Hello all.
I'm looking to upgrade my Pooch"s kernel to 3.14 65
(it's currently the 5th entry at -- https://www.kernel.org --).
I've downloaded it and I did an initial compile. Looks ok (it reported no errors).
Q. #1) Where do I tick in "make gconfig" to get PAE to use a full 4 Gb of RAM?
(I'm unsure about the difference between SMP and PAE.)
Q. #2) Is it desirable (and ok) to compile as many elements as possible as
modules? The new 3.14.65 kernel has 4.4Mb vs 3.47Mb for the 3.14.56 original.
(I read somewhere that the boot process was more efficient with a smaller kernel.)
Q. #3) Is there any way to improve module recognition from within the config?
("The Pooch" has always had a problem with it.)
Q. #4) Finally, when I tried to boot, Puppy reported that it could not find
puppy_wheezy_3.14.56.1.sfs and dropped to a bare-bones terminal.
(I have to rename all sfs's related to Puppy proper to 3.14.65, I suppose?)
(Flash, if you think I should open a separate thread for each of those four
questions, just tell me.)
Many thanks in advance.
Where to tick in kernel's "make gconfig" to get x, y, z?
Where to tick in kernel's "make gconfig" to get x, y, z?
musher0
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
I'm grinning at your sig, there 
although I'm not quite sure I deserve it...
For the record -- SMP is "Symmetric Multi-Processing" is, as I understand it, the thing that lets a multi-core CPU in a computer be treated as a multiple-processor computer -- in other words, it lets you have four processors, each doing one or two things (depending on the presence or absence of another term called "multi-threading"), rather than being one processor that's only being say 1/4 used (if it's a four core proc). More info at Wikipedia.
PAE I can go into a little more detail about, because I understand it better. Here, this should do...
Er, credit where credit is due, some of the textual imagery here is borrowed full retail plus shipping from my old trusty and dusty CompTIA A+ Certification book by one Mike Meyers (no, not the actor) -- mine's a 4th or 5th Ed. and, while it's good for holding down a desk at some 1500+ pages, the guy is brilliant on his subject, and at explaining. It's not just a doorstop-grade tome, it's a resource! (For the record -- took the course, not the exams, because I can't memorize for crap -- so I don't have the Cert.)
Imagine that your computer's memory is a vast row of mailboxes. Each mailbox is a little small, such that it can only hold one letter at a time, and you can only open one mailbox in the row at any given time. Each mailbox is given a number, just like real life -- but the first mailbox is numbered 0, and the last, say, 65535.
Each mailbox in our metaphor is a single RAM location. Each location can be --must be, if it's to be useful-- individually addressed, and each can hold what's called in memory lingo a single "word" -- a group of bits with a consistent but arbitrary length. One RAM chip might be four bits per word, another might be eight. Modern ones are sixty-four, but in the spirit of keeping it simple, let's set our mailboxes to an eight-bit word. One byte in one mailbox at any time.
Now, let's look at how those addresses are implemented. If you google up a "pinout" (connection diagram) for a CPU called the 6502 -- it's a 1975 model, which is way outdated, but it's one I know decently, so bear with me here -- you will see a number of pins (connections) with numbers starting with "A" -- A0 through A15, to be specific. These sixteen wires coming off the 6502 are its "address" lines. They connect to the mailboxes so that when a given combination of address lines are active, it tells our mailman to go to a particular mailbox, open it, and either take out an eight-bit word or put one inside.
Worth noting -- if one puts a new word inside a box, without taking out the old word first, that old word disappears forever -- if it didn't, you'd get a combination of the old and new words, the term for that being "corrupt memory"
So here's the thing... a modern operating system that is described as a "32 bit" OS, has that bit length as its maximum number of usable address wires. The postman runs out of numbers for mailboxes at the 32-bits-all-active point, which is mailbox # 4,294,967,295...that's one overworked postman (!) but he still sometimes wants more of those mailboxes -- or, at least, his customers do.
Say that 32 bit operating system, happens to be on a computer with a 64-bit address bus coming out of its CPU. That means that it can address (2^64)-1 = 18,446,744,073,709,551,615 mailboxes -- that poor, poor postman! (Before you ask -- each additional address line, because of the way binary works, doubles the number of usable mailboxes.)
That said, if you're only using half of that 64-bit address bus (bus = set of wires with a common purpose), you only get the first 4,294,967,295 mailboxes, and none of the rest, because the postman runs out of numbers for boxes at Box # 4294967295 (remember, what we'd call Box # 1, the computer calls Box # 0, because computers start counting at zero). What PAE does is it lets the operating system use more address wires on that bus -- up to the full 64-bit "width" (that's the term) of the bus, if needed -- in order to use all the memory it has.
