FP in kernel is the only way linux can be deployed in these scenarios.
the -rt kernel is not sufficient.
Xenomai is a lot better, however it does add for me up to 25 microseconds of latency on intel 2 ghz xeon.
So this means that in order to take dsp processing out of the kernel with the required audio latency we lose 15% of the cpu's processing power which is potentially 1.2 gigaflops with sse3!
so what is 'more right'? wasting 15% of your cpu due to scheduling latency even with xenomai or enabling fpu/sse in kernel?
People who believe that floating point has no place in the kernel are living in the dark ages.
The reality now is that thousands of products are shipped and in use now that process audio in real time on linux, in kernel mode, with major DSP style processing done with FPU and Altivec and SSE2 with userspace only involved to tweak coefficients.
If you ever saw a Cirque Du Soleil show, or went to disneyland, saw a metallica concert, or heard the opening ceremonies in Bejing, you have listened to audio being processed in linux kernel mode in floating point with both single fpu and vector procesing.
It would be nice to do it all in userspace.
Until the linux kernel is able to schedule real time user tasks with less than 10 microseconds of latency for a recurring interrupt every 166 microseconds, solution providers are forced to use kernel mode audio processing.
It is so unfortunate that linux audio is in such a mess with also, pulseaudio, oss. Developers should take a look at apple's CoreAudio for ideas if they can't figure things out.
level control systems, now owned by Meyer Sound, still has installations from 1996 that are running with either original BeBoxes (repackaged into rack mount cases) or older intel hardware.
People didn't upgrade because there was no need to; it was not broken and the hardware still works.
One live musical show was using one of these rack mounted original BeBox with Dual Processors. A few years ago it stopped working; they rebooted it and it worked but only one CPU was running. The problem was that they one cpu fan stopped turning, the cpu got very hot - so hot that it unsoldered itself and fell. But the system worked just fine on a reboot, just a little slower!
Yes, in fact for some time now, over 20 years ago, jet fighters utilized Head Transfer Functions (HTF's) to provide three dimensional audio to the pilot. So wherever the pilot is, and wherever his orientation, he would "hear" where the enemy was located, but the distances were scaled to make them sound closer.
On windows XP the command line "ipv6 install" is all that is necessary to enable ipv6.
Perhaps you don't know that you have have ipv4 and ipv6 enabled at the same time? People can migrate to it as they need.
The benefits are real, and are all brought about because of the ipv6 Router Advertisement protocol. Having internet addressable IP addresses is NOT the point, it is only one tiny option that you have if you are using ipv6.
Take a look at UPNP and all their NAT issues with regards to "punching holes in NAT routers" to allow for bidirection DNS networking.
Take a look at all users who have 192.168.0.x home networks trying to VPN in to their office which also has 192.168.0.x.
Take a look at the hacks that allow websites to trigger links to your internal router http admin page via "http://192.168.0.1/cgi-bin/admin.cgi?user=admin&pw=password"
IPv6 also has benefits for QoS control and multicasting.
The number of bits in an IP address is irrelevant to people wanting ipv6.
I'm sorry though that you have to manage hundreds of windows XP systems. Hopefully when they upgrade to Win7 or Mac or Linux this would free up your time to do more interesting work.
I use IPV6 over the internet at home right now all managed by a little "Apple Airport Express" that has ipv6 router capability in it already. Every computer on my wireless LAN gets a link local IPV6 address as well as a global IPV6 address, be it Windows, Linux, or Mac.
Here is the interesting thing; This $100.00 "Apple Airport Express" which already supports ipv6 has a little checkbox in the ipv6 settings labelled "Allow incoming IPv6 connections". With this checkbox off, it is a firewall, it does not allow incoming IPV6 Connections!
Isn't that neat! So the external hacker is unable to nmap every single box on your private network!
Anyways, I never need to remember ipv6 addresses either; all my computers are smart enough to discover each other themselves via avahi, bonjour, network browsing, etc.
It seems like most people here are scared of ipv6, yet never bothered to actually learn anything about it.
You mean kind of like TrollTech's (Nokia's) QT4.5 library which is licensed under the LGPL? Is that an inferior library? Oh wait it is also available under a dual licensed proprietary license! So does that make it inferior to itself?
> The spirit of the GPL does not matter. All that matters is the reality of the license.
Yes, but to take such a literal view ignores the fact that the GPL was written for a reason - and that reason is that proprietary vendors were restricting certain freedoms that the FSF believes belong to everybody.
The GPL was written to preserve those freedoms for anyone using and distributing software under it. If a vendor finds a loophole that again restricts those freedoms, they are clearly not interested in promoting the freedoms the GPL was intended to preserve, and thus the GPL is broken.
Wow, that's not true at all. The GPL is a license that has requirements. If the requirements are met the GPL is not broken. The GPLv2 and GPLv3 have different requirements and the GPLv4 when it comes out will have other requirements.
