Computers threatened by getting slowed down by up to 30% due to necessary patches... [Archive]

Skybird

01-03-18, 07:40 AM

... of the Kernel in Intel CPUs of the past 10 years.

The bug, a Kernel memory vulnerability, is at the lowest chip level and allows attackers to directly access a normally hidden part of the chip where passwords and other elemental nformation like this get stored. Intel cannot solve it on hardware level, so rewriting the kernels for Linux and Windows is inevitable. Right this is currently being done, in great secrecy.

If things do not go well, users can face really substantial losses in speed and performance that cannot be compensated for by other means.

This bug is pretty much a worst case scenario that has become real. I wonder whether newest 8th generation CPUs are affected as well. if so, starting to sell these new chips just weeks ago while this bug already is known since months, to me oversteps the line to criminal acting and fraud. 8th generation intel chips are beign sold since Novembre 2017.

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I post this here and not in the PC subforum, since I think the importance and relevance for all intel users justifies to spread this a bit beyond that subsection of the forum.

Note that Linux users are as affected by this as are Windows users. Status with Apple users is unknown.

Users of already hopelessly messy Windows 10 and its enforced patch beta testing program for everybody, can only pray.

Its my habit since many years to NEVER store any login information for anything on HD, I always manually log in, I have all options for auto assists, auto login and auto fill off, and I manually control frequently the deletion of all temporary folders known to me.

I also have private data files like images, texts, photos, on external HDs and USB sticks and DVDs, but not on HD. I connect them when I want to access them, disconnect them immeidately when they are no longer needed - and clean the HD and reboot afterwards. I did not know of this bug, but I am not surprised that somethign like this shows up. I just waited for "safe features" like the above described things getting compromised.

I used to do some brokering stuff online until 2016 when I finsihed by withdrawel from all that, completely. I would not broker online anymore these days. But professionals's brokering still takes place via their computers. Its all a mess. Its frightening, absolutely frightening how vulnerable we all have been forced to become. And many of us still are stupid enough to apllaud all this and call it their comfortable life. Its a big gamble you make there and at stake is nothing else than the economic survival of the individual.

We need cashless payment, and all people havign all their property digitally, and everybody being online and in the cloud! Lets do it! Its safe. When will people stop to allow getting fooled by stupid lies and cheap fraudulent promises like these?

Catfish

01-03-18, 08:32 AM

[...]
We need cashless payment, and all people havign all their property digitally, and everybody being online and in the cloud! Lets do it! Its safe. When will people stop to allow getting fooled by stupid lies and cheap fraudulent promises like these?

No, buy houses and cars even if you cannot afford it! And when all breaks down blame it on the government, never on your own behaviour! ;)

Dowly

01-03-18, 08:51 AM

Oh look, another thread for Skybird to preach about the dangers of all things digital, ironically he does this on an internet forum, and nope Windows doesn't get a pass this time either as he moans about Windows 10 for a Intel's mess.

For this and all the countless of other blabberings, I bestow to you, Skybird, the honor of being the sole inhabitant of my ignore list. Bye! :salute:

Skybird

01-03-18, 09:27 AM

Its a flaw that Intel cannot cure and that must be tackled by redoing the Kernel at its very lowest, most fundamental level, apart of it that usually gets acitvely hidden from all woreking software and incoming traffic, since it is most sensitive. The articles describe why this issue is so dangerous.

The fixes, if they will be fixes indeed, need to be distributed via security updates, most likely, for Windows and Linux. And Microsoft has a well documented, empirically well established reputation of - since being "turned into a service" - enforcing consumers and company customers alike to swallow many ill-designed, broken, misfunctioning "patches" that you can delay for 30 days in some higher W10 versions, but cannot avoid . Your computer is owned by Microsoft. They do not ask you anymore.

From some time on, you must pick it up and suffer from it - even if you know and there is verified documentation of that the given piece of "patch" damages your machine, your work flow, your company productivity. The past two years have been over and over full of such events. That is the bad truth about the new enforced cumulative patching policy by Microsoft. Its a mess.

And thats why I call them out over this.

The speed losses in testing of the Kernel fixes so far have shown to be in the range of 5-30%. Newer chips seem to be less heavily effected and get away with lesser speed losses, but losses nevertheless.

