IRC logs of #boinc for Friday, 2009-08-14

02:29 *** efc has quit IRC

02:34 *** linear_shift has joined #boinc

04:38 *** wdsmia_ has joined #Boinc

04:38 *** wdsmia has quit IRC

05:05 *** wdsmia has joined #Boinc

05:06 *** wdsmia_ has quit IRC

05:56 *** synapt has quit IRC

05:56 *** synapt has joined #boinc

06:01 <MTughan> PovAddict: I'll be sure to run it in a VM when I do it then. Looks interesting though.

06:04 *** arfmarf has joined #boinc

06:20 *** Celelibi_ has joined #boinc

06:24 *** wdsmia has quit IRC

06:24 *** wdsmia has joined #Boinc

06:28 *** Celelibi has quit IRC

06:28 <hawmps> 'lo

06:29 <MTughan> Morning Jay.

06:30 <hawmps> Hi Mike

06:33 *** Aeternus has joined #boinc

06:36 <hawmps> is seti just not downloading apps?

06:36 <hawmps> or supporting files?

08:46 *** yomshleeshee has joined #boinc

08:54 <hawmps> is seti down?

08:55 <MTughan> Nope.

08:55 <MTughan>

08:55 <Romulus> Title: Server status page (at

08:58 <yomshleeshee> Einstein still is :-/

08:58 <MTughan> Yeah, filesystem crash.

08:58 <MTughan> I don't see anything new on their homepage yet.

08:58 <yomshleeshee> yeah, I went to reattach after a bit of a boinc break and it was down. d'oh!

08:59 <MTughan> One second, bot tricking...

08:59 <MTughan> Einstein is at

08:59 <Romulus> Title: Einstein@Home status (at

08:59 <MTughan> Einstein?

08:59 <Romulus> I guess Einstein is my second project after SETI, MTughan

08:59 <MTughan> Gah.

08:59 <MTughan> forget Einstein

08:59 <Romulus> Got it!

08:59 <MTughan> Einstein is at

08:59 <Romulus> Title: Einstein@Home status (at

08:59 <MTughan> Einstein?

08:59 <Romulus> Somebody said Einstein is at, MTughan

08:59 <MTughan> There we go.

09:00 <yomshleeshee> lol. bots :-/

09:03 <hawmps> almost got the seti/linuc/cuda app working

09:03 <hawmps> right up to the 'download failed' part

09:03 <MTughan> It looks like the download servers are still up. Updated 3 minutes ago.

09:16 <hawmps> any way to get a more detailed error report?

09:17 <MTughan> There should be... One sec.

09:17 <MTughan>

09:17 <Romulus> <> (at

09:19 <hawmps> thanks

09:20 <hawmps> bb shortly

09:20 *** hawmps has quit IRC

09:23 *** hawmps has joined #boinc

09:23 *** ChanServ sets mode: +o hawmps

09:25 <hawmps> &wx 12020

09:25 <Romulus> hawmps: Temperature: 66.6F / 19.2C | Humidity: 100% | Pressure: 29.98in / 1015.1hPa (Steady) | Conditions: Overcast | Wind Direction: NNE | Wind Speed: 0.0mph / 0.0km/h ; Today - Patchy dense fog this morning. Mostly sunny. Highs in the upper 80s. Light and variable winds...becoming southwest around 5 mph this afternoon.; Tonight - Clear. Lows in the lower 60s. South winds around 5 mph in the (1 more message)

09:25 <hawmps> &more

09:25 <Romulus> hawmps: evening...becoming light and variable.; Saturday - Mostly sunny. Humid with highs in the upper 80s. Southwest winds around 5 mph.;

09:26 *** wdsmia_ has joined #Boinc

09:26 *** wdsmia has quit IRC

09:32 *** hawmps has quit IRC

09:33 *** hawmps has joined #boinc

09:33 *** ChanServ sets mode: +o hawmps

09:34 *** zombie67 has joined #boinc

09:51 *** wdsmia has joined #Boinc

09:51 *** wdsmia_ has quit IRC

09:54 *** wdsmia_ has joined #Boinc

09:54 *** wdsmia has quit IRC

10:05 *** wdsmia has joined #Boinc

10:06 <hawmps> conectile dysfunction?

