@AviD Well, the point is you can find good stable voltages/settings, and reduce power draw, then whenyou see how high you go, you can re-enabled HT, it shouldn't cause any instability but power+heat will be up, so you'll need to cool it adequately
Well. This does not bode well. It was starting to feel somewhat warm in the house, so I went to turn on the AC. It appears as though the compressor isn't kicking on.
@AviD with turbo you're enabling it to clock individual cores higher when others are not in use, meaning you'd be getting a lot of heat on a really small space... that's tough to cool down, so it's suggested that you disable turbo. Also, it's a bit sporadic since it's only tested for nominal speeds and I believe that the multipliers it uses can make it really unstable. I would advice against it.
@AviD Another thing to consider is also that the mobo is faulty. I wouldn't bet against it, since you had strange problems since the start (better stability at higher OC)
But I highly doubt that you permanently damaged your CPU, not with the OC you were doing (unless the heatsink isn't properly attached to it... then all is possible)
Anyway, I was actually dreaming of that mocking that I was on the receiving end of after giving my opinion on compiled code vs. interpreted languages security. Thought of another case for the former, if anyone is interested in pulling down my pants again and giving me some from behind... just tell me when to drop the soap :)) :P
@D3C4FF added my +1 on that Synergy question because it's good and valid... but I can't go through others any faster not to trip over the serial voting check
At 1335 on both Sec.SE and SO now... made one edit here, waiting for the accept... need another on SO
@D3C4FF If for example I went on your profile page and opened your questions/answers from there to vote on them, the system would detect serial upvoting and reverse them. It's also doing other checks to prevent it, but the ones from profile page as a referral are the most obvious ones. I.e. if you vote on mostly a single individual's posts, or vote too fast without taking the time to actually read it... stuff like that
@TerryChia oh cheers! didn't know that... so each time I edit I bug either two high ranking users or a mod? wheeey there's more perks to it than I initially thought hehe
I want my two 1337s now... but then 3000 here I come :)
@D3C4FF yo is this DNS spam.looptech.com.au familiar to you? had it in my access logs with referrer from my profile here... I was thinking, who do I know in Oz that would put spam in their DNS name... thought of you 1st LOL
@D3C4FF BTW... Remove the overflow:hidden; from Styles_System1.css, line 259 ;) Some of the contents on the main page are hidden under another DIV... either that, or change the style for the table after it to width: 581px;... someone messed it up there :P
If you're looking for the equivalent of OWASP for network security then I'd suggest that testing methodologies like OSSTMM or PTES (although PTES is still rather incomplete), are a good place to look for information.
However as you say both of those will have a very large amount of information i...
@RoryMcCune PTES would be less incomplete if they didn't crop their diagrams like this one LOL
@RoryMcCune Shame indeed. I still find it useful tho, especially for the diagrams and nice explanations to them. Organizing things around is with PT more of a challenge than the tests themselves sometimes are IMO
when I say organizing... I mean also reporting on findings obviously. That's where PTES comes quite handy with nearly copy/paste-able structure that then just needs actual test results... I wasn't involved with many pentests tho, so it's just my impression really
not many pentesters here tho... I expect most were consumed by UK and Germany :)
@RoryMcCune Out of curiosity, how many different versions do you include in your reports? I've only glimpsed through reports for management and mostly worked on the ones for devs, but the differences were huge
@TildalWave well we typically pass out one report with a management section at the beginning (short and high-level) and then a detail section in the middle (goes into what was tested and found) and then a "slides" section at the back which just summarizes each of the vulns found...
so it can be a bit repetitious but then things like Dradis can help automate some of it..
@RoryMcCune You do follow-ups as standard? Or that depends on the client? I imagine it's suggested and probably a standard practice with bigger customers, but what about the smaller dev teams? If you even do those?
@TildalWave TBH we very rarely get asked to do follow-ups which is a shame. I did one the other week where I did a workshop on web app security and talked through the findings for the devs
but that was the first one I've done in over a year.
