CPU Vulnerabilities, Meltdown and Spectre, Kernel Page Table Isolation Patches, and more

Discussion in 'Hardware Components and Aftermarket Upgrades' started by hmscott, Jan 2, 2018.

  1. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    4,725
    Messages:
    12,174
    Likes Received:
    2,227
    Trophy Points:
    631
    Again, I agree with your statement overall.

    The quoted text below is where we differ. I don't believe that what they will say is fixed will be true either way. Nor will it stop further exploits from being discovered weeks, months, years and decades from now.

    There is not a lot of choice going forward in 2019 as to whether or not to use an insecure platform like a computer (any computer).

    I will keep buying based on performance (Intel still rules there, even after waiting for Ryzen to catch up for a couple of years) and the security will be handled internally as usual (in my terms and with my rules).

    Awareness happened around the '70s. The only thing that is changed today and for a few years now is that a larger %age of the general public is aware (at least on some level) of the new(er) exploits too.

     
    Riley Martin likes this.
  2. ajc9988

    ajc9988 Death by a thousand paper cuts

    Reputations:
    1,501
    Messages:
    5,606
    Likes Received:
    7,935
    Trophy Points:
    681
    First, that was general advice. Second, Intel and AMD have mentioned fortifications and corrections to KNOWN vulnerabilities. Switching to unknown makes no sense here, as it is meant to over generalize and say buy now, which isn't proper for all.

    Third, and least important, Zen 2 has already shown to have caught up at stock clocks with the 8-core demo while using significantly less power and will drop mid-year. Do what you like, but facts are facts.

    Fourth, awareness isn't just hippy dippy b.s. It allows for an evaluation of your deployment environment, for an evaluation of risks, and to make an informed decision. It allows for monitoring if the risk is accepted, to a degree and only applicable to some vulnerabilities, but can lead to general monitoring for odd behavior.

    Fifth, unless you are a coding guru that writes microcode and firmware, as well as modifying OS kernels, etc., no one truly does security on their own. They can make decisions about deploying patches for vulnerabilities, which goes to knowing your deployment environment and risk tolerance, but they alone cannot do all security alone and are at the whims of hardware manufacturers and software vendors to various degrees.

    Sent from my SM-G900P using Tapatalk
     
    Riley Martin likes this.
  3. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    4,725
    Messages:
    12,174
    Likes Received:
    2,227
    Trophy Points:
    631
    My post was general advice too. Take it for what it's worth. ;)

    AMD and Intel will never mention corrections to unknown vulnerabilities. No company does. That is why my long term viewpoint is valid. The exploits will keep coming. Regardless of which brand is the popular favorite today. And why the 'unknown' is also relevant here too.

    Zen 2 doesn't interest me in my workflows. Too slow. Lower power may have been a factor, but dropping mid-year? Am I expected to take an extended hiatus while AMD keeps ramping things up? lol...

    Yeah, not a coding guru. But microcode and firmware is not the end all and be all of the security within a company either. Like I've stated previously; anything digital is up for grabs by someone just a little bit smarter than yourself.

    Anyone is fooling themselves that they're making an informed decision today when it comes to security. I would be surprised that even the great coders working at the big firms saw the exploits found in the last few months to a year.

    To take an analogy from a cash flow business problem. The cash is always the problem. Just like the potential data is in the digital domain. Seems we agree that constant monitoring is the only 'solution'.

    At the same time, you can only look for what you've been taught to see. ;)

    Thanks for the exchange.

     
    Riley Martin likes this.
  4. Papusan

    Papusan JOKEBOOK's Sucks! Dont waste your $$$ on FILTHY

    Reputations:
    22,061
    Messages:
    23,308
    Likes Received:
    40,325
    Trophy Points:
    931
    Some info for you. But I don't know what he used for testing https://forums.mydigitallife.net/threads/windows-10-hotfix-repository.57050/page-378#post-1507000
    upload_2019-3-8_11-28-40.png
     
    Riley Martin and Robbo99999 like this.
  5. Robbo99999

    Robbo99999 Notebook Prophet

    Reputations:
    3,929
    Messages:
    6,514
    Likes Received:
    5,563
    Trophy Points:
    681
    When I have time I'm gonna update my laptop 2920XM Sandybridge CPU with that retpoline registry fix that manually enables retpoline and I'll post back with my performance findings.