10:09 *** wdsmia has quit IRC

10:09 *** wdsmia has joined #Boinc

10:14 *** wdsmia has quit IRC

10:14 *** wdsmia has joined #Boinc

10:15 *** wdsmia_ has quit IRC

10:25 *** zombie67 has quit IRC

10:40 <desti>

10:40 <Romulus> <> (at www.!@#&

10:50 *** wdsmia_ has joined #Boinc

10:50 *** wdsmia has quit IRC

11:08 *** wdsmia has joined #Boinc

11:08 *** wdsmia_ has quit IRC

11:19 *** wdsmia_ has joined #Boinc

11:19 *** wdsmia has quit IRC

11:21 *** wdsmia has joined #Boinc

11:21 *** wdsmia_ has quit IRC

11:25 *** wdsmia_ has joined #Boinc

11:25 *** wdsmia has quit IRC

11:29 <quail> hii! alll

11:40 <yomshleeshee> hey

11:40 <yomshleeshee> anyone here crunch cosmo@home?

11:41 <quail> not me, I have not crunched in along time

11:42 <yomshleeshee> just wondering about their credit assignments. The granted seems high compared to claimed. Unless it's just me:

11:42 <Romulus> <> (at

11:42 <yomshleeshee> claimed 90 granted 420 !?!

11:43 <quail> have you checked what the others are getting for doin the same wu?

11:43 <yomshleeshee> well, these are min quorum = 1. so I'd have to find similar ones

11:44 <yomshleeshee> Ok

11:44 *** wdsmia has joined #Boinc

11:44 <yomshleeshee> my wuindex + 4 is a result that is back and it claimed 330 and granted the same 420

11:44 <quail> credit normally works on and average between you and couple others doin the same wu

11:44 <yomshleeshee> right, this is clearly a fixed credit

11:47 <quail> ah

11:49 <yomshleeshee> I mean I see two others that are my index+4 to 7 and these are  claiming ~300credits. They're dual cores

11:49 <quail> damn I love my new laptop, even with its little probs under linux

11:49 <yomshleeshee> so I guess they just figure give those kinds of numbers to everyone

11:49 <yomshleeshee> works for me, just seemed high

11:50 <yomshleeshee> yeah linux on laptops has come a *long* way. I remember the mandrake days where I could only do B&W graphics o.O

11:50 <quail> I use debian

11:50 <yomshleeshee> yeah, it's popular or the ubuntu equiv

11:51 <yomshleeshee> I'm a fedora guy, though I've been running ubuntu on my laptop. thought I'd give that try to see what the hype is about

11:51 *** wdsmia_ has quit IRC

11:51 *** wdsmia_ has joined #Boinc

11:52 <quail> ubuntu didn't like my laptop

11:53 <yomshleeshee> really? huh. I'm surprised

11:53 <quail> I could never get ubuntu 9.04 to shut my laptop down properly, and had to hard power it off

11:53 <yomshleeshee> Sometimes it jacks up my audio. I'll get distortion every so often. But other wise I've had no problems

11:54 <quail> this my laptop

11:54 <quail> SysInfo: Linux 2.6.30-bpo.1-amd64 |  Intel(R) Core(TM)2 Solo CPU    U3500  @ 1.40GHz 800.000 MHz | Bogomips: 2793.17 | Mem: 3412/3901M [||||||||||] | Diskspace: 282.42G Free: 261.86G | Procs: 115 | Uptime: 8 hrs 11 mins 42 secs | Load: 0.36 0.15 0.04  | Screen: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07) @ 1366x768 (32 bpp) | wlan0: In: 68.86M Out: 6.21M

11:54 <yomshleeshee> wow. yeah, that's a problem. I'm running 9.04 too. I nervously tried the suspend the other day and it worked just fine. Was stoked about that.

11:55 <yomshleeshee> ah, yeah that's beefier than mine. I'm running a *clears throat* cheap as dell computer

11:55 <quail> the laptop itself is a acer aspire timeline 4810T

11:55 <yomshleeshee> was skeptical of the hardwar compatability but have been impressed

11:55 <quail> dells are good

11:56 <yomshleeshee> i mean, I've had good experiences with them

11:56 <quail> :-)

11:56 <yomshleeshee> indeed

11:56 <yomshleeshee> and they started making a line of comps designed for running linux

11:56 <quail> reason I got this laptop is it has 8+ hours battery life

11:57 <yomshleeshee> :-o

11:57 <yomshleeshee> that'd be sweet. mines ~1hr :-/

11:57 <quail> which is good for me as I am always on the go and not always at a power outlet