Like it or not, guys, but if in the near future you are one of those noticing performance losses on your Intel-machines or Apple computers using intel chips, you now know why that is so, and whom to thank for it. It'S Intel's fault, and Microsofts, Linux and Apples varying remedies, since Intel cannot fix this: its a fundamental design flaw.

AMD users seem to be safe from this mess.

Trust in all this blindly, Catfish and Dwoly, trust in all this blindly. "Digital" always and automatically means "better". Unconditionally. And you look so cool and are so much "in" with being digital, the girls and your clique love it.

Skybird

01-03-18, 09:55 AM

And a correction, or better: an addendum: the Linux fix seems to aim at fixing AMD chips as well, but I do not find further info on that, so it is unclear whether that is to an unreported vulnerability of AMD chips as well, or just a precautionary measure.

Skybird

01-03-18, 11:50 AM

More info, thanks to Rotareneg at SBP forum for linking it:

https://www.reddit.com/r/sysadmin/comments/7nl8r0/intel_bug_incoming/

Most people are expected to be hit by losses in the one digit percent range. But in tests some get hit really hard, up to -50%, depending on the running tasks.

This is a real nasty way to start the year. Cloud computing: compromised. VM: compromised. The Linux patch seems to not have much effect on gaming under Linux, but effects on DirX remain to be seen. Certain other tasks under Linux however need to digest a massive blow in performance.

Mr Quatro

01-03-18, 02:01 PM

:oAbsorb Sky's good and leave his bad ... he's still a useful member of subsim.

What do you recommend Sky for a personal carry around USB device?

Rugged, shock proof, water proof etc 64gb min or perhaps a small ssd?

I like this one: SanDisk 128GB Ultra Flair CZ73 USB 3.0 Flash Drive, Speed Up to 150MB/s (SDCZ73-128G-G46) $37.00

never mind that one it has bad reviews I like this one better: Samsung 128GB FIT USB 3.0 Flash Drive, Speed Up to 150MB/s (MUF-128BB/AM)

I posted too quick they both have bad reviews ... I can't afford to lose my data :o

Skybird

01-03-18, 06:34 PM

HD is the best storage device (even before CD/DVD), because even if it breaks down, you still have a good chance that data can be retrieved by a specialist, if it is vital for you. USB sticks tend to break the moment when you can need it the last. SSD in principle are big, faster USB sticks. I mean, flash ROM is quite vulnerable. I do not really trust it.

If you do not desperately depend on fast speed, a HD with solid USB 3.0 connection.

And backup copies on a secondary, depending on your data value: tertiary device, DVD, stick.

When it comes to just editing texts and photos, speed is no argument, the data files are simply to small. Even a rewritable CD then will do.

Skybird

01-03-18, 07:11 PM

Google delivers statement saying that CPUs by Intel, AMD and ARM are affected.

https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.html

Skybird

01-03-18, 09:11 PM

https://www.theverge.com/2018/1/3/16846540/intel-processor-security-flaw-bug-response
https://www.theverge.com/2018/1/3/16846784/microsoft-processor-bug-windows-10-fix

Windows patches for ARM, AMD and Intel CPUs are now in the wild.

In fact a whole armada of patches is reported by some admins to come in, outside the usual timeline of Microsoft.

It also gets vrpeported by some people that if AV is not compatible with these patches, they will not install automatically. Woody Leonhard's well-known blog is at Defcon 2, so status and quality of these patches also currently is unknown and unconfirmed.

Its all a damn nasty and confused mess. Currently companies start to point fingers at each other, maybe breaking communication embargos to save themselves.

What a FUBAR mess.

Its important to understand that there can be no repair and no fixing of this problem, it is a hardware-based issue that cannot get fixed - except by replacing all affected CPUs and mainboard chips. Whatever is being applied in software fixings, is and can only be a workaround.

Skybird

01-04-18, 08:35 AM

https://www.youtube.com/watch?time_continue=434&v=_qZksorJAuY

Benchmarking the i7 8700K pre and post update. Gaming seems to be unaffected, but 4k reading from SSD takes a 20-25% dive in speed.

Desktop user sin generla may not see too worrying differences, but for server centers, single effects could add up and accumulate to severe final end effects (performance drops).

For Intel, who already had AMD breathing down their neck last year, this is a nasty kick into their family jewels.

