Posted by
timothy
on from the jamie-has-the-munchies dept.
punkmac writes "It's that time again to do the thing we all love to do, compile your shiny new kernel. This time its 2.4.20. The changelog can be found at the usual places, and please use a mirror for all your downloads"
There are several significant changes and improvements, not the least is multithreading performance for us SMP users, not to mention the improved filesystem(s) performance/support. Admittedly, not many flashy, shiny new things, but a fair number of performance/real world improvements.
-- Don't park drunk, accidents cause people.
Re:Be kind to kernel.org
by
Clue4All
·
· Score: 2, Interesting
Why? ISC increased their bandwith from 100 Mbps to 250 Mbps. It rarely reached 90% usage, often during a new Red Hat release. The main site works just fine, and will continue to do so.
Did they fix the new ptrace vulnerability?
by
bigberk
·
· Score: 3, Interesting
I haven't tried it myself yet, but I found no reference to this ptrace vulnerability in the changelog. I suspect this is still a problem (it was in 2.4.19).
zlib in the kernel?
by
hpavc
·
· Score: 2, Interesting
anyone know where i can read about this zlib(compression|decompression) library support? there is some very interesting mentioning of it in./inculde/linux/zlib.h and of course in menuconfig as well.
-- members are seeing something, your seeing an ad
Re:zlib in the kernel?
by
Anonymous Coward
·
· Score: 1, Interesting
Why would you need to read about that ? Its just a library, for things
in the kernel needing zlib (de)compression, like ppp. At this point I actually dont know anything else than ppp which uses it.
It's been there all time, now its just placed in another directory.
Re:zlib in the kernel?
by
TeknoHog
·
· Score: 3, Interesting
There is also zisofs -- read-only support of compressed ISO CDs. In read-only systems the compression actually makes sense, because the size does not change, and decompression is much faster than compression. To use these, you need additional tools, probably called mkzisofs.
Unfortunately, many CDs are used to store mp3/ogg/divx and the like, which don't really compress much. So zisofs is not necessarily very useful.
-- Escher was the first MC and Giger invented the HR department.
For anyone using reiserfs
by
Xpilot
·
· Score: 5, Interesting
There is a significant patch that speeds up write operations for large files on IDE drives by up to 15%. This made it in v2.4.20-pre6.
-- "Backups are for wimps. Real men upload their data to an FTP site and have everyone else mirror it." -- Linus Torvalds
What about Athlon prefetch bug?
by
Anonymous Coward
·
· Score: 1, Interesting
Revert 2.4.19's AMD Athlon prefetch workaround
What does this mean?
These changelogs are getting too long...
by
7-Vodka
·
· Score: 5, Interesting
They need to be put in a database so you can click on a hardware device and get a list of patches applied with other query parameters for example.
I don't know but my Wireless USB NIC Drivers don't seem to be working.:o/ Already posted it to the mailing list...
-- FLR
Broken USB mouse support fixed?
by
Anonymous Coward
·
· Score: 2, Interesting
Anybody know if this version fixes USB optical mouse support which seems to have been broken for some users with 2.4.19? (Yes, I've checked groups.google.com and know about enabling CONFIG_USB_HIDINPUT but this does NOT help). It seems Red Hat 7.3 users were affected, as well as Mandrake (although I know only what broke for me -- RH 7.3). Thank god my optical usb mouse had a usb->psmouse converter plug which I'm using as a temporary workaround.
Did they fix these showstopper bugs in 2.4.20?
by
linefeed0
·
· Score: 2, Interesting
There is at least one nasty VM-related bug which has been reported in Linux 2.4.19 and possibly earlier revisions. Rumors of more abound; this is the one that's personally happened to me.
Kernel BUG at page_alloc.c:91! (kills kswapd) Google groups for "page_alloc.c:91". Usually blamed on NVidia drivers but notable cases have occurred without them. It's not clear whether the kernel team has made a notable effort to debug them, but they are real and occur on a variety of systems. When this happens kswapd stops working, which causes more bugs later on, and eventually a system reboot is required.
There are other period reports of BUGs being flagged with an otherwise normal kernel on the linux-kernel list.