    I have actually done the same thing with my desktop Skylake 6700K CPU, and even though you can't have retpoline on it (not compatible with Skylake & later CPUs), but it does enable "Import Optimisation" which sits alongside retpoline as the two changes they've used to combat the Spectre related performance losses. I've done some testing on my desktop, and the "Import Optimisation" does increase performance a little, but it's varied. I've seen the highest Firestrike CPU score I've ever seen since Spectre fixes came about, but it's only a small 0.3% higher than the previous highest score I've seen. I did some BF1 gaming and I was taking note of the fps while on the Amiens map (often CPU limited on that one), again I saw the highest fps in certain areas than I've ever seen before since Spectre - I can't really provide a totally accurate figure, but I did see it in the 140's rather than the 130's in the opening section of spawning in Amiens - hard to be absolutely sure though because each time I play that map I don't always look at the fps. CB15 didn't show any improvements with "Import Optimisation".
     
    Papusan likes this.
  6. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    6,407
    Messages:
    19,715
    Likes Received:
    24,511
    Trophy Points:
    931
    I've been doing real IP security monitoring and mitigation since the first IP/TCP internet debuted in 1981 - implementing protocol source at the kernel level on UNIX - VMS - Tops 10/20 - and dozens of other OS's since, at the firmware and OS configuration level for most every brand of communications hardware involved in the entire communications chain of data, and I was active on the internet before then since 1978 before the TCP transition from NCP.

    And, I've continued that work for a wide range of companies, from small startup's to huge multi-national corporations and governments on worldwide networks at the hardware and software level. Designing and debugging at the hardware and software level, pre-configuring and preventing security issues throughout all of those implementations.

    In comparison, and to be kind, your comments and @Robbo99999 's comments in this thread stand out in sharp contrast to what I see from other peer professionals, and makes what you two write often look to me like infantile gibbering garbage.

    When your comments are directed at trying to make yourselves look good rather than actually trying to solve problems and help people become aware of problems, it's a further waste of our time.

    You two, and a few others, repeatedly waste our time trying to downplay the actual serious issues at hand, to the detriment of others security concerns, wasting our time posting useless meanderings instead of contributing something constructively useful.

    My perspective is much more serious and matter of fact based on the certainty of real world experience - how things look when discovered, how they look as they progress into real threats, and how they look as captured in real time as they happen to live systems, over days, weeks, months and years, transitioning to long term support for mitigations into decades of time.

    Please stop filling the thread with useless posts that dilute the useful information with garbage no one seriously interested in this thread topic wants to wade through. We don't want to hear BS that tells us "not to worry about this or that" because of some hair-brained guesses.

    You simply don't know what you are talking about and we don't need to hear your mutterings to yourselves why it's all going to be ok and there is nothing to really worry about. There certainly is, and you need to stop ignoring it like it doesn't exist.
     
    Last edited: Mar 9, 2019
    Riley Martin likes this.
  7. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    6,407
    Messages:
    19,715
    Likes Received:
    24,511
    Trophy Points:
    931
    These problems could be from coincidental changes unrelated to Retpoline - or accidental side effects of changing the associated code, as the retpoline code isn't supposed to be active by default, and should only be able to be enabled manually, which most gamers will not have done - yet they are still seeing performance problems with the patch installed.

    It's sad when fix patches create new bugs, even when the fixes aren't active, so until MS gets this patch fixed it's probably best to forgo this retpoline patch by uninstalling it.

    Microsoft Warns Latest Windows 10 October Update Can Cripple Game Performance
    by Brandon Hill — Thursday, March 07, 2019
    https://hothardware.com/news/microsoft-windows-10-october-update-cripple-game-performance

    "Microsoft released KB4482887 for Windows 10 version 1809 (October 2018 Update) on March 1st, and the headlining feature was the inclusion of Retpoline performance improvements for Spectrevariant 2 mitigations. But almost immediately, gamers began reporting problems.

