[marcelm@something marcelm]$ grep ACPI/usr/src/linux/.config # CONFIG_HOTPLUG_PCI_ACPI is not set # CONFIG_ACPI is not set
I don't know what you are referring to with "hot docking" (and Google doesn't seem to know either), but plugging in a USB/Firewire device, using it and removing it (in other words: hotplugging) works just fine. I have used an USB mouse, scanner, printer, fotocamera and webcam, and two firewire videocameras in GNU/Linux that way, so I think I can state that USB and Firewire stuff mostly works.
But is it a reboot or not?
on
Linux 3.0
·
· Score: 1
Or more precisely:
Can the currently running programs continue running, causing in effect no downtime? Or does it require stopping all processes and starting them again after the new kernel has started, making it effectively a reboot with skipping the BIOS?
That means that your packets should be spaced about 2 to 4 milliseconds apart. This is easy to schedule when your system has a 1 millisecond granularity, but is a disaster when your clocks are 10 milliseconds apart -- your packets end up going out in clumps. Your 100bT network may not care either way, but if you are pushing video over ADSL, 802.11b, or ATM, you may find your packets getting lost along the way.
That's not true.
When an application sends data over the network, it does a send() (or possibly a write()) on a socket. These are systemcalls, so the CPU switches context to the kernel, and the data send by the program is placed in the kernel network buffers. Note that this happens immediately, without waiting for another timeslice.
Then the kernel sends as much as possible (depends on the buffer size on the network card itself) of the data to the network card (after slapping on IP and TCP headers), after which the kernel returns to the application.
Now comes the difference: you suggest that when the network card is done sending the data, it'll have to wait for the next timeslice (because then a context switch to kernelspace occurs and the kernel can do some work), but this is not true!
When the network card is done sending the data, it immediately generates an interrupt (what do you think IRQs are for?). On interrupt, the CPU switches context to the kernel, and the kernel (still having the data to be send in the network buffers) can immediately replenish the buffer on the network card, allowing packets to follow very closely on eachother, regardless of timer granularity.
By the way, somewhat modern network cards can burst packets. That is, they can receive a whole batch of packets from the kernel, which they will then send at the appropriate speed of the medium, so that not everey packet will generate an interrupt. And that's a good thing (tm), because high interrupt loads (think towards 100,000 interrupts/sec for gigabit - without jumbo frames and bursts) are performance killers.
Actually, Potato has around 4500 packages, and Woody will probably be released with something like 8500 packages (and Unstable is closing in on 10,000).
SGI contributed their journaled file system. It's great, but it's still not completed. Reiserfs and ext3fs are far more advanced, and from what I've seen, are the preffered choices. RH with ext3fs, SuSE with Reiserfs. No distro that I'm aware ships with SGI's JFS(?).
SGIs filesystem is called XFS, JFS is a filesystem from IBM (GNU/Linux support for JFS is quite preliminary afaik).
Mandrake (at least the version I decided to have a look at; normally I use Debian) has an option to install on XFS.
And, as other readers have pointed out, XFS is by no means incomplete (but its GNU/Linux support is quite new, that's true).
...as far as I can see, Debian can do this easily: apt-get source packagename.
Then you can./configure and make it yourself. If you only care about the optimizations (and not about the compile-options), you can even do apt-get source -b packagename and it will be built automatically.
The required development packages can be installed with apt-get build-dep packagename by the way, so you don't have to worry about that either.
It is the X Window System. You are talking about version 11 (which is the only one in use today).
The most recent is X Window System version 11 release 6 (X11R6), which is probably the defacto standard.
I mirrored install.iso on ftp://galileo.luon.net/pub/install.iso . It's a mirror in the Netherlands, so it's probably only useful for europeans. The server has a 100mbit inet connection, so: do your worst!
If you should hire Open Source programmers for the job, then keep in mind that your experience with them (either good or bad) will probably not be representative for Open Source programmers in general, since they're all too different. They're different in choice of lifestyle, music taste, religious matters, general social behaviour, coding style, attitude towards Proprietary Software, attitude towards authorities, and whatever aspect you can think of.