How to get patches accepted?
by
caluml
·
· Score: 3, Interesting
[PATCH] : sock_writable not appropriate for TCP sockets
hi marcelo-
[ i sent this patch August 30 against 20-pre5, and it appears to have been
dropped. this is an important performance fix that should be included in
2.4.20. i apologize for not tracking this more closely. ]
sock_writeable determines whether there is space in a socket's output
buffer. socket write_space callbacks use it to determine whether to wake
up those that are waiting for more output buffer space.
however, sock_writeable is not appropriate for TCP sockets. because the
RPC client's write_space callback uses it for TCP sockets, the RPC layer
hammers on sock_sendmsg with dozens of write requests that are only a few
hundred bytes long when it is trying to send a large write RPC request.
this patch adds logic to the RPC layer's write_space callback that
properly handles TCP sockets.
patch reviewed by Trond, Alexey, and DaveM, and already accepted in 2.5.
I wonder if I can get the name changed from Linux to Evilix by sending a patch to Marcelo and just mentioning that Trond, Alexey and DaveM have reviewed it and it's already in 2.5...;)
DMCA Violations
by
12dec0de
·
· Score: 2, Interesting
As some people here on slashdot (an article IIRC) hinted at patches in 2.4.20 being in possible violation with the DMCA, and hence waiting to release.20, I guess these are all either cleaned up by now or not in the patch at all.
Can somebody give me a pointer to information on the actual problems and the resolution as well?
Re:Be kind to kernel.org
by
KjetilK
·
· Score: 3, Interesting
Hm, I read that they've got a 250 Mbits/s link now, and I was going to respond "hey, let's stress-test it!" But www.kernel.org was very slow to me, in spite of that
Current bandwidth utilization 57.87 Mbit/s
That's actually rather weird, because it is not that uncommon to have that load. If this number is correct, there has to be something wrong with the server. But probably, they are getting more traffic.
Well me? I went directly to a mirror after getting the announcement from the announcement list this morning and downloaded it from there before reading/.
The new kernel fixes the lcall DoS. While I have no untrusted users on my system, things like that is always good to have out of the way.
-- Employee of Inrupt, Project Release Manager and Community Manager for Solid
Having read both the changelog and searched the kernel archives it's interesting to note that way back in 2.4.20-pre3 the Athlon MP/XP prefetch bug work around was pulled out of the kernel.
Subsiquently, Alan Cox in the kernel mailing list archives responds to someone asking about this and says that he thinks some of the VM changes have fixed at least the AGP problems but not the mmap ones, to quote:-
-ac has it removed. I didn't know Marcelo had it removed. Andi Kleen has
a patch for doing most of the right things without trashing performance.
That may be what Marcelo merged. It fixed AGP but not O_SYNC mmap I
believe
this was posted on the 17th August and there's no further mention of the problem as far as I can see.
If this hasn't been fixed in the full 2.4.20 release then there is no way I can install this kernel on our three dual processor Athlon MP boxes we use for mathematical simulations as the machines need to be able to run for many weeks on the problems and hence have to be rock-solid stable.
Does anyone know if the new VM fixes fully fix the Athlon MP/XP problems?
I've been programming for a long time. I'm good at it, and I suppose if I really put some time into coding for a Free Software project, I could do a lot of good.
But most of my contribution to Free Software has really been in the form of writing. I have also written a lot of stuff which is not copylefted, but posted publicly on the web.
here some more as well as
this.
I have contributed some to
zoolib, but that's mostly in the form of qa, project management (for the initial open source release) and marketing.
One reason I prefer to contribute by writing is that my normal programming work is so hard, that when I get time to take a break from it, it's difficult to work up much enthusiasm to write more code, no matter how fun the project might be. I imagine that's a common problem.
Another reason is that I feel that any contribution I could make to Free Software, at least in the limited time I have, would be small. I could fix some bugs, add some features, do some testing. But how many people would benefit from my personal contribution? I don't think that many would, at least not until I had the time to develop a really serious package, and I just don't have the energy for that. I have lots of ideas, but no time.