I prepared an easier to read version of the diagram (I believe). It's following same basic principles the @D3C4FF's excelent answer does. I was tempted to go with his analogy first, but I thought it wouldn't be appropriate due to recent events and for the current global political climate.
The t...
I've added a bit more description why I think it's a suitable analogy. Obviously some didn't get it... I honestly hope it's not simply a case of them using too reflective displays LOL
@RoryMcCune That's a shame yup. What's the general feedback you get for your VAs? Are your clients overwhelmed by your reports, or they actually knew what to expect, or even gave suggestions?
@RoryMcCune There's probably some consulting going on even before they sign the contract,... I guess what I'm asking is if you negotiate to what detail you should go, or you have packages and then decide on pricing and timeframe terms based on project complexity? Or a bit of both?
@TildalWave VAs are weird. In a lot of cases I think people are having them done purely for compliance purposes, so we'll see the exact same findings repeatedly
@TildalWave well I don't see a lot of the pre-sales stuff all the time, but a lot of it is trying to find out why they want the test as much as what they're looking to cover. Ideally people want the test to improve their security, so that's the best case as you can give guidance on fixing stuff. Other times it's contractual/compliance and then it can become a point of getting as much done as possible within their budget which tends to be fixed.
@RoryMcCune I glimpsed through your webpage and have a few suggestions. You mentioned now compliance testing, so maybe it wouldn't be a bad idea to include what certification compliance testing you offer (maybe references too). And maybe in the stuff section also what certification they have? I only went there curious about what you said BTW, I'm not a certification freak or spying on you :P You've already shared the contents of your car's booth with the rest of the world anyway LOL
@TildalWave heh, no worries, all that stuff is as public as the contents of my car boot :) Most of the stuff we do at the moment is 3rd party (we work for other testing companies) so references are a little tricky, although we've had a couple recently. good point though I should put compliancy content up there to help with search traffic.
@RoryMcCune some good points there at the yes but can I crash it? about what PT actually is, possibly the simplest explanation I've ever heard (minus the accent LOL)
@RoryMcCune neah I don't have problems understanding... just teasing ;) BTW... that brownish liquid you're drinking... could that be the source of your accent? :P
Looks like a posh single malt
3/4 pint daily allowance for married with children LOL
Club-Mate () is a caffeinated carbonated mate-extract beverage made by the Loscher Brewery (Brauerei Loscher) near Münchsteinach, Germany, which originated in 1924. Club-Mate has a relatively high caffeine content (20 mg per 100 ml), low sugar content (50 g/kg), and low calories (20 per 100 ml of beverage), compared to other beverages such as cola and other energy drinks.
Club-Mate is available in 0.33 and 0.5 litre bottles.
History
Geola Beverages of Dietenhofen, Germany originally formulated and marketed Club-Mate under the name Sekt-Bronte. The drink was only known regionally unti...
@TildalWave Well, it's a universally proven facts that Rorys are into Metal music, and we all know that head bangers worship Satan and drink babies' blood.
You are the victim of branch prediction fail.
What is Branch Prediction?
Consider a railroad junction:
Image by Mecanismo, from Wikimedia Commons: http://commons.wikimedia.org/wiki/File:Entroncamento_do_Transpraia.JPG
Now for the sake of argument, suppose this is back in the 1800s - before...
@TerryChia possibly. It could also be that there is much more content, so the chances of a Q reptraining are low. But there was some research done into voting in tags and I think haskell came out as the tag to be in for votes/answer.
@Adnan Dunno for @RoryMcCune (tho possibly because he associates with @RoryAlsop), but @RoryAlsop looks young because he's so old that his appearances run into the integer overflow (the - sign only shows in the mirror reflection, though).
