VOGONS


Reply 80 of 151, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Here's an alternative to the Powershell script for Windows users:

Easily examine and understand any Windows
system's hardware and software capability to
prevent Meltdown and Spectre attacks

https://www.grc.com/inspectre.htm

DOSBox Compilation Guides
DosBox Feature Request Thread
PC Game Compatibility List
How To Ask Questions The Smart Way
Running DRM games offline

Reply 81 of 151, by KCompRoom2000

User metadata
Rank Oldbie
Rank
Oldbie
DosFreak wrote:

Here's an alternative to the Powershell script for Windows users:

Easily examine and understand any Windows
system's hardware and software capability to
prevent Meltdown and Spectre attacks

https://www.grc.com/inspectre.htm

Just downloaded that to my Windows 7 PC (HP DC7800 with a C2Q Q6600), and I get the following results:

Spectre & Meltdown Vulnerability Status

Vulnerable to Meltdown: NO
Vulnerable to Spectre: YES!
Performance: SLOWER
(full details below)

Fuck, It looks like I'm vulnerable to Specture. 🙁

I still post here, but only occasionally.

Reply 82 of 151, by Firtasik

User metadata
Rank Member
Rank
Member

Phenom II and Windows 8.1:

Spectre & Meltdown Vulnerability and Performance Status

Vulnerable to Meltdown: NO
Vulnerable to Spectre: YES!
Performance: GOOD
(full details below)

And no chance for the BIOS update. 😵

Unfortunately, this system will be open to exploitation of the Spectre vulnerability until and unless its BIOS and CPU microcode firmware are updated. 

11 1 111 11 1 1 1 1 1 11 1 1 111 1 111 1 1 1 1 111

Reply 83 of 151, by bjwil1991

User metadata
Rank l33t
Rank
l33t

On the bright side, your system's performance is better than my Pentium Dual-Core laptop.

Discord: https://discord.gg/U5dJw7x
Systems from the Compaq Portable 1 to FX-6300
Twitch: https://twitch.tv/retropcuser

Reply 84 of 151, by TryAgain

User metadata
Rank Newbie
Rank
Newbie

There are few misconceptions here.

First Meltdown is really Intel hardware flaw.
In essence, to allow speculative execution to take that branch of the code which step up on kernel memory when processor is into user mode and running unprivileged process is unacceptable error of implementation of their own architecture.
Simply that path must precluded exactly that speculative direction of execution at MMU level.
Thumb for AMD here.

Next. Pentium and Pentium MMX are more close to 486 than to Pentium Pro... These are in order execution style cpus and their branch prediction is of a rudimentary type. Think of not lately Atom cpus.
This also apply to Cyrix Mx. It's safe because out of order is of very primitive type on this cpu.
So I make a bet that these aren't vulnerable to both flaws.
The simple reason of long life of these vulnerabilities lie in hard to extract
parallelism from general x86 code.
So manufacturers sacrifice security for performance. But now pendulum returns with a terrible sign of comming frozen age ... after Meltdown.

It reminds me of question why windows 98 have had been so fast compared to Linux, because there is a little security to check in it...

Reply 85 of 151, by Scali

User metadata
Rank l33t
Rank
l33t
TryAgain wrote:
First Meltdown is really Intel hardware flaw. In essence, to allow speculative execution to take that branch of the code which s […]
Show full quote

First Meltdown is really Intel hardware flaw.
In essence, to allow speculative execution to take that branch of the code which step up on kernel memory when processor is into user mode and running unprivileged process is unacceptable error of implementation of their own architecture.
Simply that path must precluded exactly that speculative direction of execution at MMU level.
Thumb for AMD here.

It's not as simple as that. You have to define what 'taking a branch' means.
Technically the branch is not taken until the instruction is retired.
This never happens on Intel CPUs.
So to the user process, the branch is never actually taken, and you are never doing anything unprivileged.