    In one reddit thread, numerous users complained about performance problems in their games after installing KB4482887. A redditor by the name of rayw_reddit wrote:

    This patch is causing massive lag spikes in older games, like CoD4 and CoD MW2. Right after installing this update, I launch any of the two aforementioned games, moving the mouse around (yes, mouse movement) causes the game to freeze in 1 second intervals every time. If you don't move the mouse, game appears fine.

    The above gamer is running a pretty stout system with a Ryzen Threadripper 2950X and a GeForce RTX 2080 Ti.

    "I’ve got this same issue with an RTX2080 and Ryzen 2700X," added maverick26290, "Destiny 2 has some insane lag spikes when you move the mouse.

    Also uninstalled this update and all is normal again." Yet another user, jenders37, wrote, "Having issues with this on several of my games as well. Destiny 2 was a NIGHTMARE, but as you said it resolved when removing the patch."

    Hopping on over to the Destiny subreddit, it's more of the same, with jlobue10 writing:

    I logged in to play some yesterday prior to Season of the Drifter starting this Tuesday and noticed my game was unusually sluggish. It also felt as though my mouse setting and sensitivity were messed up even though I had been comfortable with the settings I have for quite a long time now.

    Now, Microsoft has officially chimed in an acknowledged that there is indeed a performance issue with KB4482887. The company confirms:

    After installing KB4482887, users may notice graphics and mouse performance degradation with desktop gaming when playing certain games (eg: Destiny 2).

    At this point, Microsoft says that it is working on a fix that would restore performance for affected gamers. In the meantime, Microsoft is suggesting that gamers who are experiencing issue uninstall KB4482887 for now, after which any performance lagging should be eliminated.

    Microsoft has not said what is causing the problems, but many are speculating that it could be in relation to the Retpoline tweaks that were implemented with this latest cumulative update."

    Another article here:
    Microsoft warns of KB4482887 game performance issues
    http://forum.notebookreview.com/threads/windows-10.762434/page-556#post-10878109

    [​IMG]
    Remember, retpoline is really only attempting to claw back Windows performance lost due to patches for Spectre / Meltdown vulnerabilities. Felt worst on older gen CPUs, the original Spectre / Meltdown patches, resulted in slowdowns particularly in disc and network access. But now we see that the magic bullet dubbed KB4482887 might cause users to "notice graphics and mouse performance degradation with desktop gaming when playing certain games, such as Destiny 2."

    ...Looking around Reddit and Twitter one can't be certain if games other than Destiny 2 are affected, but it seems likely. In the meantime if your PC is adversely affected and you can't wait for a 'resolution' from Microsoft, it is OK to "uninstall KB4482887 to regain performance," says Microsoft in the patch release notes."
     
    Last edited: Mar 9, 2019
    Riley Martin likes this.
  8. Robbo99999

    Robbo99999 Notebook Prophet

    Reputations:
    3,929
    Messages:
    6,514
    Likes Received:
    5,563
    Trophy Points:
    681
    That comes as a big surprise, you don't seem mature enough to have that kind of experience. And you talking about clogging up threads with useless posts, yet you constantly spam other threads with anti Intel/NVidia and pro-AMD postings to the point that it's just spam, and also combined with your real need for having the last word. I'm surprised to hear your experience given your 'behaviour' on here...but it is what it is. I'll continue to post my own viewpoints on the topics in these forums, and in this thread too.
     
    Last edited: Mar 9, 2019
    Ultra Male and tilleroftheearth like this.
  9. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    4,725
    Messages:
    12,174
    Likes Received:
    2,227
    Trophy Points:
    631
    Well said. Nothing more to say from me.

     
  10. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    6,407
    Messages:
    19,715
    Likes Received:
    24,511
    Trophy Points:
    931
    I am just as serious with my posts and comments about Intel, Nvidia, AMD, as I am here.

    The amount of posts coincide with the amount of information being discussed with new releases, reviews, and owners feedback. There is a cycle to new product releases, and at times there is far more information being exchanged publicly and privately than others.

    The "last word" is your imagination, as there are always ongoing discussions around the same subject going on all the time, even after my last posts.