Haggis - it's like a sausage, only better. You can use it in place of mince to make haggis lasagne, or as a topping for haggis pizza. You can have it for breakfast, lunch, (and/or) dinner
@RoryAlsop That pizza looks Scottish for more reasons than just being a haggis version of pizza Bolognese LOL the crust is wiiiiideeeee and no cheese on it! You probably don't get many Italian tourists, huh? :)))
@TildalWave definitely - I am a massive pizza fan. Generally where other bands would drink/take drugs etc between soundcheck and playing - we go find a good Italian restaurant :-)
@TerryChia dunno - I only ever liked a couple of things about her - edible clothing, and her stage craft. I'm not really that into pop music (cos I'm a fogey)
@TerryChia If John could get his hands on some he just might. He doesn't like to do odd things for the purpose of being odd, but he does try to do things unusual that taste yummy.
Oh, the systems guys in here should probably look at this.
@TerryChia Yeah, he pretty much summed up my feelings in his first post.
@TerryChia It's more about retraining your debugging process.
$THING doesn't seem to be working, is $THING running? Yes. Does $THINGS logs list any obvious failures? No. Are there any denies in /var/log/audit/audit.log?
@RoryAlsop The policies, these days, are pretty spot on. But yeah, between the policies being good and restorecon to fix issues...there's 90% of the problem solved right there.
@RoryAlsop @ScottPack How much data does Ingress use? I'm probably gonna play it with my phone tethered to my tablet but I don't want to burst my data cap.
@TerryChia The max I managed in one month was 1.5Gb, and that was using the IITC browser a lot. Get one of the battery packs I mentioned - 12000mAh is a lot!
@ScottPack Yep! Just moves right along and spawns the process though. At least output something lol. I'm trying to refine one particular alert. It comes up when Adobe tries to run an update and I've found the rule but I'm not sure if I can do an exception
A lot of discussion I see suggests that changing default ports for services is just "security by obscurity" and is easily defeated by scanning for open ports.
My question is this, though, if the port is changed to a random port, yes an attacker knows it's open, but do they know what service is ...
Technically, there is some prioritization, but not much. Largely it was one of those decisions that was designed in for future proofing, but wasn't ever really fully utilized.
@Travis The best think you could do is rewrite the signature that's firing and tweak it so it doesn't fire on your use case.
Is there any advantage in changing the SSH port, I've seen people do that, but I can't seem to find the reason why.
If you have a strong password and/or a certificate, is it useful for anything?
Edit:
I should also mention that I am using iptables rules to limit brute forcing attacks, only 5 l...
A few people around here brought up a decent point that it cuts down on the automated bruteforce type attacks you find on the internet and can help to somewhat reduce the noise on your logs.
So, weigh that against having your users remember an obscure port number.
@ScottPack It's a very vague rule. Just looks for content of PE|00 00| saying that a portable EXE file was detected. Not smart enough yet to come up with a solution on how to block everything but adobe
Will the signatures work chronologically like if I do a pass first for the adobe and alert for the current sig?
@ScottPack: Good info in that SSH port article. Very relevant for sure. I can't upvote until I get 15 rep on there though :(
@Travis It might be worthwhile to just disable, then. Or, perhaps, not not fire an alert on akamai boxes.
@Travis Not really.
So, the reason snortd takes so long to start is that it's processing all of the signatures and building out a big state table/decision tree like data structure.
So the order in which the rules are processed depends more on how they're constructed.
@Travis PulledPork will pull down the newer SharedObject rules, oinkmaster can't process those. You can also use regexes in your PP config to modify a signatures when you update. So you actually can locally modify a shipped rule and still update.
Stack Overflow has Code Review. UX is about to have Design Review.
So what do we need? Security Review. Someone needs to start a proposal to make such a site happen.
Proposal: Security Review
As far as I can tell, this proposal essentially amounts to asking StackExchange citizens to perform a penetration test on live sites.
I can see several serious issues with this:
It may be difficult to verify that the OP is:
The actual owner or representative of the ...