The problem however is that the speculative execution makes use of two things:
1) The branch prediction unit
2) The cache

By executing carefully constructed code, you can time whether a branch is predicted properly, or whether a value was already fetched into cache before.
This is what is known as a 'side-channel': the actual information is not leaked, but it can be deferred from the behaviour of the CPU.
Up to now, no CPU designer ever realized that such a side-channel could open because of speculative execution. Which is why both Intel and ARM are affected.
What we have here is a paradigm shift. From now on, CPU designers will think differently about security forever.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 86 of 151, by TryAgain

User metadata
Rank Newbie
Rank
Newbie

Yes. I shouldn't use word branch. It is reserved for Spectre flaw... But it really is irrelevant because those flaws are very similar. Manufacturers assume that nobody check that in fact modern cpu do not check memory operation BEFORE doing speculative execution.

World down example. There is a bank robbery. There are some unspecified criminals in the room with the safe. Suddenly in this room enters an innocent clerk which rush to the safe, unlocking it and say: "Well, It's all right...everything is in place"...
until when I add.

Reply 87 of 151, by Scali

User metadata
Rank l33t
Rank
l33t
TryAgain wrote:

Yes. I shouldn't use word branch. It is reserved for Spectre flaw... But it really is irrelevant because those flaws are very similar. Manufacturers assume that nobody check that in fact modern cpu do not check memory operation BEFORE doing speculative execution.

The whole point of speculative execution is that you *speculate*. One such speculation is assuming that you have access (hence you go ahead and prefetch the data in cache, so you have it cached by the time access is cleared, to speed up processing).
Just like you speculate that you take a branch, only to find out later if you should have or not. Then you roll back.
How hard is it to get the concept?

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 89 of 151, by DosFreak

User metadata
Rank l33t++
Rank
l33t++
Joey_sw wrote:

But is this whole spectre & meltdown vulnerability really harmfull for averages PC/tablets/smartphones users tho?

Why wouldn't it? Read up on the vulnerabilities there's plenty of info out there. Assuming the OS and browser patches are applied then probably not likely. Assuming you're running an updated antivirus and use application control then really unlikely but if you're talking about the "average" user then OS, browser patch and antivirus probably aren't likely so in that case they are already compromised via remote exploit or they installed an exploit themselves due to their own ignorance and no need for meltdown or spectre but mabye it's just a nice to have as one more vulnerability to add to an exploit kit.

DOSBox Compilation Guides
DosBox Feature Request Thread
PC Game Compatibility List
How To Ask Questions The Smart Way
Running DRM games offline

Reply 90 of 151, by bjwil1991

User metadata
Rank l33t
Rank
l33t

Ran the InSpectre program on my Windows 10 Pro running on an AMD FX-6300
Spectre & Meltdown Vulnerability
and Performance Status

Vulnerable to Meltdown: NO
Vulnerable to Spectre: YES!
Performance: GOOD

Well, crap. Also, my motherboard doesn't have a new BIOS whatsoever.

My ThinkPad R40 running XP Pro 32-bit only reports both meltdown and spectre exist, but the performance is good.

Discord: https://discord.gg/U5dJw7x
Systems from the Compaq Portable 1 to FX-6300
Twitch: https://twitch.tv/retropcuser

Reply 91 of 151, by TryAgain

User metadata
Rank Newbie
Rank
Newbie
Joey_sw wrote:

But is this whole spectre & meltdown vulnerability really harmfull for averages PC/tablets/smartphones users tho?

Let do speculative thinking...
If average Joey tablet uses one of ARM Cortex A7 or A53 cpu or clone, tablet is immune , because there is not speculative execution in such cpu...

But there is speculation in my thinking which is vulnerable to both, so I roll back my answer with your mind backspace, but yeah you still remember something about lucky tablet numbers 7 and 53...

I bet that Pentuim 3 Coppermine and Duron Spitfire are also immune to these flaws. They have not hardware prefetcher that screw MMU on these cpus...

Reply 92 of 151, by TryAgain

User metadata
Rank Newbie
Rank
Newbie
bjwil1991 wrote:

Ran the InSpectre program on my Windows 10 Pro running on an AMD FX-6300
Spectre & Meltdown Vulnerability
and Performance Status

.