...make sense?
EDIT 3 APR 2016 -- Corrected a figure above. (2^64)-1 = 18,446,744,073,709,551,615, not 18,446,744,073,709,551,616. Thanks, ASD, for the correction, although I wish you'd replied in the thread and not via PM. Apologies to everyone for the incorrect figure that was there before.
although I'm not quite sure I deserve it...For the record -- SMP is "Symmetric Multi-Processing" is, as I understand it, the thing that lets a multi-core CPU in a computer be treated as a multiple-processor computer -- in other words, it lets you have four processors, each doing one or two things (depending on the presence or absence of another term called "multi-threading"), rather than being one processor that's only being say 1/4 used (if it's a four core proc). More info at Wikipedia.
PAE I can go into a little more detail about, because I understand it better. Here, this should do...
Er, credit where credit is due, some of the textual imagery here is borrowed full retail plus shipping from my old trusty and dusty CompTIA A+ Certification book by one Mike Meyers (no, not the actor) -- mine's a 4th or 5th Ed. and, while it's good for holding down a desk at some 1500+ pages, the guy is brilliant on his subject, and at explaining. It's not just a doorstop-grade tome, it's a resource! (For the record -- took the course, not the exams, because I can't memorize for crap -- so I don't have the Cert.)
Imagine that your computer's memory is a vast row of mailboxes. Each mailbox is a little small, such that it can only hold one letter at a time, and you can only open one mailbox in the row at any given time. Each mailbox is given a number, just like real life -- but the first mailbox is numbered 0, and the last, say, 65535.
Each mailbox in our metaphor is a single RAM location. Each location can be --must be, if it's to be useful-- individually addressed, and each can hold what's called in memory lingo a single "word" -- a group of bits with a consistent but arbitrary length. One RAM chip might be four bits per word, another might be eight. Modern ones are sixty-four, but in the spirit of keeping it simple, let's set our mailboxes to an eight-bit word. One byte in one mailbox at any time.
Now, let's look at how those addresses are implemented. If you google up a "pinout" (connection diagram) for a CPU called the 6502 -- it's a 1975 model, which is way outdated, but it's one I know decently, so bear with me here -- you will see a number of pins (connections) with numbers starting with "A" -- A0 through A15, to be specific. These sixteen wires coming off the 6502 are its "address" lines. They connect to the mailboxes so that when a given combination of address lines are active, it tells our mailman to go to a particular mailbox, open it, and either take out an eight-bit word or put one inside.
Worth noting -- if one puts a new word inside a box, without taking out the old word first, that old word disappears forever -- if it didn't, you'd get a combination of the old and new words, the term for that being "corrupt memory"
So here's the thing... a modern operating system that is described as a "32 bit" OS, has that bit length as its maximum number of usable address wires. The postman runs out of numbers for mailboxes at the 32-bits-all-active point, which is mailbox # 4,294,967,295...that's one overworked postman (!) but he still sometimes wants more of those mailboxes -- or, at least, his customers do.
Say that 32 bit operating system, happens to be on a computer with a 64-bit address bus coming out of its CPU. That means that it can address (2^64)-1 = 18,446,744,073,709,551,615 mailboxes -- that poor, poor postman! (Before you ask -- each additional address line, because of the way binary works, doubles the number of usable mailboxes.)
That said, if you're only using half of that 64-bit address bus (bus = set of wires with a common purpose), you only get the first 4,294,967,295 mailboxes, and none of the rest, because the postman runs out of numbers for boxes at Box # 4294967295 (remember, what we'd call Box # 1, the computer calls Box # 0, because computers start counting at zero). What PAE does is it lets the operating system use more address wires on that bus -- up to the full 64-bit "width" (that's the term) of the bus, if needed -- in order to use all the memory it has.
...make sense?
EDIT 3 APR 2016 -- Corrected a figure above. (2^64)-1 = 18,446,744,073,709,551,615, not 18,446,744,073,709,551,616. Thanks, ASD, for the correction, although I wish you'd replied in the thread and not via PM. Apologies to everyone for the incorrect figure that was there before.
Last edited by starhawk on Sun 03 Apr 2016, 20:23, edited 1 time in total.
HI starhawk.
Yep, makes sense. Incredibly clear explanation, too! Many thanks!
About the sig thing, I was considering adding your unofficial title:
BFN.
Yep, makes sense. Incredibly clear explanation, too! Many thanks!
About the sig thing, I was considering adding your unofficial title:
Like that, with the wink icon at the end? Is that ok?"resident philosopher | philosophe en résidence |
BFN.
musher0
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
Hi, starhawk.starhawk wrote:Shameless flattery will get you nowhere, musher0!