If you think that the application store violates the spirit of the GPL, then surely you must think that the requirement of 95% of all Intel motherboards to contain a closed source BIOS in order to boot a violation as well.
If there is a spirit of the GPL it is clearly the spirit of the great RMS.
Tell me, what does He think about this topic? I'm not interested in an A.C.'s interpretation of what the GPL's spirit might be.
Yes exactly, My Linux box is completely free and costs me nothing to run! Even the hard disks are free, because it is not apple!!! And since I told my electric company that I'm running GPL software, my electric company is giving me free electricity for running my linux system!
The app store restrictions do not violate the GPL in any way. The author of the application has to make the source code available. If the author does not, then he is violating the GPL.
When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs; and that you know you can do these things.
Even though apple sends goons who force you to sign papers , strip search you, and take your lunch money in order to download binaries to a real iphone, that has no bearing on the freedom of the source code. The GPLv2 or GPLv3 explicitly allows anyone to sell GPL'd binaries as long as they make the source code available to anyone at no more than your cost of duplication and shipping.
Imagine I personally create a handheld prototype hardware device and port Xpilot to it. There are only two of these made and cost me $2000 each to manufacture these prototypes. I sell one of the device with Xpilot in it. I include a CD with the Xpilot source code with the device. If anyone else wants to compile Xpilot on my prototype device, they would have to buy one from me. They would have to pay me more than $2000 to make it worth my while to bother. And if I say no, I'm not building them anymore because I want to backpack around the world now, am I violating the 'Spirit' of the GPL because I distributed the binaries and sources for a platform which is no longer available? And how would this scenario be different if apple decided to stop selling iphone SDKs? Are you requiring apple's SDK to be open source? or requiring apple's hardware to be open source? why don't you require your intel motherboard to have an open source BIOS as well, seeing that you can't play Xpilot on a PC without paying some BIOS company like Pheonix for the right to boot your computer.
If the XPilot authors did not want anyone selling their software they should have used a different license.
Perhaps the GPLv4 will have an explicit anti-apple clause.
I believe that precision millisecond stock trading globally is the real reason behind the IEEE 1588v2 precision time protocol. The cisco 9000 enterprise switch supports it. Support has been lacking in smaller switches. The only other group using PTPv2 is the cell phone industry.
The interesting part of PTPv2 for me is that it is used in the 802.1AS protocol ( http://www.ieee802.org/1/pages/802.1as.html ) which is one of the foundations of Audio Video Bridging (AVB) http://www.ieee802.org/1/pages/avbridges.html - Which allows for real time low latency low jitter media streams transported via ethernet with guaranteed bandwidth.
Just yesterday I was joking with friends: Forget about stealing the rounded pennies from bank accounts, criminals could re-program the PTPv2 implementation in switches to steal milliseconds of time during trading!
Anyways, back on the original question, no, network speed is not so crucial once all of your packets are properly timestamped.
Interestingly, amazon web services now has a import/export service, where they will accept your USB drive via courier and import it into their "Simple Storage Service" aka S3.
The thing is that there was a 75 megabyte file on the computer which contained the IP addresses, hostnames, user names, and email passwords and ftp passwords of users and computers in the same room. I watched the file grow in size as it collected passwords!
This is another reason why self-signed ssl certificates can be more secure than having no certificate at all, despite the warnings firefox gives you.
--jeffk++
Re:Amusingly..
on
R.I.P. FTP
·
· Score: 2, Insightful
I have a 'friend'... yeah, that's it... Who back in 1997 or so had a co-located server running RedHat linux with a vulnerable dns server. It was attacked. Multiple times. I only noticed - I mean my 'friend' only noticed when a further attack caused/etc/passwd to be broken. When the box was taken down and analyzed, it was noted that the system put the ethernet port in promiscuous mode and was sniffing all of the traffic for the co-located boxes that were connected to the same switch. A file on the disk contained all the user names and protocols and passwords of the people with accounts on those co-located systems.
So when you log in with unencrypted FTP, how do you know that the server that you are connecting to is not sitting in a rack next to a compromised system sniffing the traffic? How often do people misconfigure their servers and switches like this, and how would you know?
Slashdot Mirror
The option really is to not use linux at all, which would be more un-sane!
--jeffk++
Define 'right' ?
FP in kernel is the only way linux can be deployed in these scenarios.
the -rt kernel is not sufficient.
Xenomai is a lot better, however it does add for me up to 25 microseconds of latency on intel 2 ghz xeon.
So this means that in order to take dsp processing out of the kernel with the required audio latency we lose 15% of the cpu's processing power which is potentially 1.2 gigaflops with sse3!
so what is 'more right'? wasting 15% of your cpu due to scheduling latency even with xenomai or enabling fpu/sse in kernel?