Personally, I (as a Free Software lover/user/programmer) would have little problems writing Proprietary Software for a Proprietary Software company (with the notable exception of Microsoft; I've had it with them), providing the paycheck would be fair and the working environment good (I'd require that in any job).
I would have some demands though, but they aren't very extreme. The most important ones include:
-I should be allowed to work on Free/Open Source software in my spare time, and GPL that software.
-I should be allowed to use Free/Open Source software to do the job you require me to do (note that this means I only want to use what works best for the job, be it tools you provide or otherwise; also note that using GPLd software to write program X does not mean that X needs be GPLd)
-It should be clear what I'm exactly allowed to do; for example: am I allowed to open-source a tool I wrote to aid me in the job you required me to do (say, a text editor to write source in)?
Mozilla is slowish and has a large footprint mainly due to the fact that its gui is done double: first it has to render all the widgets in gtk (and possibly your gtk-theme), and *then* it also renders its own gui on top of that. Also the fact that mozilla is somewhat bloated (it includes a mail client, for example) make it slower/larger.
Mozilla's rendering engine, gecko, however is quite fast and becoming better every release. The nice thing about gecko is that it can be used by browsers other than mozilla itself. Galeon is such a browser: It uses gecko for rendering its pages, but its gui need not resemble that of mozilla in any way (except for the scrollbar to scroll your view-window, it seems).
Galeon is a nice browser; It just uses gtk (which makes its gui render quite a lot faster than mozilla's), it includes no mail client, it is fast (because it's simple and uses gecko), and it has some nice features that make it a quite usable browser. It has in fact been my main browser ever since I tried it.
Note: because galeon (and skipstone as well) use gecko, you need to have mozilla installed in order to use galeon of skipstone (or any other browser using gecko).
Although you are right about the fact that your company would probably make less money if it would opensource the program it develops, this does not mean you cannot use Free software to write that very program.
This is an important point. You don't *have* to opensource the program, even if all the software you used to create it was licensed under the GPL. That is an important point.
What the GPL boils down to is that if you use GPL'd source code, and put it in your programs source code, you have to GPL your program as well. This is to prevent anyone from stealing an GPL'd application, and selling it as proprietary software.
----
Re:Meanwhile, on the Debian front...
on
Mozilla 0.9 Out
·
· Score: 1
Ben-Oh is right. There have been unofficial mozilla 0.8 debs for a long time on several places. Mozilla 0.9 will probably be packaged before it's 2 weeks old. Besides of that, nothing stops you from installing the binaries yourself (or compiling it, if you feel like that).
Either you are referring to the rollover to 1,000,000,000, which is no problem (except perhaps for a few ill-written programs), just a number some people get excited about.
The other case is that you're referring to the future overflow of the currently 32bits signed time counters, somewhere in 2038. This is no problem either, because the size (in bits) of the time counter is not specified, although it is currently 32 bits. In C for example (and this is quite a relevant example, since the most UNIX kernels/libs/programs are written in C), this is accomplished by using the type 'size_t' for time counters. In the standard header files, size_t is defined to be 32bits. If this is changed to 64bits, and all programs using time counters are recompiled, we are set to go for another approximately 280,000,000,000 years... at which point we'd have to switch to 128bits (if the human race still exists and unices are still used).
An athlon tbird 1.33 ghz with DDR RAM is already faster than a p4 1.6ghz. And with the palomino coming out soon (planned debut at 1.5ghz, afaik), I know my next computer won't be a p4, and it won't carry RDRAM...
A pIII-733 with 64 MB is *way* overkill for a gateway/router. A T3-line can easily be routed/firewalled by a 486/100 with 8 MB of ram, running GNU/Linux or one of the BSDs of course. Also, the rather powerful 3D-card will not be very useful in a router...
Slashdot Mirror
There has been ACPI support since Linux 2.4.0.
/usr/src/linux/.config
[marcelm@something marcelm]$ grep ACPI
# CONFIG_HOTPLUG_PCI_ACPI is not set
# CONFIG_ACPI is not set
I don't know what you are referring to with "hot docking" (and Google doesn't seem to know either), but plugging in a USB/Firewire device, using it and removing it (in other words: hotplugging) works just fine.