11:57 <yomshleeshee> right. Battery's an important consideration

11:58 <quail> and it 1.9Kg which is great for carrying around

11:58 <yomshleeshee> nice

11:59 <quail> I want to get one them mil spec panisonic laptops

12:00 <yomshleeshee> hmmm

12:00 <hawmps> 'toughbook'

12:00 <yomshleeshee> ah

12:01 <quail> only prob with them is they around the 5Kg heavy

12:01 <yomshleeshee> yeah, that's pretty crazy

12:02 <yomshleeshee> that'd be like carrying a gallon of milk around everywhere. not fun

12:03 <quail> and the price is another factor as they 3000AUD+

12:05 *** zombie67 has joined #boinc

12:05 *** wdsmia has quit IRC

12:06 <zombie67> got my ram yesterday.  Up to 20gb.    Now all 8 TLP tasks are running smoothly.

12:06 <yomshleeshee> yeah...

12:06 <PovAddict> holy crappers

12:06 <yomshleeshee> 20gb! crapola

12:06 <quail> zombie67: fscking nice

12:06 <zombie67> It was 8x1gb, now it's 4x4gb + 4x1gb.

12:07 <yomshleeshee> yeah, that's pretty rediculous. I have HD's smaller than that o_O

12:07 <MTughan> 4GB sticks? Those are expensive.

12:07 <quail> MTughan: yeah they are

12:08 <quail> and I only got 4GB ram in my laptop :-(

12:08 <quail> hehe

12:08 <yomshleeshee> I only have 1Gb :-( x 2

12:08 <MTughan> I have 4GB in my laptop and 8GB in my desktop, but they're all 2GB sticks.

12:08 <quail> I have 4GB in my desktop too

12:09 <yomshleeshee> brb

12:10 <zombie67> these are FB DIMMs, so even more expensive.  Still, WAY cheaper than they used to be.  I got tired of BOINC bringing my machine to a grinding hault.

12:10 <zombie67> halt

12:10 <MTughan> Xeon procs?

12:10 <zombie67> yeah

12:10 <zombie67> only way to get to 8-way with a mac

12:11 <MTughan> Only ones I know of besides some Opterons that'd require FB DIMMS.

12:11 <zombie67> no, FB DIMM is an intel spec.  I seriously doubt ANY AMDs use FB DIMMs.  Also, AMD uses hypertransport.

12:11 <MTughan> Ah.

12:12 <zombie67> another spec that Intel created and then killed within a single generation.  Thanks.

12:12 <yomshleeshee> back

12:12 <zombie67> nehalem based xeons are ddr3

12:12 <zombie67> normal dimms

12:12 <desti> fbdimms are also used by suns sparc

12:12 <zombie67> yep

12:13 <zombie67> we went with that design when intel told everyone that they were the wave of the future

12:13 <zombie67> and now Sun will be left holding the bag

12:15 <quail> zombie67: no it amiga, lmao

12:18 *** mnx has joined #boinc

12:20 *** jasongopher1 has joined #boinc

12:29 *** wdsmia has joined #Boinc

12:29 *** wdsmia_ has quit IRC

12:30 <MTughan> PovAddict: Wow, nice chain of events on the kernel exploit there...

12:30 <quail> where?

12:31 <MTughan> Pov posted a link on how the recent kernel exploit discovered occurred.

12:31 <Romulus> Title: Fun with NULL pointers, part 1 [] (at

12:31 *** wdsmia has quit IRC

12:31 <MTughan> I think it was the recent one anyway.

12:31 *** wdsmia has joined #Boinc

12:32 <yomshleeshee> "This vulnerability, which affects the 2.6.30 kernel "

12:32 <yomshleeshee> this should be interpreted as it's is specific to 2.6.30, correct?

12:32 <yomshleeshee> as opposed to <= 2.6.30

12:33 <MTughan> I guess so. It's apparently already been fixed.

12:33 <yomshleeshee> oh, that's good

12:33 <yomshleeshee> not that I'm even 2.6.30 up-to-date. ha!

12:34 <MTughan> Actually, I think this is a different one... There was an exploit revealed a couple days ago. A bug that apparently affected all 2.4 and 2.6 kernels dating back to 2001 allowed someone to run code with kernel priveleges.

12:34 *** wdsmia_ has joined #Boinc

12:34 <yomshleeshee> oh, this isn't the same

12:34 <MTughan> Although it's similar.