--jeffk++
People who believe that floating point has no place in the kernel are living in the dark ages.
The reality now is that thousands of products are shipped and in use now that process audio in real time on linux, in kernel mode, with major DSP style processing done with FPU and Altivec and SSE2 with userspace only involved to tweak coefficients.
If you ever saw a Cirque Du Soleil show, or went to disneyland, saw a metallica concert, or heard the opening ceremonies in Bejing, you have listened to audio being processed in linux kernel mode in floating point with both single fpu and vector procesing.
It would be nice to do it all in userspace.
Until the linux kernel is able to schedule real time user tasks with less than 10 microseconds of latency for a recurring interrupt every 166 microseconds, solution providers are forced to use kernel mode audio processing.
It is so unfortunate that linux audio is in such a mess with also, pulseaudio, oss. Developers should take a look at apple's CoreAudio for ideas if they can't figure things out.
--jeffk++
Now you are just crazy rumor mongering...
--jeffk++
First Flu Shot?
I wonder what happens if a worker has an existing health immune system based condition that can be adversely affected by a flu shot?
--jeffk++
Thanks for the link.
I wonder how these charts would represent the Island of Stability ?
--jeffk++
Yes this blog posting of his was disappointing because he showed his opinions in areas of technology where he has willful ignorance.
--jeffk++
You leave your computers on and logged in without any screen saver password?
level control systems, now owned by Meyer Sound, still has installations from 1996 that are running with either original BeBoxes (repackaged into rack mount cases) or older intel hardware.
http://testou.free.fr/www.beatjapan.org/mirror/www.be.com/developers/developer_news/spotlight/lcs.html
People didn't upgrade because there was no need to; it was not broken and the hardware still works.
One live musical show was using one of these rack mounted original BeBox with Dual Processors. A few years ago it stopped working; they rebooted it and it worked but only one CPU was running. The problem was that they one cpu fan stopped turning, the cpu got very hot - so hot that it unsoldered itself and fell. But the system worked just fine on a reboot, just a little slower!
--jeffk++
But the reality is that no one will care to read your tweets unless you are a politician or a star.
--jeffk++
Yes, in fact for some time now, over 20 years ago, jet fighters utilized Head Transfer Functions (HTF's) to provide three dimensional audio to the pilot. So wherever the pilot is, and wherever his orientation, he would "hear" where the enemy was located, but the distances were scaled to make them sound closer.
--jeffk++
On windows XP the command line "ipv6 install" is all that is necessary to enable ipv6.
Perhaps you don't know that you have have ipv4 and ipv6 enabled at the same time? People can migrate to it as they need.
The benefits are real, and are all brought about because of the ipv6 Router Advertisement protocol. Having internet addressable IP addresses is NOT the point, it is only one tiny option that you have if you are using ipv6.
Take a look at UPNP and all their NAT issues with regards to "punching holes in NAT routers" to allow for bidirection DNS networking.
Take a look at all users who have 192.168.0.x home networks trying to VPN in to their office which also has 192.168.0.x.
Take a look at the hacks that allow websites to trigger links to your internal router http admin page via "http://192.168.0.1/cgi-bin/admin.cgi?user=admin&pw=password"
IPv6 also has benefits for QoS control and multicasting.
The number of bits in an IP address is irrelevant to people wanting ipv6.
I'm sorry though that you have to manage hundreds of windows XP systems. Hopefully when they upgrade to Win7 or Mac or Linux this would free up your time to do more interesting work.
--jeffk++
I use IPV6 over the internet at home right now all managed by a little "Apple Airport Express" that has ipv6 router capability in it already. Every computer on my wireless LAN gets a link local IPV6 address as well as a global IPV6 address, be it Windows, Linux, or Mac.
Here is the interesting thing; This $100.00 "Apple Airport Express" which already supports ipv6 has a little checkbox in the ipv6 settings labelled "Allow incoming IPv6 connections". With this checkbox off, it is a firewall, it does not allow incoming IPV6 Connections!
Isn't that neat! So the external hacker is unable to nmap every single box on your private network!
Anyways, I never need to remember ipv6 addresses either; all my computers are smart enough to discover each other themselves via avahi, bonjour, network browsing, etc.
It seems like most people here are scared of ipv6, yet never bothered to actually learn anything about it.
--jeffk++
But the data IS secure! there is a little padlock on my web browser window that says so! ;-)
--jeffk++
yay! nom nom nom
jeffy
How do you prove to law enforcement that your computer does NOT have encrypted files on it?
jeff
You mean kind of like TrollTech's (Nokia's) QT4.5 library which is licensed under the LGPL? Is that an inferior library? Oh wait it is also available under a dual licensed proprietary license! So does that make it inferior to itself?
--jeffk++
> The spirit of the GPL does not matter. All that matters is the reality of the license.