Also, if Snorby is doing any kind of internal tracking, it may have already created an internal entry for that signature ID and mapped it back to the invalid name.
We're run into the latter problem with our SIEM. It creates a new log message record in it's own signature database every time it sees a new kind of log message, but it's keyed off the SID (for snort) and the message field doesn't get updated.
So what I ended up doing was configuring my snort puppet module to restart barnyard if the rules get updated.
@AJHenderson It's amusing really to watch it again after growing up and realizing that whitehats don't fly through the cyberspaces of their own computer systems with high speed VR glasses.
While looking for solutions to entropy pool depletion on virtual machines, I came across an interesting project called haveged, which is based on the HAVEGE algorithm (HArdware Volatile Entropy Gathering and Expansion). It makes a pretty fantastic claim.
HAVEGE is a random number generator th...
@ThomasPornin Maybe it was one of those random things? :) "It is not due to timer drift, but the aggregation of the asynchronous behaviors in a modern processor" doesn't read as true random to me. If the RNG is determined by fluctuation in code execution time, then what happens to it when you run two or more of these in parallel? Is f(1->N)[0,1] mod N = RND? Maybe I'm missing something, but I'm not so sure it maintains the randomness with high N.
@TildalWave Well, there is randomness, and there is unpredictability.
Apparently, the RNG feeds itself on the exact timing of some operations which may take a variable time depending on "a lot of factors"
mostly cache misses, TLB misses, and branch prediction
If several such entropy harvesters run in parallel, it is likely that the values they obtain will not be independent of each other.
In particular, if the attacker can run some code of his own on the machine, and the processor uses hyperthreading, then the attacker can more or less obtain the same kind of value as the entropy harvester
i.e. the harvester will indeed obtain values which are more or less random (in a physical sense) but the attacker will get the same, which means that the attacker can still "predict" them (a posteriori prediction, which still counts as an attack if the attacker does not get to see the PRNG output)
it seems that the HAVEGE designers are not completely aware of the difference between randomness and unpredictability, the latter being crucial to security (e.g. if you use the PRNG to generate a RSA key)
@ThomasPornin That was my initial thoughts when I was reading that reply. I didn't delve deeper into it, because... well it's Friday evening LOL... but what you describe seems exactly to be the problem with it. But not just that it isn't random (i.e. it is unpredictable, but not random), I fear that the resolution of the function isn't infinite either, meaning you can overload it to return predictable results. How random it is, when it depends on execution time, but you can control this? It isn't.
@ThomasPornin Still, I'm not sure what you meant with answerer did it again? You gather it should be converted to a comment again? Or were you just surprised that his earlier answer was converted? (I didn't see it before this alleged conversion to a comment, but seems rather strange to me - never seen it before on a post that at least attempts on answering the question)
@TildalWave The guy first post an answer, which consisted of comments on the other two answers. Some mod converted the answer into comments (mods can do that). Apparently, he was not happy with the conversion, and reposted an answer. I wanted to make the mod (supposedly @Rory the Old) aware of it, so as he should decide whether this requires mod action again.
On that specific subject, I don't have a definite answer. Answers-which-should-be-comments are bad answers, but I am quite happy with leaving bad answers just be bad answers, without any mod action to erase them.
@ThomasPornin Oh then it was justified. This latter post seems OK to me, albeit a bit strange to explain the difference between HAVAGE and havaged, say that he's working on havaged, then only deal with how HAVAGE works internally. But it is in a form of an answer, maybe just the first comment in it isn't really needed. (I'll just edit it to remove the first paragraph and a few typos, then we'll see what mods will think of it)
@ThomasPornin yes, if I get things correctly, two are needed to approve edits then... I'll re-read it as well to check for errors/typos ... shouldn't take a minute ;)
@ThomasP - it looks like his previous answer was split up into a range of comments. I'm not sure how to do that - quite a useful tool. Re the answer, I think @Tildal's edit sounds the way to go. Probably not mod needed