12:34 <yomshleeshee> hmmm

12:38 *** jasongopher has quit IRC

12:49 *** wdsmia has quit IRC

12:50 <PovAddict> MTughan: that link is *not* the recently-discovered vulnerability

12:50 <MTughan> So I figured out afterwards. :P

12:50 <yomshleeshee> meaning the 2.4 & 2.6 kernel vulnerability

12:51 <PovAddict> but many details about the vuln and the exploit (null pointer deference, using mmap, abusing SVR4 personality, using a suid root program like pulseaudio) are similar

12:51 <PovAddict> so it's... educational

12:52 <MTughan> I find it particularly interesting that GCC removes the NULL check.

12:52 <PovAddict> the idea is that if tun is NULL, then tun->sk would have already segfaulted

12:53 <PovAddict> so the if (!tun) wouldn't be reached

12:53 <MTughan> But in there, NULL is valid as the pointer to 0x0.

12:53 <PovAddict> and if it's not null, then it's obvious what the result of the check is, so it's optimized away

12:54 <MTughan> So should it be checking for NULL and 0x0?

12:54 <PovAddict> NULL is 0

12:54 <PovAddict> by a #define, so they are indistinguishable by the compiler

12:54 <PovAddict> (also, (!tun) isn't even using 'NULL' explicitly)

12:55 <PovAddict> thing is, if you *couldn't* mmap a page of memory into 0x0, there would be still a bug

12:55 <PovAddict> if tun is null, you get a kernel panic at tun->sk

12:59 *** Aeternus has quit IRC

12:59 <MTughan> Better than an exploit.

12:59 <PovAddict> well, that's a denial of service..

12:59 <MTughan> Yeah, but you're not running kernel level code.

13:00 <PovAddict> get a shared hosting account, do the needed magic to cause that kernel code to be run with tun=null, and...

13:00 <PovAddict> normally, a non-root user can't crash the entire system

13:00 <MTughan> Would you like the person to be able to get total control over the machine instead?

13:00 <PovAddict> (root easily can... heck, you can corrupt kernel memory from the shell)

13:01 <MTughan> It may not be the best situation, but it's preferable to an exploit IMO.

13:01 <PovAddict> of course ring0 code execution is worse than a denial of service

13:01 <PovAddict> but checking for tun nullness before doing tun->sk would solve both situations

13:01 <MTughan> Yep.

13:01 <PovAddict> (for this particular piece of code9

13:01 <MTughan> Line of code was just a little too high.

13:07 *** wdsmia_w has joined #boinc

13:47 *** mnx has quit IRC

13:48 *** siofwolves has joined #boinc

13:50 *** Aeternus has joined #boinc

15:00 *** siofwolves has quit IRC

15:15 *** linear_shift has quit IRC

15:17 *** linear_shift has joined #boinc

15:41 *** kathryn_ has joined #boinc

15:43 *** freakazoid0223 has quit IRC

15:46 *** KathrynM has quit IRC

15:55 *** teh_failerer has joined #boinc

15:56 <teh_failerer> hi

15:59 <teh_failerer> boinc does not detect my gpu :( i'm on jaunty with a gts250 and 190.x driver (cuda 2.3)

16:01 <PovAddict> what version of BOINC?

16:01 <teh_failerer> 6.2.18

16:01 <PovAddict> that version has no GPU support at all

16:01 <MTughan> You'll need 6.4.x or later for GPU.

16:02 <teh_failerer> ok thank you

16:03 *** yomshleeshee has quit IRC

16:06 <teh_failerer> it works !

16:23 *** teh_failerer has left #boinc

16:31 *** wdsmia_w has quit IRC

16:42 <xcamel> 'lo

16:42 <MTughan> 'lo Jay

16:43 <xcamel> Hio Mike

16:49 *** efc has joined #boinc

16:54 <xcamel> &wx 12074

16:54 <Romulus> xcamel: Temperature: 88.2F / 31.2C | Humidity: 59% | Pressure: 30.11in / 1019.5hPa (Falling) | Conditions: Clear | Wind Direction: WSW | Wind Speed: 10.0mph / 16.1km/h ; Tonight - Clear. Patchy fog after midnight. Lows in the lower 60s. Southwest winds around 5 mph this evening...becoming light and variable.; Saturday - Patchy fog in the morning. Mostly sunny. Highs in the upper 80s. Southwest (1 more message)

16:54 <xcamel> &more

16:54 <Romulus> xcamel: winds around 5 mph.; Saturday Night - Clear. Patchy fog after midnight. Lows in the lower 60s. Southwest winds around 5 mph in the evening...becoming light and variable.;

16:55 <xcamel> toasty.

