Re: double checking Debian's 686-pae vs. -486 probe
On Tue, 2011-05-24 at 07:59 +0800, jida...@jidanni.org wrote: Recently Debian sid split the kernel into these two packages, linux-image-2.6.39-1-486_2.6.39-1_i386.deb linux-image-2.6.39-1-686-pae_2.6.39-1_i386.deb My Thinkpad ended up being told by the installation scripts to use -486. But my Thinkpad isn't really very old... how can one double check to see if the decision was correct? I can't figure it out even after probing the .debs to find just where they probe the choice. The clue is in the name. And expanded in the description, in case you don't know what PAE is. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Re: double checking Debian's 686-pae vs. -486 probe
BH == Ben Hutchings b...@decadent.org.uk writes: BH The clue is in the name. And expanded in the description, in case you BH don't know what PAE is. $ apt-cache show linux-image-2.6.39-1-686-pae|grep Celeron supported by the Intel Pentium Pro/II/III/4/4M/D, Xeon, Celeron, Core and That's me. Celeron. Odd how whatever script it was told me I had to use -486. I looked in the preinst and postinst parts of /var/cache/apt/archives/linux-image-2.6.39-1-686-pae_2.6.39-1_i386.deb to try to find what probed my Celeron and said it couldn't use 686-pae and had to use -486, but it is onebigblack mystery so sorry. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87ipt0ze1n@jidanni.org
Re: double checking Debian's 686-pae vs. -486 probe
Le mardi 24 mai 2011 à 16:41 +0800, jida...@jidanni.org a écrit : BH == Ben Hutchings b...@decadent.org.uk writes: BH The clue is in the name. And expanded in the description, in case you BH don't know what PAE is. $ apt-cache show linux-image-2.6.39-1-686-pae|grep Celeron supported by the Intel Pentium Pro/II/III/4/4M/D, Xeon, Celeron, Core and That's me. Celeron. Odd how whatever script it was told me I had to use -486. How about trying a “grep pae /proc/cpuinfo” instead of looking for approximate CPU descriptions? -- .''`. Josselin Mouette : :' : `. `' `- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1306227621.3872.275.camel@pi0307572
Re: double checking Debian's 686-pae vs. -486 probe
JM == Josselin Mouette j...@debian.org writes: JM How about trying a “grep pae /proc/cpuinfo” instead of looking for JM approximate CPU descriptions? And indeed lshw says product: Intel(R) Celeron(R) M processor 1.40GHz capabilities: fpu fpu_exception wp vme de pse tsc msr mce cx8 sep mtrr pge mca cmov clflush dts acpi mmx fxsr sse sse2 ss tm pbe up bts I.e., no pae. So I would add often in the Description: This kernel requires PAE (Physical Address Extension). This feature is **often** supported by the Intel Pentium Pro/II/III/4/4M/D, Xeon, Celeron, Core and Atom; AMD Geode NX, Athlon (K7), Duron, Opteron, Sempron, Turion or Phenom; Transmeta Efficeon; and VIA C7. or may be or often available or something like that. Also the -486 Description should mention the name of the -686-pae package. Say, for the sake of sharing a single .deb offline, what bad thing might happen if I run the -486 kernel on my other machines where I could run the -pae kernel? Namely other single Celerons having the pae capability. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87ei3ozbon@jidanni.org
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, May 24, 2011 at 05:32:08PM +0800, jida...@jidanni.org wrote: JM How about trying a “grep pae /proc/cpuinfo” instead of looking for JM approximate CPU descriptions? And indeed lshw says product: Intel(R) Celeron(R) M processor 1.40GHz capabilities: fpu fpu_exception wp vme de pse tsc msr mce cx8 sep mtrr pge mca cmov clflush dts acpi mmx fxsr sse sse2 ss tm pbe up bts I.e., no pae. So I would add often in the Description: Or specifically mention early Centrino CPUs not having PAE support. -- WBR, wRAR signature.asc Description: Digital signature
Re: double checking Debian's 686-pae vs. -486 probe
AR == Andrey Rahmatullin w...@wrar.name writes: AR Or specifically mention early Centrino CPUs not having PAE support. Celeron™ too! So just say may or might like the pros do. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/871uzoz6we@jidanni.org
Re: [ltp] Re: double checking Debian's 686-pae vs. -486 probe
On mar., 2011-05-24 at 09:58 -0300, Stefan Monnier wrote: Say, for the sake of sharing a single .deb offline, what bad thing might happen if I run the -486 kernel on my other machines where I could run the -pae kernel? Nothing serious, unless they have a lot of RAM (in which case the kernel may not be able to make use of all of it). And you lose NX too. Regards, -- Yves-Alexis -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1306243091.25324.16.camel@oban