Skybird

01-07-18, 06:58 AM

For the time being: avoid, avoid, avoid.

https://www.askwoody.com/2018/multiple-reports-of-blue-screens-bsods-0x000000c4-when-installing-the-january-win7-monthly-rollup-kb-4056894/

propbeanie

01-07-18, 04:59 PM

... and I keep getting prompted on my Win7 systems to install the dad-blamed thing, even though I have MS updates turned off... Skybird, have you read anything about this affecting older Core 2 Duo machines? All I can find is stuff on the iCore stuff... ?? :salute:

Skybird

01-07-18, 07:02 PM

Wowh, that is an old CPU. I remembere that some days ago there was something on KB4056894 endlessly reinstalling from beginning, but I cannot find or remember anymore where it was, I think it was under Windows 7. But probably not related to that old processor.

What I remember is thnat the guy said he was prompted to install the opatch, but could not hide or suprerss it. Cannot recall anymore whether he was on W7 or W10. He finally got it fixed, somehow, the prompt got superssed/hidden - but he was some IT pro with lots of knowledge.

Windows 7 has gotten increasing problems and troubles with regular updates over the past 12 months or so, however. It seems Microsoft is determined to mess patching up wherever it can.

Sorry, I cannot help.

Skybird

01-07-18, 07:06 PM

I wonder what all the AMD users do who got locked out their systems after installing KB4056894, which makes many AMD systems BSODing on booting - the patch comes as regular Windows monthly quality rollup patch, mind you, not evcen clearly mentioning in its description what it is, just the usual stereotypic empty phrases and text blocks that mean nothing anymore these days.

Gargamel

01-07-18, 08:16 PM

https://imgs.xkcd.com/comics/meltdown_and_spectre.png

Skybird

01-08-18, 08:05 AM

The reports on AMD systems booting right into a BSOD after installing KB4056982, widened and now are not just about Win7 anymore, but also Win10 1709. Most affected are owners of Athlon, Sempron, Opteron and Turion CPUs. But generally all AMD chips tend to be at risk.

If you are owning an AMD, stay away from KB4056982 and make sure your Windows updater does not install it behind your back.

The problem manifestates itself by that when you boot your system warm or cold, you get a BSOD after just one or two seconds. You get an error code either 0X000000C4 or 0X800F0845.

If you own an intel system, stay away from KB4056982 either, at least as long as unvoluntary Beta testing for Microsoft is not your passion.

mapuc

01-08-18, 01:23 PM

I have AMD Radeon on my computer.

I don't know if the blue screen I got some weeks ago have anything to do with this patch

When I got this blue screen, I was expecting to read "Thread stock in device driver" but instead I read

"dpc watchdog violation"

This was a new type of stop code.

Made a search for this and discovered it had something to do with my graphic card.

Markus

Catfish

01-08-18, 01:44 PM

^ this may be related to the latest Radeon card 'crimson' driver update. :hmmm:

Jimbuna

01-08-18, 01:51 PM

^ this may be related to the latest Radeon card 'crimson' driver update. :hmmm:

I could give an alternative potentialcause but then I'd have to infract mesel :)

Mr Quatro

01-08-18, 02:02 PM

My amd radeon card is always trying to get me to update, but I don't respond.

If it works don't fix it ... :yep:

How do you stop Win 7 update? It doesn't ask me it just does it, but not that one sky posted yet ... :o

STEED

01-09-18, 12:52 PM

Microsoft pulls security update that freezes PCshttps://news.sky.com/story/microsoft-pulls-security-update-that-freezes-pcs-11201393

What a bunch of spanners. :haha:

Too busy snooping than fixing. :03:

propbeanie

01-09-18, 03:12 PM

Now, how do they propose a person "roll-back" their computer if it doesn't boot? When I tried one of their fancy "snapshots", the "emergency start-up" couldn't find one... fine job Uncle Bill, fine job... I going to convert the box to Linux anyway. Just doing it a little earlier than planned... My favorite part about all of this is that I cannot find any documentation on Core 2 Duo machines being affected, so my OS did NOT ~need~ an update. I have automatic updates turned OFF... ~HOW~ did it install an update?... but I can see it from the command prompt... I was much more successful in thwarting their Win10 install than I was this.