Yes, but to take such a literal view ignores the fact that the GPL was written for a reason - and that reason is that proprietary vendors were restricting certain freedoms that the FSF believes belong to everybody.
The GPL was written to preserve those freedoms for anyone using and distributing software under it. If a vendor finds a loophole that again restricts those freedoms, they are clearly not interested in promoting the freedoms the GPL was intended to preserve, and thus the GPL is broken.
Wow, that's not true at all. The GPL is a license that has requirements. If the requirements are met the GPL is not broken. The GPLv2 and GPLv3 have different requirements and the GPLv4 when it comes out will have other requirements.
If you think that the application store violates the spirit of the GPL, then surely you must think that the requirement of 95% of all Intel motherboards to contain a closed source BIOS in order to boot a violation as well.
If there is a spirit of the GPL it is clearly the spirit of the great RMS.
Tell me, what does He think about this topic? I'm not interested in an A.C.'s interpretation of what the GPL's spirit might be.
--jeffk++
Yes exactly, My Linux box is completely free and costs me nothing to run! Even the hard disks are free, because it is not apple!!! And since I told my electric company that I'm running GPL software, my electric company is giving me free electricity for running my linux system!
--jeffk++
The app store restrictions do not violate the GPL in any way. The author of the application has to make the source code available. If the author does not, then he is violating the GPL.
The spirit of the GPL does not matter. All that matters is the reality of the license.
From the GPLv2:
Even though apple sends goons who force you to sign papers , strip search you, and take your lunch money in order to download binaries to a real iphone, that has no bearing on the freedom of the source code. The GPLv2 or GPLv3 explicitly allows anyone to sell GPL'd binaries as long as they make the source code available to anyone at no more than your cost of duplication and shipping.
Imagine I personally create a handheld prototype hardware device and port Xpilot to it. There are only two of these made and cost me $2000 each to manufacture these prototypes. I sell one of the device with Xpilot in it. I include a CD with the Xpilot source code with the device. If anyone else wants to compile Xpilot on my prototype device, they would have to buy one from me. They would have to pay me more than $2000 to make it worth my while to bother. And if I say no, I'm not building them anymore because I want to backpack around the world now, am I violating the 'Spirit' of the GPL because I distributed the binaries and sources for a platform which is no longer available? And how would this scenario be different if apple decided to stop selling iphone SDKs? Are you requiring apple's SDK to be open source? or requiring apple's hardware to be open source? why don't you require your intel motherboard to have an open source BIOS as well, seeing that you can't play Xpilot on a PC without paying some BIOS company like Pheonix for the right to boot your computer.
If the XPilot authors did not want anyone selling their software they should have used a different license.
Perhaps the GPLv4 will have an explicit anti-apple clause.
--jeffk++
I believe that precision millisecond stock trading globally is the real reason behind the IEEE 1588v2 precision time protocol. The cisco 9000 enterprise switch supports it. Support has been lacking in smaller switches. The only other group using PTPv2 is the cell phone industry.
The interesting part of PTPv2 for me is that it is used in the 802.1AS protocol ( http://www.ieee802.org/1/pages/802.1as.html ) which is one of the foundations of Audio Video Bridging (AVB) http://www.ieee802.org/1/pages/avbridges.html - Which allows for real time low latency low jitter media streams transported via ethernet with guaranteed bandwidth.
Just yesterday I was joking with friends: Forget about stealing the rounded pennies from bank accounts, criminals could re-program the PTPv2 implementation in switches to steal milliseconds of time during trading!
Anyways, back on the original question, no, network speed is not so crucial once all of your packets are properly timestamped.
--jeffk++
Interestingly, amazon web services now has a import/export service, where they will accept your USB drive via courier and import it into their "Simple Storage Service" aka S3.
http://aws.amazon.com/importexport/
--jeffk++
Yes, I think it was doing arp poisoning.
The thing is that there was a 75 megabyte file on the computer which contained the IP addresses, hostnames, user names, and email passwords and ftp passwords of users and computers in the same room. I watched the file grow in size as it collected passwords!
This is another reason why self-signed ssl certificates can be more secure than having no certificate at all, despite the warnings firefox gives you.
--jeffk++
I have a 'friend'... yeah, that's it... Who back in 1997 or so had a co-located server running RedHat linux with a vulnerable dns server. It was attacked. Multiple times. I only noticed - I mean my 'friend' only noticed when a further attack caused /etc/passwd to be broken. When the box was taken down and analyzed, it was noted that the system put the ethernet port in promiscuous mode and was sniffing all of the traffic for the co-located boxes that were connected to the same switch. A file on the disk contained all the user names and protocols and passwords of the people with accounts on those co-located systems.
So when you log in with unencrypted FTP, how do you know that the server that you are connecting to is not sitting in a rack next to a compromised system sniffing the traffic? How often do people misconfigure their servers and switches like this, and how would you know?
--jeffk++