16:55 <MTughan> Wow...

16:55 <MTughan> &wx YOW

16:55 <Romulus> MTughan: Temperature: 86°F / 30°C | Humidity: 43% | Pressure: 30.12in / 1020hPa | Conditions: Partly Cloudy | Wind Direction: SW | Wind Speed: 8mph / 13km/h | Updated: 4:00 PM EDT; Tonight - Clear. Low 15C(59F).; Saturday - Sunny. High 31C(88F). UV index 8 or very high. Saturday night clear. Low 15C(59F).; Sunday - Sunny. High 31C(88F).;

16:55 <efc> Moo

16:55 <MTughan> Glad I'm inside... May be going underground to dinner.

17:02 <efc> I'm going to try to close a bank account. Applying crash helmet now.

17:03 *** yoyo[RKN] has joined #boinc

17:04 <efc> You around Pov ?

17:05 <PovAddict> sorta

17:10 <efc> Was wondering if there was a way to locate an Atom Ion in the Seti@home records and see how they do..

17:11 <efc> Since you are the DB guru..

17:12 <PovAddict> would this help for a start?

17:12 <Romulus> <> (at

17:12 <efc> I don't know how to pick an Atom out of there

17:13 <efc> 1000s of records..

17:13 <efc> Ideally I'd like to find a record in the seti@home computer page thingies

17:15 <efc> Like mine here

17:15 <efc>

17:15 <Romulus> <> (at

17:18 <xcamel> hey pov...

17:19 <PovAddict> how do I tell it's an Ion?

17:19 <PovAddict> N270?

17:20 <efc> Possible that it has a geforce 9300 (i think thats the number)

17:20 <PovAddict> this bastard estimates 1 hour download time remaining

17:21 <PovAddict> so far only 39 machines with a reported GPU at all

17:21 <efc> Yeah, I can't imagine many Atoms having one

17:21 <wdsmia_> re-hi

17:21 <PovAddict> I mean globally, not only Atoms

17:22 *** desti_T2 has joined #boinc

17:22 <efc> Oh. Thats alarming.

17:22 *** wdsmia_ is now known as wdsmia

17:22 <PovAddict> that's how slow it was downloading, I guess :)

17:22 <efc> moo wdsmia

17:22 <PovAddict> I had downloaded like 1% of the DB by then

17:22 <PovAddict> SPD:122.61KiB/s ETA:50m43s

17:23 <efc> Don't bother doing an hour download for this.. I thought you might have one laying around

17:23 <PovAddict> not from SETI

17:23 <efc> I just haven't seen any decent measurements of what it does as a cruncher

17:25 <PovAddict> no Atoms yet

17:26 <efc> If its in sequential order

17:26 <efc> they'll mostly be at the end..

17:26 <efc> There are around 5 million hosts (as you can see from the ID above)

17:27 <PovAddict> they seem mostly sequential, yes

17:27 <efc> I'd search myself but I don't know much of a way to navigate the S@H data

17:27 <PovAddict> one Atom found, no <coprocs>

17:28 <PovAddict> host created Sept 2004, heh

17:28 <efc> Thats an early adopter.. Very early

17:28 <PovAddict> did atoms exist back then?

17:29 <PovAddict> I'm thinking he might have upgraded the CPU and kept the host record

17:29 <PovAddict> after all, that host is running Windows Server 2008...

17:29 <PovAddict> so at least the OS upgrade didn't cause a new hostID

17:30 <efc> Atoms came out in 2007 I think

17:30 <efc> You see the occasional engineering sample out before official release. But not that far before.

17:35 <efc> Found a few people asking about the combo on the boards

17:35 *** yoyo[RKN] has quit IRC

17:36 <PovAddict> SPD:115.68KiB/s ETA:39m53s

17:37 <efc> I suppose I could be the first and post my results..

17:38 *** desti has quit IRC

17:39 *** infinisoft has joined #boinc

17:50 *** desti_T2 is now known as desti

17:58 <PovAddict> efc: an i7 with GeForce 9300

17:58 <Romulus> <> (at

17:59 <PovAddict> are you sure that's the number?