I have used an USB mouse, scanner, printer, fotocamera and webcam, and two firewire videocameras in GNU/Linux that way, so I think I can state that USB and Firewire stuff mostly works.
Or more precisely:
Can the currently running programs continue running, causing in effect no downtime?
Or does it require stopping all processes and starting them again after the new kernel has started, making it effectively a reboot with skipping the BIOS?
When an application sends data over the network, it does a send() (or possibly a write()) on a socket. These are systemcalls, so the CPU switches context to the kernel, and the data send by the program is placed in the kernel network buffers. Note that this happens immediately, without waiting for another timeslice.
Then the kernel sends as much as possible (depends on the buffer size on the network card itself) of the data to the network card (after slapping on IP and TCP headers), after which the kernel returns to the application.
Now comes the difference: you suggest that when the network card is done sending the data, it'll have to wait for the next timeslice (because then a context switch to kernelspace occurs and the kernel can do some work), but this is not true!
When the network card is done sending the data, it immediately generates an interrupt (what do you think IRQs are for?). On interrupt, the CPU switches context to the kernel, and the kernel (still having the data to be send in the network buffers) can immediately replenish the buffer on the network card, allowing packets to follow very closely on eachother, regardless of timer granularity.
By the way, somewhat modern network cards can burst packets. That is, they can receive a whole batch of packets from the kernel, which they will then send at the appropriate speed of the medium, so that not everey packet will generate an interrupt. And that's a good thing (tm), because high interrupt loads (think towards 100,000 interrupts/sec for gigabit - without jumbo frames and bursts) are performance killers.
You might as well read the interview. The "quoted" sentences do not appear in the interview.
That said, the 2.4 kernels (2.4.18 mostly) are doing their work excellently here, on about 20 (mostly different) computers.
when today there are over 2500 packages....
Actually, Potato has around 4500 packages, and Woody will probably be released with something like 8500 packages (and Unstable is closing in on 10,000).
SGI contributed their journaled file system. It's great, but it's still not completed. Reiserfs and ext3fs are far more advanced, and from what I've seen, are the preffered choices. RH with ext3fs, SuSE with Reiserfs. No distro that I'm aware ships with SGI's JFS(?).
SGIs filesystem is called XFS, JFS is a filesystem from IBM (GNU/Linux support for JFS is quite preliminary afaik).
Mandrake (at least the version I decided to have a look at; normally I use Debian) has an option to install on XFS.
And, as other readers have pointed out, XFS is by no means incomplete (but its GNU/Linux support is quite new, that's true).
...as far as I can see, Debian can do this easily: apt-get source packagename .
./configure and make it yourself. If you only care about the optimizations (and not about the compile-options), you can even do apt-get source -b packagename and it will be built automatically.
Then you can
The required development packages can be installed with apt-get build-dep packagename by the way, so you don't have to worry about that either.
Curious... I've been playing .wmv (and other) files with mplayer (and no, that's not windows media player) for nearly half a year...
For those in Europe, especially near/in the Netherlands:
ftp://galileo.luon.net/linux/
It is the X Window System. You are talking about version 11 (which is the only one in use today). The most recent is X Window System version 11 release 6 (X11R6), which is probably the defacto standard.
----
ACK! both the primary and secondary luon.net nameservers are down... Galileo is still reachable on galileo.tte.ele.tue.nl
----
I mirrored install.iso on ftp://galileo.luon.net/pub/install.iso . It's a mirror in the Netherlands, so it's probably only useful for europeans. The server has a 100mbit inet connection, so: do your worst!
----
But yours doesn't fit in a 120-char limited .sig ;O
hehe... other than that, you're right. (bows gracefully)
----
>This thing can flash-fry a whole cow in 10 seconds. Ohhh! But I want in it 5!!! - Homer Simpson
:p
Actually, the line was:
"This thing can flash-fry a whole cow in 40 seconds."
"Aw... But I want it *NOW*!"
And, yes, I'm a nitpicker
----
....wishing to devote their free time for OSS. Do you all agree or disagree?
This is slashdot... how could we all agree on anything?