But I feel that passing on my experience by writing can benefit others far out of proportion to the effort I put in. That is because I aim my writing to enable others to do better. By writing well, I enable many other developers to code a little better, and many users to do better testing and bug reporting.
I could lead by example by writing good code, but how many people would learn by reading it? When's the last time you studied the source for some package you weren't really actively involved with? Prose is much more accessible.
This is all the more important because so few engineers of any sort are good writers. When my father was a civilian electrical engineer for the Navy, the shipyard sent him to a writing class so he could write better test plans! The man has a master's degree! But the Navy put lots of people through that class because so many of their engineers didn't know how to write.
What is funny is that I find writing much more difficult than programming. With software, you know when you're wrong. It's not always so clear with writing. The main reason I write so well is because most of my effort is put into editing - and I still find lots of mistakes later.
Even more ironic is that I used to hate writing with a passion. One major reason is that I have terrible handwriting - I can't read my own handwriting, and can't imagine how anyone else could. But the schools always used to require handwritten essays. They used to send me to remedial handwriting classes, which I really hated because they made my hand hurt. It's painful for me to write much by hand.
What did it for me was two things - a composition class I took at the community college during the summer when I was sixteen, that was just really well taught, and being able to type. I type really fast now, and there's no pain.
My senior year Advanced Placement English Teacher asked me to drop the class because of my poor handwriting. He was quite taken aback when I started screaming at him. I'd had enough of teachers criticizing my handwriting, I didn't need to hear it again when I was seventeen years old.
He was concerned that I couldn't pass the exam (which could get me college credit) because the judges wouldn't be able to read my essays.
He proposed a compromise. He suggested that I block print.
I had no problem with that. And at the exam at the end of the year, I turned in my exam neatly block-printed in all capital letters. I just used bigger capitals for where a capital was really required.
I was the only student in my school that year to get a 5 on the english AP exam (a perfect score).
mirrors and P2P
by
silence535
·
· Score: 3, Interesting
..can be found at the usual places, and please use a mirror for all your downloads"
Apologies if this has been discussed before and I missed it. (If so, ignore and silently send me the link to the discussion please.)
Why don't the kernel guys distribute the kernel sources over the well established P2P networks to reduce load on their servers and to speed up downloads for everyone?
I mean, why not put it into the donkey and publish the e2k link on the official kernel.org web page and also submit it to the sharereactor?
MD5 or other hashing can guarantee that the official release has not been tampered with.
Am I missing something here?
silence
-- Dyslectics of the world, untie!
Re:use the mirrors to download the changelog
by
suwain_2
·
· Score: 3, Interesting
What I've always wondered is this -- don't they notice that their program doesn't compile, and fix it before releasing it? It's easy to forget a semicolon on the end of a line, but I can't say I've ever released a program that had fatal compile errors. Is there something behind the scenes I'm missing?
-- ________________________________________________
suwain_2:: quality slashdot p
VESA Local Bus
by
Zymurgy
·
· Score: 3, Interesting
Reading the changelog, I was just happy to see they updated the drivers for the Adaptec AHA-2840 VLB SCSI controller and the SMC 91C92 Ethernet chip, which is found on my VESA Local Bus Ethernet cards. Ah, I'm thrilled because I just so happen to be building a file server right now that uses both.
Hey, you can't tell me I'm the only person in here who uses the VESA Local Bus on his servers!!!
Slashdot Mirror
There are several significant changes and improvements, not the least is multithreading performance for us SMP users, not to mention the improved filesystem(s) performance/support. Admittedly, not many flashy, shiny new things, but a fair number of performance/real world improvements.
Don't park drunk, accidents cause people.
Why? ISC increased their bandwith from 100 Mbps to 250 Mbps. It rarely reached 90% usage, often during a new Red Hat release. The main site works just fine, and will continue to do so.
Is your browser retarded?
I haven't tried it myself yet, but I found no reference to this ptrace vulnerability in the changelog. I suspect this is still a problem (it was in 2.4.19).
anyone know where i can read about this zlib(compression|decompression) library support? there is some very interesting mentioning of it in ./inculde/linux/zlib.h and of course in menuconfig as well.
members are seeing something, your seeing an ad
There is a significant patch that speeds up write operations for large files on IDE drives by up to 15%. This made it in v2.4.20-pre6.