18:00 <MTughan> Wikipedia says the ION has a 9400M.

18:00 <PovAddict> looking

18:01 <xcamel> geez.. I just might get kyzyl rebuilt yet..

18:03 <PovAddict> the 'M' doesn't seem to show up in the coprocs tag

18:03 *** wdsmia has quit IRC

18:03 <PovAddict> how much RAM do those cards have?

18:03 <MTughan>

18:03 <Romulus> Title: NVIDIA ION - Wikipedia, the free encyclopedia (at

18:05 <PovAddict> I have 1087503 hosts so far

18:05 <PovAddict> but I'm downloading multithreaded

18:05 <PovAddict> so as soon as the first thread finishes, the number will suddenly double

18:06 <PovAddict> as it will include what is downloaded so far of the second chunk

18:12 <PovAddict> I'll be done in about 20 seconds

18:14 <PovAddict> efc: *your* machine doesn't report its GPU

18:32 <efc> Back

18:34 <efc> maybe my boinc version is too old

18:34 <efc> 5.10.45

18:34 <PovAddict> doesn't even know what a GPU is

18:34 <PovAddict> bah, I can't just load this into a text editor and search for things...

18:35 <PovAddict> zombie67 could, though, on his machine with 20GB of RAM :P

18:38 <MTughan> PovAddict: I have 8GB here.

18:38 <PovAddict> I have 2GB physical, and this is 2.6GB of XML...

18:39 <MTughan> I mean I can search for something if you'd like.

18:43 *** Rookie_69 has joined #boinc

18:44 <Rookie_69> &wx yaw

18:44 <Romulus> Rookie_69: Temperature: 69.6°F / 20.9°C | Humidity: 58% | Pressure: 30.06in / 1017.8hPa | Conditions: Clear | Wind Direction: WSW | Wind Speed: 7.6mph / 12.2km/h | Updated: 7:44 PM ADT; Tonight - A few clouds. Fog patches developing overnight. Low 15C(59F).; Saturday - Sunny. Fog patches dissipating in the morning. High 32C(90F) except 24C(75F) along the coast. UV index 8 or very high. Saturday (1 more message)

18:44 <Rookie_69> &more

18:44 <Romulus> Rookie_69: night a few clouds. Low 18C(64F).; Sunday - Sunny. High 27C(81F).;

18:46 <Rookie_69> Wait... They forgot to say rain. After all, it is the weekend coming up.

18:55 *** Rookie_69 has quit IRC

19:00 <xcamel> 'lo Glenn

19:06 *** wdsmia has joined #Boinc

19:42 <wdsmia> &ping

19:42 <Romulus> pong

20:07 *** Zeeno has joined #boinc

20:15 *** Zeeno has quit IRC

20:27 *** Aeternus has quit IRC

20:40 *** wdsmia_ has joined #Boinc

20:40 *** wdsmia has quit IRC

20:46 <PovAddict>

20:46 <Romulus> Title: Changeset 18846 - BOINC - Trac (at

20:46 <PovAddict> ATI GPU detection code added

20:46 <wdsmia_> about time

20:47 <MTughan> Not written by BOINC people though.

20:50 <MTughan> Also supports only 2 ATI cards.

20:51 <MTughan> From the looks of it, anyway...

20:52 *** wdsmia has joined #Boinc

20:53 <MTughan> Oh, nvm... Weird naming of variables.

20:53 *** wdsmia has quit IRC

20:57 *** wdsmia has joined #Boinc

21:08 *** wdsmia_ has quit IRC

21:13 <efc> this is naive of me

21:13 <efc> but isn't naked sprintf bad?

21:13 <MTughan> "naked"?

21:14 * linear_shift watches C porn.

21:14 <MTughan> Make a useful C program that looks like ASCII porn. :P

21:14 <efc> ie, not protected against buffer overflow

21:15 <efc> I thought you didn't use strcat, sprintf, etc in anything remotely public-facing

21:15 <efc> there are length restricted versions of all those

21:15 <linear_shift> Yeah, like fprintf.

21:16 <linear_shift> Just write to stdout.

21:16 <MTughan> That prints to a file.

21:16 <MTughan> I think the length limited version is something like snprintf.

21:16 <efc> strcatn, etc

21:16 <MTughan> Yeah, snprintf.

21:17 <PovAddict> efc: you haven't read much of BOINC code, have you?