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, May 24, 2011 at 07:15:29PM +0800, jida...@jidanni.org wrote: AR == Andrey Rahmatullin w...@wrar.name writes: AR Or specifically mention early Centrino CPUs not having PAE support. Celeron™ too! So just say may or might like the pros do. Early Celerons do have pae: (a Pentium 2 class one): flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pse36 mmx fxsr the one mentioned earlier was a Celeron M. But in general, Celeron is an insanely overused brand, it carries no information other than a crippled version of a CPU from Intel as it can come from about any other brand they make. I think it might be far easier to list CPUs that do not have pae or might not have it rather than the other way around. It appears that at least as Intel is concerned, the only processors =686 without pae are certain models of Pentium M (and thus relevant Celerons as well). Not sure about AMD and the rest. -- 1KB // Microsoft corollary to Hanlon's razor: // Never attribute to stupidity what can be // adequately explained by malice. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110524143454.ga26...@angband.pl
Re: [ltp] Re: double checking Debian's 686-pae vs. -486 probe
Say, for the sake of sharing a single .deb offline, what bad thing might happen if I run the -486 kernel on my other machines where I could run the -pae kernel? Nothing serious, unless they have a lot of RAM (in which case the kernel may not be able to make use of all of it). And you lose NX too. I wouldn't lose any sleep over it. Stefan -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/jwvmxicfauw.fsf-monnier+gmane.linux.hardware.think...@gnu.org
Re: double checking Debian's 686-pae vs. -486 probe
Adam Borowski, le Tue 24 May 2011 16:34:55 +0200, a écrit : But in general, Celeron is an insanely overused brand, it carries no information other than a crippled version of a CPU from Intel as it can come from about any other brand they make. Indeed. The name could probably be simply dropped from the list. Samuel -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110524144824.gg4...@const.bordeaux.inria.fr
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, 2011-05-24 at 17:05 +0600, Andrey Rahmatullin wrote: On Tue, May 24, 2011 at 05:32:08PM +0800, jida...@jidanni.org wrote: JM How about trying a “grep pae /proc/cpuinfo” instead of looking for JM approximate CPU descriptions? And indeed lshw says product: Intel(R) Celeron(R) M processor 1.40GHz capabilities: fpu fpu_exception wp vme de pse tsc msr mce cx8 sep mtrr pge mca cmov clflush dts acpi mmx fxsr sse sse2 ss tm pbe up bts I.e., no pae. So I would add often in the Description: Or specifically mention early Centrino CPUs not having PAE support. The Centrino brand refers to a combination of Pentium M, Intel chipset and approved WLAN adapter which all have good power-saving facilities. It doesn't refer to a processor. You are correct that most Pentium M processors do not have PAE (or do not advertise it in CPU feature flags). I didn't know some of the Pentium Ms were also called 'Celeron'; I'll see if I can rephrase the description accordingly. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Re: double checking Debian's 686-pae vs. -486 probe
jida...@jidanni.org writes: Say, for the sake of sharing a single .deb offline, what bad thing might happen if I run the -486 kernel on my other machines where I could run the -pae kernel? Namely other single Celerons having the pae capability. Nothing happens. You just won't be able to use pae capabilities. Which means you are limited to ~3.5GiB physical memory. MfG Goswin -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87vcx0dpwi.fsf@frosties.localnet