----
If you should hire Open Source programmers for the job, then keep in mind that your experience with them (either good or bad) will probably not be representative for Open Source programmers in general, since they're all too different. They're different in choice of lifestyle, music taste, religious matters, general social behaviour, coding style, attitude towards Proprietary Software, attitude towards authorities, and whatever aspect you can think of.
Personally, I (as a Free Software lover/user/programmer) would have little problems writing Proprietary Software for a Proprietary Software company (with the notable exception of Microsoft; I've had it with them), providing the paycheck would be fair and the working environment good (I'd require that in any job).
I would have some demands though, but they aren't very extreme. The most important ones include:
-I should be allowed to work on Free/Open Source software in my spare time, and GPL that software.
-I should be allowed to use Free/Open Source software to do the job you require me to do (note that this means I only want to use what works best for the job, be it tools you provide or otherwise; also note that using GPLd software to write program X does not mean that X needs be GPLd)
-It should be clear what I'm exactly allowed to do; for example: am I allowed to open-source a tool I wrote to aid me in the job you required me to do (say, a text editor to write source in)?
----
Mozilla is slowish and has a large footprint mainly due to the fact that its gui is done double: first it has to render all the widgets in gtk (and possibly your gtk-theme), and *then* it also renders its own gui on top of that. Also the fact that mozilla is somewhat bloated (it includes a mail client, for example) make it slower/larger.
Mozilla's rendering engine, gecko, however is quite fast and becoming better every release. The nice thing about gecko is that it can be used by browsers other than mozilla itself. Galeon is such a browser: It uses gecko for rendering its pages, but its gui need not resemble that of mozilla in any way (except for the scrollbar to scroll your view-window, it seems).
Galeon is a nice browser; It just uses gtk (which makes its gui render quite a lot faster than mozilla's), it includes no mail client, it is fast (because it's simple and uses gecko), and it has some nice features that make it a quite usable browser. It has in fact been my main browser ever since I tried it.
Note: because galeon (and skipstone as well) use gecko, you need to have mozilla installed in order to use galeon of skipstone (or any other browser using gecko).
----
If I'm not mistaken, 1024 exobyte would be a zettabyte, and 1024 zettabyte is a yottabyte.
----
Although you are right about the fact that your company would probably make less money if it would opensource the program it develops, this does not mean you cannot use Free software to write that very program.
This is an important point. You don't *have* to opensource the program, even if all the software you used to create it was licensed under the GPL. That is an important point.
What the GPL boils down to is that if you use GPL'd source code, and put it in your programs source code, you have to GPL your program as well. This is to prevent anyone from stealing an GPL'd application, and selling it as proprietary software.
----
Ben-Oh is right. There have been unofficial mozilla 0.8 debs for a long time on several places. Mozilla 0.9 will probably be packaged before it's 2 weeks old. Besides of that, nothing stops you from installing the binaries yourself (or compiling it, if you feel like that).
----
Either you are referring to the rollover to 1,000,000,000, which is no problem (except perhaps for a few ill-written programs), just a number some people get excited about.
The other case is that you're referring to the future overflow of the currently 32bits signed time counters, somewhere in 2038. This is no problem either, because the size (in bits) of the time counter is not specified, although it is currently 32 bits. In C for example (and this is quite a relevant example, since the most UNIX kernels/libs/programs are written in C), this is accomplished by using the type 'size_t' for time counters. In the standard header files, size_t is defined to be 32bits. If this is changed to 64bits, and all programs using time counters are recompiled, we are set to go for another approximately 280,000,000,000 years... at which point we'd have to switch to 128bits (if the human race still exists and unices are still used).
----
or "kernal"
----
An athlon tbird 1.33 ghz with DDR RAM is already faster than a p4 1.6ghz. And with the palomino coming out soon (planned debut at 1.5ghz, afaik), I know my next computer won't be a p4, and it won't carry RDRAM...
And I didn't even mention the price difference.
----
Drill some holes in the case of the Xbox, mount a fan against it, and you should be free of any risk of overheating.
----
A pIII-733 with 64 MB is *way* overkill for a gateway/router. A T3-line can easily be routed/firewalled by a 486/100 with 8 MB of ram, running GNU/Linux or one of the BSDs of course. Also, the rather powerful 3D-card will not be very useful in a router...
----