    The Pro-AMD comments are trying to keep AMD in the conversation while Intel and Nvidia are flooding the mind space with their marketing BS and deceptive spin. If thoughtful and objective people that consider all options and motivations for purchase didn't do that then Intel and Nvidia would have brainwashed everyone.
    Sounds good, thanks to both of you for ending this useless OT dialog in this thread.

    Please let's keep the security discussions serious, constructive, and on point as they should be for the sake of their scope and effect on people looking for information and help here.

    There is plenty of Intel BS trying to minimize the real situation and try to downplay the problem, while caring nothing for their customers security exposure - only as far as they are forced to by the level of awareness and demand for fixes from their customers.

    Raising the awareness and providing information is what we are trying to do here and in other threads here on NBR so members can know what's what when their only other source of information is Intel and Microsoft.

    Posting minimization's and poo-pooing the potential effect of exploits not even out of the gate - before we know their full effect - so as to drive it out of the mind for real thought and consideration, is a disservice and a waste of everyone's time.

    This threads discussion addresses the vast majority of computers used today, and wasting time mentioning air-gapped or otherwise "unassailable" computers as being unaffected is wasted discussion. As it is well known that "unassailable" vectors are also compromisable.

    As with all security exploits access is key to deployment, and there are few if any consumers with computers that aren't connected to the internet, or don't use USB or other media - or the internet - to install software, all vectors providing access for infection through unpatched or unknown exploits.

    All successful malware coming through browsers use exploits unpatched or unknown, which can deliver a payload that contains other unpatched or unknown exploits - including these new to public disclosure discoveries which may not be new at all, and in fact might already be used in the field.
    * Here is the original post I made, 22 posts back:

    http://forum.notebookreview.com/thr...atches-and-more.812424/page-120#post-10875985

    SPOILER alert, literally: Intel CPUs afflicted with simple data-spewing spec-exec vulnerability
    'Leakage ... is visible in all Intel generations starting from first-gen Core CPUs'
    By Thomas Claburn in San Francisco 5 Mar 2019 at 06:34
    https://www.theregister.co.uk/2019/03/05/spoiler_intel_processor_flaw/

    148 Comments

    Note: "The issue is separate from the Spectre vulnerabilities, and is not addressed by existing mitigations. It can be exploited from user space without elevated privileges."

    Updated: Further demonstrating the computational risks of looking into the future, boffins have found another way to abuse speculative execution in Intel CPUs to steal secrets and other data from running applications.

    --------

    Intel CPUs Reportedly Vulnerable To New "SPOILER" Speculative Attack
    Written by Michael Larabel in Security on 5 March 2019 at 09:28 AM EST. 66 Comments
    https://www.phoronix.com/scan.php?page=news_item&px=Intel-SPOILER-Attack

    "SPOILER is the newest speculative attack affecting Intel's micro-architecture.

    Researchers out of the Worcester Polytechnic Institute and University of Lubeck discovered this new speculative attack dubbed SPOILER, Speculative Load Hazards Boost Rowhammer and Cache Attacks.

    Intel was notified of this issue a few months ago but no software/hardware fix appears ready yet, while the researchers claim there might not be an effective software solution available at least anytime soon -- and any mitigation would likely come at a performance cost, as we've seen with Spectre and Meltdown over the past year.

    AMD and ARM CPUs aren't believed to be impacted by SPOILER.

    The SPOILER research paper can be read here:
    https://arxiv.org/pdf/1903.00446.pdf

    66 Comments
    ------

    SPOILER alert, literally: Intel CPUs afflicted with simple data-spewing spec-exec vulnerability
    submitted 5 days ago by alexeyr
    https://www.reddit.com/r/programmin...er_alert_literally_intel_cpus_afflicted_with/
    https://www.reddit.com/r/hardware/comments/axi1pn/spoiler_alert_intel_chips_hit_with_another/
    https://www.reddit.com/r/intel/comments/axjk51/spoiler_alert_literally_intel_cpus_afflicted_with/
    https://www.reddit.com/r/Intelligen...er_alert_literally_intel_cpus_afflicted_with/
     
    Last edited: Mar 12, 2019
    joluke, Riley Martin and ajc9988 like this.
Loading...
Similar Threads - Vulnerabilities Meltdown Spectre
  1. Starlight5
    Replies:
    14
    Views:
    1,022

Share This Page