Mr Quatro

01-09-18, 05:16 PM

MS tried to install KB4056894 today, but it said that it failed ... so thanks to Sky I turned off auto update :yep:

d@rk51d3

01-09-18, 05:48 PM

MS tried to install KB4056894 today, but it said that it failed ... so thanks to Sky I turned off auto update :yep:

Double check. In many cases, the fail message is a fail in itself. :up:

Skybird

01-09-18, 09:02 PM

I still think private users probably do not miss anything if delaying this patch. Its not as if the vuln erabilties intorduced by Meltdowna nd Spectre, of which there are now over one dozewn attack options known I think, can hit orivate users easily. The attacker needs his code on your PC in order to extract that data from forbidden memory sources.

Behave and by that reduce the probability of this code reaching you. That way your risk probably is smaller than if you risk to trust in Microsoft.

Linux Ubuntu and Mint should start to get new Kernels these days. But not even there I am in a hurry.

What I noticed, however, is that on quite soem websites in the past two days, response times by servers have gone way up. Occasional slow traffic at timers odf a day's high traffic phase, that is something one is used to, but this is too widespread and too general currently as if I would accept that reference as an explanation. I think those patches they applied make themselves known by degrading server performance. These patches, if they should have an mitigating effect, necessarily must slow down processor performances, you cannot avoid that, its the nature of this kind of patching that is needed. Private PCs so far do not feel much - server farms however: that is something else, there many small degradations add up.

They have plenty, plenty, plenty of optimization work to do there. Or replace hardware.

Skybird

01-09-18, 09:10 PM

https://news.sky.com/story/microsoft-pulls-security-update-that-freezes-pcs-11201393

What a bunch of spanners. :haha:

Too busy snooping than fixing. :03:

https://www.computerworld.com/article/3246188/microsoft-windows/microsoft-yanks-buggy-windows-meltdown-spectre-patches-for-amd-computers.html

How unexpected. :oops:

Eichhörnchen

01-10-18, 04:58 AM

Catfish

01-10-18, 05:17 AM

^ USB sticks, like SSDs, can store data for a long time; but the problem are changes of the data. If you use such a device for everyday changes, they will fail at some point. Data on such "fixed" devices (or better their memory blocks) can only be overwritten for a limited number of times (like with rewritabled CDs/DVDs, the technology is a bit different but has similar problems).

Best is really to store important data on HDs. The problem here is that the adaptors/ports/connectors change over time. This is due to improved hardware (inside computers back then IDE, SCSI, SATA, next .. what ?), or USB 2, 3, ..next? a.s.o..

So to keep data you have to regularly update the data storage devices, and copy the data to the newer ones. This is b.t.w. a big problem for museums etc., some say our times will become known as the "low information age", because data of this time stored in contemporary devices will not be accessible in a few decades from now – either because the storage containers broke down (like with CDs appx. 20 years of usability, or HDs with appx. 5 years until first errors and later failure) or because there will be no new computers being able to connect to the old hardware.
Microsoft has already given up on downward compatibility in a lot of cases, and this is the only company that at least used to think about that.

Eichhörnchen

01-10-18, 05:25 AM

I didn't know that about USB sticks... I presume you mean after the storage capacity has been reached?

https://i.imgur.com/bO0qBqK.png

This one of mine is not even half full of squirrels, so I guess you mean that once it's full then I can only change the squirrels for new ones a limited number of times?

Catfish

01-10-18, 05:39 AM

[...] I presume you mean after the storage capacity has been reached?

Yes, in a way. As long as it is not full, old blocks will not be overwritten if you add new data. But if you change old data like a new version of a Word document/template or whatever, the old info will be wiped and the new written (changed). And this only works for a limited number of writing events.

This one of mine is not even half full of squirrels, so I guess you mean that once it's full then I can only change the squirrels for new ones a limited number of times?Yes. However if you keep it at 50 percent data volume, and often change the existing 50 percent, it will also fail after the critical number has been reached. It really depends on how often you change/rewrite existing data.

Eichhörnchen

01-10-18, 05:39 AM

... some say our times will become known as the "low information age", because data of this time stored in contemporary devices will not be accessible in a few decades from now – either because the storage containers broke down (like with CDs appx. 20 years of usability, or HDs with appx. 5 years until first errors and later failure) or because there will be no new computers being able to connect to the old hardware.
Microsoft has already given up on downward compatibility in a lot of cases, and this is the only company that at least used to think about that.