This is not flattery. It IS a very beautiful thought. Seriously! It also shows newbies
that this forum is not only technical, that real living and thinking people frequent
this forum.
Everybody should indeed think twice about closing their minds to events, people,
or ideas coming into their life and which they may not fully understand at present.
French philosopher Alain could have said something like this.
Or UK philosopher Alfred Whitehead.
Finally, "no good deed should go unpunished!"
So.... tough! It's done!
For the time being anyway! I do change my sigs every now and then.
If the "torture"
gets too much, tell me, and I'll make the sentence anonymous again. But I would like to give credit where credit is due.
BFN.
Last edited by musher0 on Wed 23 Mar 2016, 01:57, edited 1 time in total.
musher0
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
@ ASD -- a recent PM from you made me look back at this thread.
I rather think I owe you an apology... I was weary from typing that huge long post for musher0, and I was in a bad mood besides -- but that is by no means any reason or excuse for snapping at you, as I did above.
So -- please accept my humble apologies
If you would, please, suggest a correction, and I will edit the original post to fix the apparent typo. I already see what I probably did wrong -- forgot to subtract the one -- but I'll wait for you if I could remember how, I'd get out my trusty old TI-83+ and see if I could get something out of it that wasn't in scientific notation -- ah, but that has long been forgotten, if it was ever known to begin with.
Worth noting, the calculator I got that number from was a random Google hit that was not itself Google Calculator (from which I could extract only scientific notation, and I felt that such a result was too imprecise) -- so, while unlikely, it's possible that the entire number was incorrect, first digit through last. It's probably accurate enough to qualify as back-of-the-envelope type stuff (which is certainly well within the limits of acceptance here) but if we're going to get it right, let's get it right, you know?
Do be in touch
I rather think I owe you an apology... I was weary from typing that huge long post for musher0, and I was in a bad mood besides -- but that is by no means any reason or excuse for snapping at you, as I did above.
So -- please accept my humble apologies
If you would, please, suggest a correction, and I will edit the original post to fix the apparent typo. I already see what I probably did wrong -- forgot to subtract the one -- but I'll wait for you
if I could remember how, I'd get out my trusty old TI-83+ and see if I could get something out of it that wasn't in scientific notation -- ah, but that has long been forgotten, if it was ever known to begin with.Worth noting, the calculator I got that number from was a random Google hit that was not itself Google Calculator (from which I could extract only scientific notation, and I felt that such a result was too imprecise) -- so, while unlikely, it's possible that the entire number was incorrect, first digit through last. It's probably accurate enough to qualify as back-of-the-envelope type stuff (which is certainly well within the limits of acceptance here) but if we're going to get it right, let's get it right, you know?
Do be in touch
-
gcmartin
Hi Christian.
Just saw your appeal for kernel compilation assist. I wonder if @StemSee's utility would achieve what your needs are for your kernels. I also wonder if his work on having a BM which boots either the distro into 32bit (PAE) mode or 64bit mode as a user selectable item of a single distro.
If not, I understand. These are just some ideas where posting on his thread may get attention and a guide to what you desire.
Hope this is helpful. Looking forward to your next distro.
Just saw your appeal for kernel compilation assist. I wonder if @StemSee's utility would achieve what your needs are for your kernels. I also wonder if his work on having a BM which boots either the distro into 32bit (PAE) mode or 64bit mode as a user selectable item of a single distro.
If not, I understand. These are just some ideas where posting on his thread may get attention and a guide to what you desire.
Hope this is helpful. Looking forward to your next distro.
ASD replied to me via PM. I will not post the contents here directly, but the gist is that I forgot the subtraction for the figure in my rather lengthy post above, where I remarked about (2^64)-1 -- the result originally given was 2^64, without the -1 part!
That has been corrected, with an added EDIT remark at the bottom of the post explaining the change.
gc, I see you've posted here -- if you ever want to know, this is how you do it that's all I think needs be said on that topic. On another note -- I'd be quite interested in hearing how stemsee's bootmanager works... I could see that being incorporated, potentially, into a mainline Pup -- imagine, instead of having separate downloads for PAE and non-PAE, having a single Pup that lets you choose could that be done with his contraptionry?
That has been corrected, with an added EDIT remark at the bottom of the post explaining the change.
gc, I see you've posted here -- if you ever want to know, this is how you do it
that's all I think needs be said on that topic. On another note -- I'd be quite interested in hearing how stemsee's bootmanager works... I could see that being incorporated, potentially, into a mainline Pup -- imagine, instead of having separate downloads for PAE and non-PAE, having a single Pup that lets you choose could that be done with his contraptionry?