v2.6 vs v3.0
Linus,
***** Please BCC me in on any reply, not CC me.
Two reasons, I am not on the Mailing List,
and second I am suffering BIG time with SPAM
from posting to mailing lists/Newsgroups.
Instructions on real address at bottom.
Thanks in advance. *****
Quoted 14 lines
From: Linus Torvalds
Subject: Re: v2.6 vs v3.0
Date: Sat, 28 Sep 2002 18:31:45 -0700 (PDT)Quoted 8 lines · Ingo Molnar
i consider the VM and IO improvements one of the most important things that happened in the past 5 years - and it's definitely something that users will notice. Finally we have a top-notch VM and IO subsystem (in addition to the already world-class networking subsystem) giving significant improvements both on the desktop and the server - the jump from 2.4 to 2.5 is much larger than from eg. 2.0 to 2.4.
Quoted 4 lines
Hey, _if_ people actually are universally happy with the VM in the current 2.5.x tree, I'll happily call the dang thing 5.0 or whatever (just kidding, but yeah, that would be a good enough reason to bump the major number).
Just a comment, I suggest the version should stay in the 2.x domain. Being called a 2.6 makes sense as it follows the established version naming. If there is a trend to do a version 2.42.x or 2.62.x I suggest that the respective development versions would be 2.32.x and 2.52.x.
Quoted 7 lines
However, I'll believe that when I see it. Usually people don't complain during a development kernel, because they think they shouldn't, and then when it becomes stable (ie when the version number changes) they are surprised that the behabviour didn't magically improve, and _then_ we get tons of complaints about how bad the VM is under their load.
The reason is simple why this happens. People and organizations do not have time to do such testing, let alone create a parallel test system. This means real and serious effort needs to be taken in developing well focused test cases for the various elements of the Kernel, run each RC and final of Kernel through these battery of tests, then releae the Kernel RC for evaluation by the community to find the type of bugs that cannot be found in formal QA/Testing due to obvious nature of limited hardware combinations that can be tested. Point is, people can be really shy of testing something that is "development" based and not been formally tested. I know from first hand experience, I have gone back to the 2.2.x kernel simply because the 2.4.x kernel has had too many stability issues for my modest day to day workstation use as much as I wanted some of the improvements in the 2.4.x kernel.
Even when the Kernel is released as "production" level many sit on fence to see what others experience and flush out. This happens to greater degree as there is no formal Kernel testing. This has to change. It cannot be started overnight, but in steps it can and grow to base. I know there are projects and efforts to do formal Kernel testing, but it needs to be part of the process and fully supported by the community. Otherwise the Kernel (the heart and soul of the system) is just not taken as serious as fast as it should and deserves from all the fine effort and talent that gives the Kernel its life.
Quoted 9 lines
Am I hapyy with current 2.5.x? Sure. Are others? Apparently. But does that mean that we have a top-notch VM and we should bump the major number? I wish.
The block IO cleanups are important, and that was the major thing _I_ personally wanted from the 2.5.x tree when it was opened. I agree with you there. But I don't think they are major-number-material.
Agreeded.
Quoted 5 lines
Anyway, people who are having VM trouble with the current 2.5.x series, please _complain_, and tell what your workload is. Don't sit silent and make us think we're good to go.. And if Ingo is right, I'll do the 3.0.x thing.
Linus, I have not been able to get my system in state to do the testing of the VM subsystem due mostly to other issues. If there is someone in Toronto that can allow me access to one system that has both IDE and SCSI on it and at least 256MB of RAM I have lots of special testing I can do on the VM subsystem. CPU wise it would be good to have a uniprocessor as well as SMP system, where CPU is at least in the 400Mz or above range. I reatehr have a slower level CPU system that too fast. If both a fast and slow CPU system can be made available that is great. If only SMP system can be provided or meets other requirements, that is ok, I will just disable SMP and compile a kernels without SMP to do tests. Do not think this is all a strange and trivial request, as there are several combination even in the very very basic sense that should be done. If the system can have more RAM to validate the corner that seems to exist with larger memory configurations that would be great. It would be most helpful to have a tape backup to allow easy save and restore of test images without having to rebuild them each time. A capacity of at least 4 MB uncompressed per tape would be most helpful. A DVD writer would be an ok alternative to tape.