"Backups are for wimps. Real men upload their data to an FTP site and have everyone else mirror it." -- Linus Torvalds
Revert 2.4.19's AMD Athlon prefetch workaround
What does this mean?
just my $2*10^-2.00
Liberty.
I don't know but my Wireless USB NIC Drivers don't seem to be working. :o/ Already posted it to the mailing list...
FLR
Anybody know if this version fixes USB optical mouse support which seems to have been broken for some users with 2.4.19? (Yes, I've checked groups.google.com and know about enabling CONFIG_USB_HIDINPUT but this does NOT help). It seems Red Hat 7.3 users were affected, as well as Mandrake (although I know only what broke for me -- RH 7.3).
Thank god my optical usb mouse had a usb->psmouse converter plug which I'm using as a temporary workaround.
There is at least one nasty VM-related bug which has been reported in Linux 2.4.19 and possibly earlier revisions. Rumors of more abound; this is the one that's personally happened to me.
Kernel BUG at page_alloc.c:91! (kills kswapd)
Google groups for "page_alloc.c:91". Usually blamed on NVidia drivers but notable cases have occurred without them. It's not clear whether the kernel team has made a notable effort to debug them, but they are real and occur on a variety of systems. When this happens kswapd stops working, which causes more bugs later on, and eventually a system reboot is required.
There are other period reports of BUGs being flagged with an otherwise normal kernel on the linux-kernel list.
[PATCH] : sock_writable not appropriate for TCP sockets
hi marcelo-
[ i sent this patch August 30 against 20-pre5, and it appears to have been dropped. this is an important performance fix that should be included in 2.4.20. i apologize for not tracking this more closely. ]
sock_writeable determines whether there is space in a socket's output buffer. socket write_space callbacks use it to determine whether to wake up those that are waiting for more output buffer space.
however, sock_writeable is not appropriate for TCP sockets. because the RPC client's write_space callback uses it for TCP sockets, the RPC layer hammers on sock_sendmsg with dozens of write requests that are only a few hundred bytes long when it is trying to send a large write RPC request. this patch adds logic to the RPC layer's write_space callback that properly handles TCP sockets.
patch reviewed by Trond, Alexey, and DaveM, and already accepted in 2.5.
I wonder if I can get the name changed from Linux to Evilix by sending a patch to Marcelo and just mentioning that Trond, Alexey and DaveM have reviewed it and it's already in 2.5...
Get your own free personal location tracker
As some people here on slashdot (an article IIRC) hinted at patches in 2.4.20 being in possible violation with the DMCA, and hence waiting to release .20, I guess these are all either cleaned up by now or not in the patch at all.
Can somebody give me a pointer to information on the actual problems and the resolution as well?
That's actually rather weird, because it is not that uncommon to have that load. If this number is correct, there has to be something wrong with the server. But probably, they are getting more traffic.
Well me? I went directly to a mirror after getting the announcement from the announcement list this morning and downloaded it from there before reading /.
The new kernel fixes the lcall DoS. While I have no untrusted users on my system, things like that is always good to have out of the way.
Employee of Inrupt, Project Release Manager and Community Manager for Solid
Subsiquently, Alan Cox in the kernel mailing list archives responds to someone asking about this and says that he thinks some of the VM changes have fixed at least the AGP problems but not the mmap ones, to quote:-
-ac has it removed. I didn't know Marcelo had it removed. Andi Kleen has a patch for doing most of the right things without trashing performance. That may be what Marcelo merged. It fixed AGP but not O_SYNC mmap I believe
this was posted on the 17th August and there's no further mention of the problem as far as I can see.
If this hasn't been fixed in the full 2.4.20 release then there is no way I can install this kernel on our three dual processor Athlon MP boxes we use for mathematical simulations as the machines need to be able to run for many weeks on the problems and hence have to be rock-solid stable.
Does anyone know if the new VM fixes fully fix the Athlon MP/XP problems?