Yes I heard a radio programme recently about our over-reliance on current storage hardware which may not be accessible in years from now.

Eichhörnchen

01-10-18, 05:43 AM

Thanks for all that, Catfish. Are you at work?? :haha:

Eichhörnchen

01-10-18, 06:03 AM

Is your boss there? Oh....

Catfish

01-10-18, 06:10 AM

^
a) yes
b) no
But i don't admit anything, also i could be lying :haha:

Skybird

01-10-18, 07:02 AM

Note that what Catfish says about USB sticki memory, is also true for SSDs. Thats why it is recommendable that if yiu use a single SSD for your OS alone, to pick one of comfortably bigger size, not one that just is big enough to carry the OS and its to be expected updates. SSDs have a certain ammount of hidden reserve blocks that get used once the ordinary, open" ones all are used or dead, but they are limited in numbers, of course. You here have the explanation why SSDs shold not be defragmentised. That suffles a lot of data around the memory blocks, and that reduces the overall lifetime of the SSD. Moden OS use the TRIM command instead.

With USB sticks, you may also get low quality of the used components, and they break once the thing gots hot, or power going through them for too long in a straigth time. Simply bad component quality. even well known brands can be affected by this.

And never buy these sticks blindly in a store. Do research and check customer feedbacks, many USB 3.0 sticks that promise high speeds, may indeed deliver speeds in the range of USB 2.0, or even worse, if you are unlucky.

Compare read and writr speeds, and decide on your priorities according to what you need the stick for. If you need it as a long time data safe, low writing times may be okay, but reading times never should send you into the dumbs. If you expect to read and write to thr stick a lot, you need good writing times as well.

Do not buy too cheap if it is USB 3.0.

Also note that USB 2.0 -extension cables plugged into an USB 3 port, will turn every USB 3 stick into an USB 2.0 speed stick. You need dedicated USB 3 extension cables, and these again vary in quality and speed. I had some people who wonderd why their 3.0 device was slow while they attached it ti a USB 3 connector, but via an USB 2 cable. The quality of USB 3 cables also can widely vary in speed, even may not transport all data correctly, especially when the sent signla already is more on the weak side.

Skybird

01-10-18, 07:55 AM

SitRep and assessment of the status quo, outlook.

https://www.riskbasedsecurity.com/2018/01/the-slow-burn-of-meltdown-and-spectre-exploits-lawsuits-and-perspective/

Many very good links in that text.

Eichhörnchen

01-10-18, 10:27 AM

Thanks for that great info, Sky. Is there power going through your USB stick only when you have it 'open' on the desktop, or is power going through it as soon as it's plugged into your pc?

Skybird

01-10-18, 10:32 AM

Does that LED in your stick light up without power? ;) In my understanding every USB connected device is under power when it is plugge din. Even passive pedals of your hardware controller. How else could the OS detect and list them in the hardware manager? Different it might only be with electric devices connected that have an on-off siwtch, say a digital camera. My small one or my smartphone only gets reocngised when connected when I indeed switch them on (from cold and dark, not from standby, for standby means they are on).

I said "when power moves thorugh the device", but mind you: I was referruing to bad quality component. Usually, with good stick that shoudk be no real problem.Consider all the professional software suits that only run with a plugged-in dongle... That dongle is left in by many users even if that software is not run. We would have a flood of complaints, if that kills these dongles in masses. In principle a dongle is just a USB-stick.

Eichhörnchen

01-10-18, 12:21 PM

Thanks... although none of my usb sticks has an LED

mapuc

01-10-18, 12:54 PM

I have a few USB the smallest is 4GB and the biggest is 1TB

The one with 1 TB is connected to my TV (PVR)

Markus

vienna

01-10-18, 03:37 PM

^
a) yes
b) no
But i don't admit anything, also i could be lying :haha:

You could have a promising future in politics... :03:

<O>

Jimbuna

01-11-18, 07:18 AM

My preferred kit:

Kingston 64GB DataTraveler 100 G3 USB 3.0 Flash Drive - 100MB/s

https://i.imgur.com/jzKpnzb.jpg

Skybird

01-11-18, 08:43 AM

I used Transcend Jetflash USB 2.0 for man yyears, and many sticks on various systems by various owners, with 8, 16 and 32 GB. Maybe 15 sticks - and not one fail.