I am not fussy what is availavle as backup as long as it does job, so long as there is one and not hard drive based. Some of this testing could uncover some file system issues at stress levels based on past experience. I therefore need backups to be able to shorten time to restore and retest for any issues found and also to always start with "exact" same reference point in case there is an additive element of bugs to the testing that will distort the testing.
A native network connection of at least 10 mbits to allow FTP installs would be most helpful, as well as a CDWriter.
I have created various programs and test cases, but still need to refine them in terms of making them more automated. For not I can easily create a number of test cases, but has much duplicate effort on coding until I can distill the elements and then ayutomate many of the elements. I honestly believe what I have developed more effectively isolates the VM subsystem so that other kernel functions and demands do not cloud the ability to evaluate a VM subsystem.
The tests I would conduct would at a minimum test and compare the VM behaviour of Linux Kernel versions 2.2.22, 2.4.18, 2.4.19, 2.4.20, current 2.5.x, FreeBSD 4.6, 4.6.2, 4.5. All the Linux versions would be compiled done on the same base distribution and configuration. Side effects of VM testing would require more variables to be tested in combination to ensure no ill effects of system VM stress affect the other elements of the Kernel, i.e. to ensure Web Servers, Database servers, etc can be stable under peak stress conditions with respect to file systems and no accumulating loss of system performance as function of time and system stress.
Quoted 2 lines
Linus
Regards,
John L. Males
Software I.Q. Consulting
Toronto, Ontario
Canada
30 September 2002 14:20
==================================================================
According to Steve McConnell in:
After the Gold Rush: Creating a True Profession of Software Engineering
About 50% of the current software engineering body of knowledge is stable and will still be relevant 30 years from now.
***** Please BCC me in on any reply, not CC me.
Two reasons, I am not on the Mailing List,
and second I am suffering BIG time with SPAM
from posting to mailing lists/Newsgroups.
Instructions on real address at bottom.
Thanks in advance. *****
Please BCC me by replacing after the "@" as follows:
TLD = The last three letters of the word "internet"
Domain name = The first three letters of the word "theory",
followed by the first three letters of the word "offsite", followed by the first three letters of the country "Iceland".
My appologies in advance for the jumbled eMail address and request to BCC me, but SPAM has become a very serious problem. The eMail address in my header information is not a valid eMail address for me. I needed to use a valid domain due to ISP SMTP screen rules.
Please note: You may experience delays in my replies. I will
reply.
This is due to major restoration activity to my riser section of the building.
My internet access will be limited to a couple weekday evenings. Weekend access will be a bit better but limited as an indirect consequence of the restoration work.
If for any reason you need a more immediate reply or fail to receive a reply, please phone and leave a message. If you do leave a phone message, please note that I may not be able to hear the message and/or reply until the evening. This will be due to the extensive noise levels of the restoration work activity drowning out most other volume levels of sound.
The work starts 22 July 2002. Based on similar experience to other sections of the building I would estimate the major noise element of the work to be about 4+ weeks for my riser. There will be secondary instances once other ajacent riser sections are done that will have a similar impact, but for lessor periods of time.
My appologies in advance, but this is mandated work that must be done.
A message that is not in the archive (sent off-list or not yet received): [email protected]
John,
***** Please BCC me in on any reply, not CC me.
Two reasons, I am not on the Mailing List,
and second I am suffering BIG time with SPAM
from posting to mailing lists/Newsgroups.
Instructions on real address at bottom.
Thanks in advance. *****
I am serious. If you cannot respect that wish then so be it. Do you think anyone willing to make an offer of time and effort such as I have would make light of this? The fact you have made the effort to decompose the eMail address demonstrates the fact it can be done and not that much effort. It is not like I am posting with no valid eMail address or without indicating what my real eMail address is. What I have done is not that much different that other people have done when posting to the LKML. I have just made it difficult for eMail bots to harvest my eMail address but deleting/changing the .at., at, NOSPAM, extra levels in domain addresses, etc items that others use to effect the same result.