The problem with Spectre is that it is cpu dependent i.e. if your cpu has microcode changes it is not anymore the cpu you bought...

Software that make this check must be mandatory of code morphing piece of type which is abandoned years ago for the reason that it is not fit well in pipeline of performance kind cpu and so on and so forth...

It must be of special breed that convert the near zero probability to somewhat lottery winning magnitude due to differences in such called homebrew architecture...

As scientific paper ends: "...Spectre will hunt us in many years to come...."

BTW one of the author is 22 years old, just three years younger then flaw itself... Poor Intel engeneers, but not so poor CEO...

That is why Russia and China start to reinvent the wheel, they want to know that wheel is the wheel and ain't focus-pocus under the hood...

Reply 93 of 151, by psychz

User metadata
Rank Member
Rank
Member

Come on, we're on vogons! Regarding retrocomputing, worrying about spectre and meltdown or the associated performance loss caused by the patches, is really pointless imho. I mean, the OSes we run on such (vulnerable) CPUs are flawed by design on oh-so-many well-known, popular and widely exploited security issues. But who cares about, for example, OS/2 Warp 3 or Windows 98 in 2018, running on computers used mainly for gaming or offline hobbyist tasks? On the other hand, who in their right mind, would run Win10 on a P2, even if it wasn't dependent on newer CPU features?

Stojke wrote:

Its not like components found in trash after 20 years in rain dont still work flawlessly.

:: chemical reaction :: athens in love || reality is absent || spectrality || meteoron || the lie you believe

Reply 94 of 151, by dr.zeissler

User metadata
Rank l33t
Rank
l33t

For me a patch for shutting-down the speculative-execution could be a very good thing for slowing the cpu down for retro-gaming 😀

Retro-Gamer 😀PowerMac 6100-66/Houdini 486/66 - G4 Cube 450/Rage128pro OS9.0.1 - Macintosh LC/Apple IIe Card OS6.0.8 - Acorn A4000 Archimedes - Unisys CWD 486/66 + Aztech Washington

Reply 95 of 151, by 386SX

User metadata
Rank l33t
Rank
l33t

What about the K6-2?
Anyway it looks like the same old story. After so many years someone find a bug, that in next version will be patched until in the very next years some other guy will find another one that has been out there forever. Software bug or hardware it doesn't change the usual news story.
it would be interesting to think if it's better to update or downgrade for security reasons.
Anyway i don't understand how these vulnerability would run on a machine, trough a javascript on the browser? If people usually dont run untrusted software how could that realistically happen?
Also on a 16bit kernel based win95/98/Me machine, would we expect someone would really think to code something for such an old os?

Reply 96 of 151, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Yes javascript is one vector.

Define "people usually don't run untrusted software". Unless it's a repo or app store...and even then.

DOSBox Compilation Guides
DosBox Feature Request Thread
PC Game Compatibility List
How To Ask Questions The Smart Way
Running DRM games offline

Reply 99 of 151, by dr_st

User metadata
Rank l33t
Rank
l33t
386SX wrote:

Also on a 16bit kernel based win95/98/Me machine, would we expect someone would really think to code something for such an old os?

No, they wouldn't, and this is why I said earlier that this thread is stupid.

The only unusual thing about these exploits is that they have the ability to leak kernel memory without explicitly attaining privilege escalation. Hence, it's really a big deal only for systems that have a built-in privilege separation mechanism, that people actually use. It's completely irrelevant for Windows 9x because there is no separation. It's also irrelevant for people who run as Administrators on 2K/XP, or for people who run Vista and later with UAC disabled, because any malicious piece of software can just automatically elevate itself to kernel mode, and then have access to everything anyways.

It is also irrelevant for people who run Linux as root (although this is a less accepted practice in Linux compared to Windows).

If you are one of the people who runs as root or administrator without UAC, ask yourself - did you feel vulnerable before Meltdown/Spectre were discovered? If the answer is no, then you can just forget about the whole thing - the new vulnerabilities don't make you more or less exploitable.

https://cloakedthargoid.wordpress.com/ - Random content on hardware, software, games and toys