For 2.0 they are no longer build, unfortunately, but I now have two USB 3.0 sticks of this series, wich work since over one year, without problems.

In the past ten years or so, I have lost maybe 8 or 10 sticks by other brands. Lost to techncial malfunctioning, I mean. All were no no-names, but well- known brands.

Its difficult to find lasting, good ones, and you need a bit of luck, too, it seems. The proven brands may fall victim for one charge of weak componentsm, say ROM bars, and the next charge they get from the same sub-contractor then again work flawless.

If you need long-time conservation of archives, lets say family pictures, go with a proven model of HD. And make not one but two or three copies, can be to DVD if file size allows. For long time storage, USB sticks would only be my last choice, also for economical reasons.

P.S. If the speed is not needed, stay away from so-called extreme speed models. Their speed gain comes at a cost: heat. And heat and electronics are no close friends. No every action with a stick lasts long enough to produce that heat, but if you write large files, and many such files in one session, then it can happen. Its a risk factor.

propbeanie

01-11-18, 01:44 PM

https://www.computerworld.com/article/3246188/microsoft-windows/microsoft-yanks-buggy-windows-meltdown-spectre-patches-for-amd-computers.html

How unexpected. :oops:

This would explain my computer... the dual-core AMD "whatever-it-is" core... No known way to recover... So I guess I'll just install Linux on top of it...

The USB, or any "electronic" storage device is very finite due to its physical characteristics. They're like little capacitors that hold a charge, but eventually physically wear out from use. It used to be that you could get one thousand state changes out of early compact cards, but I haven't kept up with what their MTBF rate is now - much less than a HD, but not a BD-RW or similar.

Eichhörnchen

01-11-18, 03:29 PM

Forget your usb stick... go look for your whacking-stick

vienna

01-11-18, 05:54 PM

https://i.pinimg.com/474x/30/02/85/300285f3ddb23774a30e5d5c351cf786--omnia-paratus-stars-hollow.jpg

<O>

Mr Quatro

01-11-18, 06:47 PM

Well this is a subsim :o

Skybird

01-11-18, 07:51 PM

This would explain my computer... the dual-core AMD "whatever-it-is" core... No known way to recover... So I guess I'll just install Linux on top of it...
Welcome to Linux. Show MS the middle finger.

Be advised that if you want to use VM/VBox under linux to get some Windows stuff or games running, many have their VBox currently messed up due to the new Linux Kernels which contain the mitigations for Spectre/Meltdown. At least under Mint I read in the major German Mint forum that several such Kernels yesterday and the day before caused problems with VBox. Probably gets fixed with time as well.

Skybird

01-12-18, 11:25 AM

We have received reports from a few customers of higher system reboots after applying firmware updates. Specifically, these systems are running Intel Broadwell and Haswell CPUs for both client and data center. We are working quickly with these customers to understand, diagnose and address this reboot issue. If this requires a revised firmware update from Intel, we will distribute that update through the normal channels.

https://www.computerworld.com/article/3247788/computer-hardware/intel-says-new-firmware-patches-trigger-reboots-in-haswell-and-broadwell-systems.html

If you use Haswell and Broadwell: don't touch these upcoming or out-in-the-wild firmware updates.

Skybird

01-22-18, 04:36 PM

After patches for broken patches for broken patches by which Microsoft wanted to tackle it sbroken AMD stuff", now comes Intel and tells people to immediately stop upodating processor firmware.

https://www.computerworld.com/article/3250250/malware-vulnerabilities/belay-that-order-intel-says-you-should-not-install-its-meltdown-firmware-fixes.html

This news made my evening. :har: Never again shall I be called "computer paranoid"

BarracudaUAK

01-25-18, 02:13 AM

Thanks for that great info, Sky. Is there power going through your USB stick only when you have it 'open' on the desktop, or is power going through it as soon as it's plugged into your pc?

USB "back in the day" was rated for "500 mah" per port.
Which is why things like "Force Feedback" Joysticks and Racing Wheels needed external power supplies, but a Mouse or Keyboard would not.

Any time you plug in anything into a powered outlet, it has power.
And since most (all?) usb "jump drives" are not switched. It is on.
Whether or not it is "mounted" (i.e. hardware is made access-able via the file system) is independent of the device being "powered".