21:17 <efc> not lately

21:17 <efc> Maybe it doesn't matter unless in a web server or kernel or something

21:17 <efc> if you have to own the machine to use the exploit, its not much of an exploit

21:17 <MTughan> int snprintf(chat *restrict c, size_t n, const char *restrict format, ...);

21:17 <MTughan> From "man snprintf"

21:18 <linear_shift> MTughan: You can print to stdout, same deal as printf.

21:18 <linear_shift> With fprintf.

21:18 <MTughan> linear_shift: Sure you can, but that's not the length limited version.

21:18 <MTughan> sprintf formats a string, but instead of sending it to an output device, it stores it in a string.

21:19 <MTughan> snprintf does the same thing, but truncates the string to size_t n bytes if longer than n.

21:19 <linear_shift> Oh.

21:19 <linear_shift> Heh.

21:19 <efc> I haven't written much c lately.. just every time that an exploit is discussed on slashdot, using the non-checked version is looked on with scorn and derision

21:20 <linear_shift> I thought for some reason fprintf had a length restriction. *-.-*

21:20 <MTughan> efc: If you can know how long the string is, sprintf is fine. But if you're using any user-given input in it, you should definitely use snprintf instead.

21:20 <MTughan> Even better, if you're using C++, use std::string.

21:22 <efc> I'm still in counseling after using the C++ standard library

21:22 <MTughan> Or an interesting printf method I've never heard of before... aprintf.

21:22 <MTughan> Apparently, allocates a new string for you.

21:23 <MTughan> Sorry, asprintf.

21:24 <MTughan> Ah, GNU extensions, also available in *BSD.

21:24 <PovAddict> linear_shift: fprintf has no problems with length because it's not writing to a fixed-size buffer

21:24 <PovAddict> well, it is...

21:24 <efc> fixed to the size of your hard disk hehe

21:24 <PovAddict> but it already checks for it, and returns the proper error code: disk full :)

21:24 * PovAddict -> dinner

21:25 <MTughan> asprintf looks to be a useful method, but snprintf is a bit more portable.

21:27 <efc> $70 on a black toner cartridge that lasts 2,000 pages is bad, right?

21:28 <efc> our old printer was about $200 for 50,000 pages

21:35 *** wdsmia_ has joined #Boinc

21:35 *** wdsmia has quit IRC

21:37 <CoderForLife> just damn

21:38 <CoderForLife> blown partition table

21:38 * CoderForLife falls into bed

21:44 *** wdsmia has joined #Boinc

21:48 <PovAddict> MTughan: asprintf uses a nonstandard feature of the *compiler*

21:48 <MTughan> So you just need to compile it with gcc and it'll work on any platform?

21:48 <MTughan> s/platform/OS/

21:49 <PovAddict> yes

21:49 <MTughan> Nice.

21:49 <PovAddict> it does some messing with the stack pointer that you usually can't do from C

21:50 <PovAddict> oh wait, that's not at all what I thought it was

21:51 <PovAddict> I thought asprintf used alloca() (which is gcc only)

21:55 *** wdsmia_ has quit IRC

21:57 *** freakazoid0223 has joined #boinc

22:02 <PovAddict> asprintf depends on GNU libc, not on the compiler

22:02 <PovAddict> which means it won't work on Windows (which uses msvcrt)

22:02 <MTughan> That's what I had thought originally.

22:03 <PovAddict> I thought asprintf used alloca, because of its name

22:03 <PovAddict> not malloc

22:03 <MTughan> I think the a just stands for a generic allocate.

22:03 <MTughan> allocating sprintf?

22:03 <PovAddict> I have actually seen functions in GNU libc that start with 'a' and use alloca

22:03 <PovAddict> and give the usual warnings about not using it inside a function call

22:04 <PovAddict> like bar(afoo())

22:04 <PovAddict> because afoo calls alloca, and that makes a mess with the arguments

22:05 <MTughan> I hadn't heard of alloca before, but I could see how it could be problematic if it messes with the stackframe.

23:00 *** wdsmia_ has joined #Boinc

23:00 *** wdsmia has quit IRC

23:06 *** wdsmia has joined #Boinc

23:06 *** wdsmia_ has quit IRC

23:10 *** wdsmia_ has joined #Boinc

23:10 *** wdsmia has quit IRC

23:12 *** CoderForLife has quit IRC

23:16 <efc> Another bank kablooey

Generated by 2.4 by Marius Gedminas - find it at!