If it was not for the experiences I have already had with SPAM to the LKML I be more than happy to include my real eMail address in my eMail headers. It was in fact that I did that has me being hit by all sorts of SPAM. Please do not tell me all about the tools to delete SPAM. I know of some, and many I cannot use for various reasons, mostly to do with turning back eMail I need to receive.
I was not aware you spoke on behalf of Linus.
If you feel I am sending "junk mail" then I kindly suggest you report that to your ISP, my ISP and other reporting authorities. Let them determine if my posting is against AUP of internet and providers. Oh yes, of course people who send "junk mail" have and use a GPG key, that is on the keyserver as well!
Regards,
John L. Males
Software I.Q Consulting
Toronto, Ontario
Canada
30 September 2002 18:02
********** Reply Seperator **********
On (Mon) 2002-09-30 21:59:26 +0100
[email protected] wrote in Message-ID:
[email protected]
To: [email protected]
From: [email protected]
Subject: Re: v2.6 vs v3.0
Date: Mon, 30 Sep 2002 21:59:26 +0100 (BST)
Quoted 16 lines
Hi,
Quoted 11 lines
Please BCC me by replacing after the "@" as follows:
TLD = The last three letters of the word "internet"
Domain name = The first three letters of the word "theory",
followed by the first three letters of the word "offsite", followed by the first three letters of the country "Iceland".
My appologies in advance for the jumbled eMail address and request to BCC me, but SPAM has become a very serious problem. The eMail address in my header information is not a valid eMail address for me. I needed to use a valid domain due to ISP SMTP screen rules.Are you serious?
Quoted 8 lines
Please don't waste the time of 100+ developers, especially Linus who you cc'ed your message to.
If you are not prepared to include a real return-address, please don't post to the kernel dev list, otherwise *you are sending junk mail yourself*.
John.
==================================================================
According to Steve McConnell in:
After the Gold Rush: Creating a True Profession of Software Engineering
About 50% of the current software engineering body of knowledge is stable and will still be relevant 30 years from now.
***** Please BCC me in on any reply, not CC me.
Two reasons, I am not on the Mailing List,
and second I am suffering BIG time with SPAM
from posting to mailing lists/Newsgroups.
Instructions on real address at bottom.
Thanks in advance. *****
Please BCC me by replacing after the "@" as follows:
TLD = The last three letters of the word "internet"
Domain name = The first three letters of the word "theory",
followed by the first three letters of the word "offsite", followed by the first three letters of the country "Iceland".
My appologies in advance for the jumbled eMail address and request to BCC me, but SPAM has become a very serious problem. The eMail address in my header information is not a valid eMail address for me. I needed to use a valid domain due to ISP SMTP screen rules.
Please note: You may experience delays in my replies. I will
reply.
This is due to major restoration activity to my riser section of the building.
My internet access will be limited to a couple weekday evenings. Weekend access will be a bit better but limited as an indirect consequence of the restoration work.
If for any reason you need a more immediate reply or fail to receive a reply, please phone and leave a message. If you do leave a phone message, please note that I may not be able to hear the message and/or reply until the evening. This will be due to the extensive noise levels of the restoration work activity drowning out most other volume levels of sound.
The work starts 22 July 2002. Based on similar experience to other sections of the building I would estimate the major noise element of the work to be about 4+ weeks for my riser. There will be secondary instances once other ajacent riser sections are done that will have a similar impact, but for lessor periods of time.
My appologies in advance, but this is mandated work that must be done.
Whoops! Looks like you just blew your cover... ;)
Nick
*BCC'ed to [email protected]*
Quoted 11 lines · John L. Males
[snip] ********** Reply Seperator **********
On (Mon) 2002-09-30 21:59:26 +0100 [email protected] wrote in Message-ID:
[email protected]To: [email protected]
From: [email protected]
Subject: Re: v2.6 vs v3.0
Date: Mon, 30 Sep 2002 21:59:26 +0100 (BST)
[snip]
Quoted 9 lines · John L. Males
John,
***** Please BCC me in on any reply, not CC me. Two reasons, I am not on the Mailing List, and second I am suffering BIG time with SPAM from posting to mailing lists/Newsgroups. Instructions on real address at bottom. Thanks in advance. *****
Then don't post to public lists.