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, May 24, 2011 at 08:58:43AM -0700, Ben Hutchings wrote: JM How about trying a “grep pae /proc/cpuinfo” instead of looking for JM approximate CPU descriptions? And indeed lshw says product: Intel(R) Celeron(R) M processor 1.40GHz capabilities: fpu fpu_exception wp vme de pse tsc msr mce cx8 sep mtrr pge mca cmov clflush dts acpi mmx fxsr sse sse2 ss tm pbe up bts I.e., no pae. So I would add often in the Description: Or specifically mention early Centrino CPUs not having PAE support. The Centrino brand refers to a combination of Pentium M, Intel chipset and approved WLAN adapter which all have good power-saving facilities. It doesn't refer to a processor. I know, Centrino is just easier to remember than 'Pentium M'. You are correct that most Pentium M processors do not have PAE (or do not advertise it in CPU feature flags). I didn't know some of the Pentium Ms were also called 'Celeron'; I'll see if I can rephrase the description accordingly. Wikipedia doesn't have it either, But Intel site has: http://ark.intel.com/ProductCollection.aspx?codeName=1788 (Banias) http://ark.intel.com/ProductCollection.aspx?codeName=2643 (Dothan; some of them support PAE, but apparently no Celeron ones) -- WBR, wRAR signature.asc Description: Digital signature
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, 2011-05-24 at 18:28 +0200, Goswin von Brederlow wrote: jida...@jidanni.org writes: Say, for the sake of sharing a single .deb offline, what bad thing might happen if I run the -486 kernel on my other machines where I could run the -pae kernel? Namely other single Celerons having the pae capability. Nothing happens. You just won't be able to use pae capabilities. Which means you are limited to ~3.5GiB physical memory. Also no SMP. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Re: double checking Debian's 686-pae vs. -486 probe
Ah, http://en.wikipedia.org/wiki/NX_bit http://en.wikipedia.org/wiki/Physical_Address_Extension http://en.wikipedia.org/wiki/Symmetric_multiprocessing you know I am learning more and more from you fellows every day. I have some specific recommendations: forget about listing individual brands, just say available in many newer CPUs, check your /proc/cpuinfo . Nothing happens. You just won't be able to use pae capabilities. Which means you are limited to ~3.5GiB physical memory. BH Also no SMP. Nor NX. Excellent to know. Do mention that on the Description. Knowing that I can now switch my sneakernet to a pure -486 implementation, as I won't be missing a thing. 3.5GiBs of physical memory? They're not as cheap as gum sticks, you know. Speaking of which, http://www.youtube.com/user/jidanni2#p/c/6E40919035151385/10/hB2b3Ce5op0 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8739k3ve3m@jidanni.org
double checking Debian's 686-pae vs. -486 probe
Recently Debian sid split the kernel into these two packages, linux-image-2.6.39-1-486_2.6.39-1_i386.deb linux-image-2.6.39-1-686-pae_2.6.39-1_i386.deb My Thinkpad ended up being told by the installation scripts to use -486. But my Thinkpad isn't really very old... how can one double check to see if the decision was correct? I can't figure it out even after probing the .debs to find just where they probe the choice. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8739k5ynn6@jidanni.org
Re: double checking Debian's 686-pae vs. -486 probe
On Tue, 24 May 2011, jida...@jidanni.org wrote: But my Thinkpad isn't really very old... how can one double check to see if the decision was correct? I can't figure it out even after probing the .debs to find just where they probe the choice. Why not just install the other kernel and try booting it? -- My Main Blog http://etbe.coker.com.au/ My Documents Bloghttp://doc.coker.com.au/ -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105241037.40889.russ...@coker.com.au
Re: double checking Debian's 686-pae vs. -486 probe
RC == Russell Coker russ...@coker.com.au writes: RC Why not just install the other kernel and try booting it? Maybe it will lead to subtle data loss. You never know. That's why I was hoping to dig out of the .debs just how they determine which Thinkpads are hip, and which to skip. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqn8yj1f@jidanni.org
Re: double checking Debian's 686-pae vs. -486 probe
jida...@jidanni.org writes: RC == Russell Coker russ...@coker.com.au writes: RC Why not just install the other kernel and try booting it? Maybe it will lead to subtle data loss. You never know. That's why I was hoping to dig out of the .debs just how they determine which Thinkpads are hip, and which to skip. No. The kernels check the cpu features on boot. They either work or not. As for your initial question: Depending on how you install the installer has different sets of kernels to choose from. Due to space not all kernels are on all images so sometimes you get the fallback option because the right one simply isn't available. Also try the -amd64 kernel. If supported that is generally preferable to 686-pae. MfG Goswin -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87aaec3hia.fsf@frosties.localnet