Is AMD64 page out of date about W^X?
According to: http://www.openbsd.org/amd64.html W^X will not work on Intel's 64 bit chips. I for one chose to go with i386 on my Core 2 because of this fact alone. Then I saw this: http://www.xbitlabs.com/news/cpu/display/20041011182310.html and scores of other pages that refer to the XD bit. On another page (sorry.. lost the link) a person claimed that newer chips (I believe around the beginning of 2005) started shipping it and that it would work on OpenBSD. As this person is who-knows-who I can't really put much to that. http://processorfinder.intel.com/ shows the Core 2's having "Execute Disable Bit" Then I see dmesg like the following: http://www.webservertalk.com/archive249-2007-1-1783328.html http://readlist.com/lists/openbsd.org/misc/10/54208.html Here next to the Core 2 we see NXE So I dunno... it looks to me like it is supported. I haven't had a chance to look at how the code functions... but would the kernel use W^X based on NXE being available? Or does it have some other code that might see it is Intel (or use some particular method of checking for the bit that might not work on Intel's implementation). Basically... I can't confirm if it works or not. And if I switch my server over to AMD64... will NXE in the dmesg really let me know that it is indeed working correctly? If indeed W^X is now supported on newer Intel chips... could someone update the AMD64 page? I know that when I was buying my current hardware I considered going AMD for this comp because I saw that. Then the prices fell on Core 2's and I went ahead with it because it does indeed seem faster. I know that Intel has been lame by not giving good documentation and perhaps this could sting them back a bit by putting people off. But it seems at this point (2-3 years after they started adding NXE) it would be good to go ahead and say if it is supported. Perhaps I am wrong and their version of NXE is really a bunch of bull and is not applicable, making the statement true. But if indeed it is supported on newer chips... it seems fair to be honest about it so users of OpenBSD don't make uninformed decisions.
Re: Is AMD64 page out of date about W^X?
I am not so sure of that.If you go here: http://processorfinder.intel.com/Default.aspx and then select Core 2 Duo or some such... then filter by "Execute Disable Bit" under supported features... you will see a bunch of Core 2s. The Core 2 is ia32e. It is not EM64T. According to some sites... if anyone really cares I will find the links... Intel started putting ia32e chips out in late 2004.. this includes some Pentium 4's. I believe Theo was expressing his disappointment around Feb of 2004. Anyway... the page does seem to be updated semi-regularly.. if the date at the bottom is accurate.. it was last changed on 2007/08/10 If Intel did indeed start including it on chips in early 2005... it would be nice to know that instead of a blanket statement that Intel does not support the NXE bit at all. It is important when making purchasing decisions and architecture choices. Darren Spruell <[EMAIL PROTECTED]> wrote: > On 9/20/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > According to: > > http://www.openbsd.org/amd64.html > > W^X will not work on Intel's 64 bit chips. I for one chose to go with i386 > > on my Core 2 because of this fact alone. > > Intel produces 2 families of 64-bit processors; the EM64T and an AMD64 > family chip. You're probably misinterpreting what is meant to indicate > the former. > > http://en.wikipedia.org/wiki/64-bit#Current_64-bit_microprocessor_architectures > http://www.xbitlabs.com/news/cpu/display/20040310223922.html > > DS
Re: Is AMD64 page out of date about W^X?
Well I'll be durned.. apparently ia32e is EM64T(Intel's marketing name for it). I was thinking it was the itanium arch which is actually ia64. But either way... EM64T is supposed to run on AMD64... the only question is will OpenBSD respond accordingly when NXE is present during dmesg. And if so.. it would be nice to change this on the AMD64 page so people are aware of it. Darren Spruell <[EMAIL PROTECTED]> wrote: > On 9/20/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > According to: > > http://www.openbsd.org/amd64.html > > W^X will not work on Intel's 64 bit chips. I for one chose to go with i386 > > on my Core 2 because of this fact alone. > > Intel produces 2 families of 64-bit processors; the EM64T and an AMD64 > family chip. You're probably misinterpreting what is meant to indicate > the former. > > http://en.wikipedia.org/wiki/64-bit#Current_64-bit_microprocessor_architectures > http://www.xbitlabs.com/news/cpu/display/20040310223922.html > > DS
Re: Is AMD64 page out of date about W^X?
I sent a message and it looks like it got rejected... basically I found out that ia32e is EM64T(Intel's marketing name for it). I was thinking it was the itanium arch which is actually ia64. But either way... EM64T is supposed to run on AMD64... and it appears that the Intel chips do support the NXE bit since around 2005. Can anyone confirm that the newer ia32e chips (made after early 2005) are actually supporting W^X? It seems that just because NXE is shown in the dmesg wouldn't necessarily mean that OpenBSD would then use it. If it is indeed supported.. could someone change the message on the AMD64 page? http://www.openbsd.org/amd64.html It does seem useful to know this information for this platform and I have veered away from AMD64 for the last year because of it.
Re: Unable to map phys mem on Intel D945G motherboard
Your message header seems to point to an issue that has come up a few times. However.. your message body doesn't really give any good clues. I would suggest looking at a past post with the subject "OpenBSD 4.1 install issue??" from early May of this year. If my guess is correct.. when you use the boot cd.. it flashes text by very quickly and then halts? If it is indeed the same problem (I have the same motherboard) then you need to recompile the kernel. This is all detailed in the previously mentioned thread. You can email me directly if you are having problems. Girish Venkatachalam <[EMAIL PROTECTED]> wrote: > Dear friends, > > I am not able to produce a dmesg for you because neither the CD nor > the hard disk would boot on the cutting edge Intel D945G > motherboard. I tried changing the RAM with no effect. It is a brand > new motherboard. > > NetBSD does not boot either. > > But FreeBSD and linux work. > > Any clue? > > I have another motherboard of the same make on which an IDE hard > disk with an old OpenBSD install works fine but CD booting does not > work. > > I have three machines all of the same make. > > What could be going wrong? > > Thanks. > > regards, > Girish
Anyone seen the quantis rng available?
It looks pretty interesting and I know support for it has been worked on for OpenBSD. The only problem is that is seems next to impossible to find in the U.S. There site shows very few distributors and of the three emails that I have sent them over the last year... I have yet to hear from them. Someone did tell me that they are expensive. Anyone know of a source that can get them? What kind of prices are they running?
Re: assembler noob question
This link helped me when I was learning about x86 on OpenBSD... http://www.phiral.net/openbsdasm.htm I decided not to use the GCC __asm__ deal for various reasons that I don't remember. I think in particular.. I did not like the look at AT&T syntax vs. Intel syntax. I wrote an assembly version of AES, Serpent and Twofish algorithms using NASM under OpenBSD and everything worked. To me.. it seemed nice to be able to write a program in C that calls functions in a NASM created obj file. This seemed very portable and my code does indeed work under windows with only slight modifications. My project is here: http://asmaes.sourceforge.net/ Tobias Ulmer <[EMAIL PROTECTED]> wrote: > On Fri, Jul 25, 2008 at 01:12:58PM +, Matthew Szudzik wrote: > > On Thu, Jul 24, 2008 at 07:16:55PM -0600, Philip Guenther wrote: > > > If you're just trying to learn the x86 instruction set, then why not > > > put your code in an __asm__() block inside a C program? That lets the > > > compiler do all the heavy lifting. > > > > Is there a man page that describes the __asm__ block? The gcc man page > > only describes how to turn it on or off, and since it's nonstandard C, > > it's not described in any of the standard reference books like K&R. > > > > > > http://gcc.gnu.org/onlinedocs/gcc-4.3.0/gcc/Extended-Asm.html#Extended-Asm