| |
| |
| |
|
Page: 1 2 3 4 5 6
Comments:
<0> damn, geist.
<0> 'security guys break the Mac every single day. Every single day, they come out with a total exploit, your machine can be taken over totally. I dare anybody to do that once a month on the Windows machine.' --Bill Gates
<0> so pwned
<1> Once a month?
<2> Q: What does an Adam Vinatieri field goal have in common with Mike Vanderjagt?
<1> There were times I couldn't even leave my computer connected to the internet without a firewall (lets say directly after the installation process for Windows Update), without being totally ****ed up already by some exploits.
<3> JimPanic: there's still those times.
<1> I suppose so.
<3> probably they start being with windows vista after few weeks. ;)
<1> Though, with Vista there comes a reaaalllyyyy improved 'firewall'.
<3> wait and you'll see.
<1> But I think the firewall is more between the user and the PC, as opposed to actual firewalls that are for networking..
<3> I'm sure they have trice the bugs than there should have been, because it feels like vista's releasing were still a bit early for coders.
<3> JimPanic: you mean DRM? ;)
<1> Yea..the restrictions management :P
<1> Apart from that: time to market is very important. (No.)
<3> yes, they missed xm*** sales.
<1> They fail, nevertheless..
<1> Though, I like their advertisement. At least the one where no Vista or Microsoft logos are printed..
<3> I'm a bit relieved their product weren't better than what it is, it would have been ****ed to actually get DRM paradox or anything else like that.
<1> The introductory pictures, where some cool phrases are placed above some cool picture :D
<3> in other hand, if it were been better, there wouldn't been system wide DRM in it.
<4> my laptop came with a coupon for free upgrade to vista
<4> Ha
<1> nn: Sell it! :D
<4> JimPanic: who would buy that trash?
<3> good idea, put the misery forward! :D
<1> nn: Idiots?
<3> nn: there's always some idiots you can cheat.
<1> nn: I dunno. Just take the money. ;D
<4> i dont know anyone who would waste their money on that
<1> nn: Tell them it's the new SuSE/suse/Suse/SUSE/...
<1> Cheery: LOL
<4> JimPanic: they'd still say "eww NO! GO AWAY YOU TARD!"
<3> say it's XP.
<3> that'd may work.
<4> doubt it
<1> Mh
<1> Say it's pr0n
<4> 37M /
<4> Woo
<3> JimPanic: but it would then go out as boumerang, even 400kg fat ladies are more horny than vista.
<1> Don't say that........
<1> Because neither of both is ***y.
<1> Just. Not. ***y.
<3> well, those fat ladies would go mad if they'd hear what I said
<1> That's ok.
<3> because saying that so weakly is too light expression.
<3> vista is WAAAY less horny than them.
<3> ok, maybe that's enough from that comparison.
<4> hmm
<5> mmh
<6> hi
<5> ho
<4> w00t w00t w00t
<5> t00w t00w t00w
<4> that kid was annoying
<7> jump_protected: call gate.dpl < selector.rpl
<7> when i try to load the gdt
<8> exactly on the lgdt instruction? sure doesn't look like it
<9> hmm, every time when I modify the entry of GDT, should I reload GDT again ?
<8> only if you need to
<9> geist: ?
<8> well, okay. what exactly do you mean by 'reload the GDT'?
<9> geist: do you mean reload the gdt is unnecessary ?
<8> do you mean lgdt?
<9> geist: yes
<9> I mean lgdt
<8> ah. no you dont
<8> you only have to do that if you change the base of the gdt
<8> or length
<9> geist: oh, I see. so if I only modity a entry in the table, lgdt is unnecessary
<9> thank you ! geist
<8> lgdt is just setting an internal hidden register in the cpu
<9> geist: I just compare it with the TLB and pagetables
<8> if you change a gdt entry that is active, it wont take effect until you reload the segment descriptor though
<8> loading a segment descriptor causes a fetch from the gdt
<7> i think on the long jump
<8> iammisc: ah now, that makes sense
<7> how do you you do a far jump in at&t syntax anyway?
<8> a long jump reloads cs
<9> geist: good point !! so once I modify a entry, I have to reload the segment register by a descriptor, the new modification can come :-)
<7> i did mov $0x08,%ebx mov (far_ptr),%eax ljmp *(%ebx,%eax)
<9> thank you !
<7> is that right?
<8> no
<7> what should i do?
<8> that syntax is completely incorrect.
<7> i am using at&t syntax?
<8> also, iirc, the target segment can only be hard coded
<7> on gas
<8> hang on a minute and i'll get you an example
<7> it compiles fine
<7> ok thanks geist
<7> this is probably my problem
<8> sure, but it's not a long jump
<8> that's just a double indirect with some sort of indexing
<7> i got it from http://sig9.com/articles/att-syntax
<8> okay, it's something like
<8> ljmp $0x08, address
<8> the first bit is the segment register
<8> er the new code segment
<8> http://newos.org/WebSVN/filedetails.php?repname=NewOS&path=%2Fnewos%2Fboot%2Fpc%2Fi386%2Fbootblock.S&rev=0&sc=0
<8> see ljmp$0x8,$(code32) // flush prefetch queue and enter 32-bit mode
<8> though what i have is not a great example
<8> it shows you a bit of a bug in gas and 16-bit code
<8> i had to stick a size prefix in front of the ljmp to get it to work
<8> the .byte 0x66
<4> is it normal for my mucus to dissolve plastic?
<8> thansk for trying to derail the conversation
<7> now i get check_cs: not a valid code segment
<8> well, is 0x08 valid for your gdt?
<7> yes it should be
<4> geist: you might be interested, i'm updating the qemu-curses patch for qemu-cvs
<7> 0x9a in the access field is a code segment right?
<8> i think so
<7> then yes 0x08 should be a code segment
<8> then you might want to check on the lgdt stuff
<7> how do i get bochs to give more messages?
<8> make sure the gdt descriptor is okay
<8> iammisc: get a debug version, step through it line at a time
<8> it's good anyway, you can for exaple after the lgdt dump the internal state of the cpu, see if it loaded
<7> ok
<7> i'm going to have to do that tomorrow
<8> it's really the only useful part of bochs anyway
<7> if my gdt pointer is called gptr
<7> lgdt (gptr) should work right?
<8> not necessarilly
<8> depnds on the setup of the segments
<8> if you dont specify a segment override, I dont know which segment it by default works from. probably ds
<8> oh god, I forgot how much 16-bit code ****ed
<7> this is 32 bit code. i am being booted by grub
<5> mjam, Eierkuchen
<8> ah, well then you probably dont have to worry about that then
<7> so should that instruction work
<8> what instruction?
<7> lgdt (gdtr)
<7> or should it be lgdt gptr
<8> probably the latter
<7> ok
<8> also, what is in your gptr
<7> it is a packed c struct
<8> length first, i ***ume
<7> unsigned short limit, unsigned int base
<7> limit =(sizeof(gdt_entry)*3)-1
<7> base=gdt_entries
<7> where gdt_entries is a c array containing a packed representation of a gdt entry
Return to
#osdev
or
Go to some related
logs:
pecl-pdo is blocking
#sed
ubuntu make-kpkg suspend2
gentoo stage tarball uri gli
suse 10.1 spdif alsa
#centos
php get variable cross domain
zynaptix
#css
what does insecure string pickle
|
|