! My processor is too slow now. I'm currently trying to move the cursor to the first '('... it'll be there eventually. Maybe that's the punishment for being cheap and getting only 32GB memory.
Nice to meet you, too. Btw: Just don't read my posts. Simple and efficient solution.
That is NOT what I am seeing. Where did you take that minimum of 25?
Any x86-64, yes, meaning some kind of thing that kept compatibility with x86 while adding 64-bit extensions. But I'm really unconvinced that Intel's original plan wasn't to let the radically different IA-64 takeover as the "universal" ISA once the time was ripe enough for that, except that they were preempted by AMD coming up with the alluringly compatible amd64...
After all, Apple's two incompatible processor transitions showed that it can be done, with software-layer compatibility only... and with some degree of annoyances to the users, sure, but Macs are still alive and kicking.
Because 10% of the marketplace drives the entire market.
I made it up.
I'm honestly not sure how that follows from, or is even related to, what I said. "Showed it can be done" somehow equates to "drives the entire market" now?
Nope. As for apple, that's a pretty much closed and controlled eco system with, Pardon my being frank, strange utterly clueless and very "experience"-driven users/customers who would accept pretty much anything if only it had an apple logo on it and was expensive. You can't transfer that to the general market.
As for intel, Nope again. It would have simply been stupid from intel. Think about it: they had the vast part of the total computer market and no reason to play with that ("never change a well running machine"). It would have been much more attractive and sensible to extend their market and to enter the attractive "professional" (and non commodity) turf of sun and ibm.
Lastly you seem to have overlooked a factor that is vitally decisive: software. It is - and was - the plethora of software available for x86 that was the real treasure. Had intel positioned the itanium as the 64-bit successor to the x86 they would have a) pissed off a very significant part of the customers and b) taken the high risk that e.g. amd came up with a x86 compatible 64 bit version of the x86 which in turn might have very badly damaged intel; intel would even have lost their own turf.
Ditto.
Such a contribution.
I patched the first part on all my boxes, pti enabled, ibrs and ibpb seems to be disabled - despite having the recent microcode loaded and the recent CL7 kernel which according to changelogs fixes all 3 of them, so.. either microcode for the CPU's (E5-v4) isn't yet available, or this is jackshit.
However, I see about no performance hit with pti at least, no increased response times nor increased CPU load in a rather high traffic shared hosting environment.
If ibrs and ibpb whenever available will cause a hit.. We'll see :-D