Agrajag: "Oh no, not again!"
But most of my contribution to Free Software has really been in the form of writing. I have also written a lot of stuff which is not copylefted, but posted publicly on the web. here some more as well as this.
I have contributed some to zoolib, but that's mostly in the form of qa, project management (for the initial open source release) and marketing.
One reason I prefer to contribute by writing is that my normal programming work is so hard, that when I get time to take a break from it, it's difficult to work up much enthusiasm to write more code, no matter how fun the project might be. I imagine that's a common problem.
Another reason is that I feel that any contribution I could make to Free Software, at least in the limited time I have, would be small. I could fix some bugs, add some features, do some testing. But how many people would benefit from my personal contribution? I don't think that many would, at least not until I had the time to develop a really serious package, and I just don't have the energy for that. I have lots of ideas, but no time.
But I feel that passing on my experience by writing can benefit others far out of proportion to the effort I put in. That is because I aim my writing to enable others to do better. By writing well, I enable many other developers to code a little better, and many users to do better testing and bug reporting.
I could lead by example by writing good code, but how many people would learn by reading it? When's the last time you studied the source for some package you weren't really actively involved with? Prose is much more accessible.
This is all the more important because so few engineers of any sort are good writers. When my father was a civilian electrical engineer for the Navy, the shipyard sent him to a writing class so he could write better test plans! The man has a master's degree! But the Navy put lots of people through that class because so many of their engineers didn't know how to write.
What is funny is that I find writing much more difficult than programming. With software, you know when you're wrong. It's not always so clear with writing. The main reason I write so well is because most of my effort is put into editing - and I still find lots of mistakes later.
Even more ironic is that I used to hate writing with a passion. One major reason is that I have terrible handwriting - I can't read my own handwriting, and can't imagine how anyone else could. But the schools always used to require handwritten essays. They used to send me to remedial handwriting classes, which I really hated because they made my hand hurt. It's painful for me to write much by hand.
What did it for me was two things - a composition class I took at the community college during the summer when I was sixteen, that was just really well taught, and being able to type. I type really fast now, and there's no pain.
My senior year Advanced Placement English Teacher asked me to drop the class because of my poor handwriting. He was quite taken aback when I started screaming at him. I'd had enough of teachers criticizing my handwriting, I didn't need to hear it again when I was seventeen years old.
He was concerned that I couldn't pass the exam (which could get me college credit) because the judges wouldn't be able to read my essays.
He proposed a compromise. He suggested that I block print.
I had no problem with that. And at the exam at the end of the year, I turned in my exam neatly block-printed in all capital letters. I just used bigger capitals for where a capital was really required.
I was the only student in my school that year to get a 5 on the english AP exam (a perfect score).
Request your free CD of my piano music.
..can be found at the usual places, and please use a mirror for all your downloads"
Apologies if this has been discussed before and I missed it. (If so, ignore and silently send me the link to the discussion please.)
Why don't the kernel guys distribute the kernel sources over the well established P2P networks to reduce load on their servers and to speed up downloads for everyone?
I mean, why not put it into the donkey and publish the e2k link on the official kernel.org web page and also submit it to the sharereactor?
MD5 or other hashing can guarantee that the official release has not been tampered with.
Am I missing something here?
silence
Dyslectics of the world, untie!
What I've always wondered is this -- don't they notice that their program doesn't compile, and fix it before releasing it? It's easy to forget a semicolon on the end of a line, but I can't say I've ever released a program that had fatal compile errors. Is there something behind the scenes I'm missing?
________________________________________________
suwain_2
Reading the changelog, I was just happy to see they updated the drivers for the Adaptec AHA-2840 VLB SCSI controller and the SMC 91C92 Ethernet chip, which is found on my VESA Local Bus Ethernet cards. Ah, I'm thrilled because I just so happen to be building a file server right now that uses both.
Hey, you can't tell me I'm the only person in here who uses the VESA Local Bus on his servers!!!
-Lawrence
Visit Zymurgy Records!
I'll try to take you up on that.
Don't give me none of this "nature theme" business.