Note: I may have misread some of the post on this line of discussion, My apologies if this is the case!

This would explain my computer... the dual-core AMD "whatever-it-is" core... No known way to recover... So I guess I'll just install Linux on top of it...

The USB, or any "electronic" storage device is very finite due to its physical characteristics. They're like little capacitors that hold a charge, but eventually physically wear out from use. It used to be that you could get one thousand state changes out of early compact cards, but I haven't kept up with what their MTBF rate is now - much less than a HD, but not a BD-RW or similar.

Base on what I've read, my understanding is this:
(In addition to the reasons Propbeanie listed)

The drive/storage starts with little or no "charge", current is used to flip the "gate" from "off" to "on", i.e. "0" to "1".
Current is now stored in the "gate"/"switch".
"Data" is now stored on the SSD.

(all the "1"s representing a "bit" and that part of the data, 'write' all "0"s to a drive and it is "erased")

When the drive is erased/data is deleted, current is used to flip the "gate" from "1" to "0".
But there is still residual "charge" in "gate".

When new data is written to the drive, it flips the "gate" again.... to "on"/"1".
Deleting data repeats the switch back to "off"/"0".

But there is slighting more "charge" than last time.

This process keeps repeating until the drive controller can no longer supply enough voltage to flip the "gate" to the other state (off/on 0/1).

Similar to how a car battery charger must output greater VOLTAGE (very basically: electrical "pressure") to charge a car battery (car batteries are usually 12.99 volts when "dead" and 14.5 volts when "charged").
Which is why most car alternators are around 15 volts.

Also a problem with SSDs: If the current "decays" then the drive "forgets" the data.
-----------------------

I've been following the Linux kernel patches since this started.
Most of the performance reduction seems to be related to primarily one task...
I/O, some links to benchmarks:

Initial (I think) batch of benchmarks:
https://www.phoronix.com/scan.php?page=article&item=linux-415-x86pti&num=1

More test, also he explains a bit more:
https://www.phoronix.com/scan.php?page=article&item=linux-more-x86pti&num=1

This gentleman is always running test and benchmarking...
I find it useful to see what's coming up soon.
He test mostly on Ubuntu (unless it a comparison between distros), usually with the latest "git" code (i.e the latest code from the development "branch").

Barracuda

Skybird

01-25-18, 07:06 AM

Everybody, stay away from any BIOS or UEFI firmware updates.

Skybird

01-25-18, 07:23 AM

Since years...? I admit this was new news to me today. Although i repeatedly said that MS's design of W10 may be due to orders by the NSA.

https://threatpost.com/intel-confirms-its-much-loathed-me-feature-has-a-kill-switch/127739/

Positive Technologies researchers Mark Ermolov and Maxim Goryachy said they believed the kill switch was introduced by Intel at the behest of the National Security Agency which it said also viewed the Intel Management Engine as a possible weak spot of a system’s endpoint defenses.
Concerns over the Intel Management Engine (ME) have been ongoing for years. In May, Intel patched a critical vulnerability (https://threatpost.com/intel-patches-nine-year-old-critical-cpu-vulnerability/125331/) that dated back nine years in the company’s Active Management Technology, which is based on Intel ME. That vulnerability could allow an attacker to gain remote access to AMT services such as the keyboard, video and mouse (KVM), IDE Redirection, Serial over LAN, and BIOS setup and editing.
Suspicions date back to 2012 over Intel’s implementation of Active Management Technology (AMT) with some labeling it a “backdoor enabled by default (https://forums.lenovo.com/t5/Security-Malware/Intel-AMT-backdoor-enabled-by-default/td-p/824749).” A reported flaw identified in June 2016 by researcher (http://boingboing.net/2016/06/15/intel-x86-processors-ship-with.html) Damien Zammit claimed that there was a remotely exploitable security hole in the Intel Management Engine that created a secret backdoor allowing a third party to use undetectable rootkits against Intel PCs. Intel denied such claims.

Lets go digital, everybody must be digital, be digital to maximise your vulnerability, all your mone yonly digital, give access to your life and belongingsa to all world!

Its safe. Always assume the best intentions. And whats more important: its so mega-cool to be all digital and pay with your credit card or your cellphone!