I agree somewhat. I think that if the monolithic kernel is working for now, we should keep it--there's no reason to rewrite something that works. A little look into any possible performance gains vs. the cost of added complexity might help. However, I wholeheartedly agree with you about the binary module idea. Eventually, perhaps by 2.8, there should be a solid set of API's that have stabilized, and there should be a way to load modules with a more abstract interface that only works with those API's. It is a pain to have to use a specific kernel with specific config options just to make use of a binary kernel module. (I'd like to point out, however, that between major versions of Windows there is almost never binary driver compatibility, or at the very least problems can show up.)
There are a few reasons USB keys don't make sense:
They are much more expensive. A 512 MB USB key costs at least $50, while a 650-700 MB CD costs pennies.
They are slower. I know from experience that booting from a USB key can be very slow. (Sometimes the planets align and it works well...) CD drives are a lot faster reading, and somewhat faster writing as well.
Only very new computers boot to USB. My Asus M2400Ne, a beautiful (if a bit flimsy) Centrino laptop, boots to USB like a charm. It doesn't need to be pampered into booting a floppy image or anything. However, most computers can't do that. It's almost guaranteed that they can boot from CD, though.
They are harder to replace. CD-RW's are everywhere, but you have to buy a USB key in a store.
However, a USB key does have the advantage that it is more "live"--CD's would be re-burnt at shutdown while USB keys would be edited in real time. (Because of this, the CD would have a longer lifetime, and also allow you to revert to where you started by rebooting and not saving the home directory.)
I'm looking forward to trying this--it looks really cool.
This is a really great idea, to see how secure Linux is.
Some people have suggested that we test vanilla Linux distributions out of the box. I think this is a dumb idea, and I'll explain why: in any OS, there has to be a balance between security and convenience. Many features that a Linux desktop requires, for example (such as low-level hardware access) would be a security risk if exploited or messed with remotely. In a desktop Linux OS, the balance will be set to a convenient desktop that requires authentication for more important system stuff, while a server OS would have a stricter balance, requiring su or sudo for almost anything. Desktops are also less exposed to the world than servers, and don't run mission-critical services.
This will be fun to see. I'm guessing the server will succumb to some exploit discovered while it is running, that could be patched. What would be interesting is to calculate the percentage of time that exploits (root or otherwise, depending on what you're concerned about) are known against an OS before a patch is available. This is where M$'s patch-once-a-month policy would hurt them a lot.
They wouldn't be obvious enough to do this. There's probably an explanation, like they were re-organizing their pages or something. This also isn't as bad as other sites do--feeding the search engines an entirely different page--but is only adding keywords to the title. (Neither page had a keyword meta tag, oddly enough.)
Can you have more than a person using a Windows computer at the same time with a GUI available?
Yes. The Jetway Magic Twin (review at AnandTech) is a small form factor box running XP that allows you to have two users at once. (It's designed for businesses with users who aren't using nearly as much CPU power as they have, so it can be shared. I think it's some kind of hack on Terminal Service that makes the second monitor a terminal or something.)
That's what I love about Gentoo... it adapts quickly, but never so quickly that it's broken. I'm running it on my laptop and I love it.
Among Intel's chips, the one I'm really drooling over and looking forward to is the dual-core Pentium-M coming out in Q1 2006. It's really going to be awesome--667 MHz FSB, DDR2 (of course), and the graphics chipset that comes in Centrino bundles is getting better too.
...I'm using the DSL ISP I use partly because they give out static addresses and partly because they've got a strong philosophical position that if you've got an Internet connection, all the bits are yours, and you should be able to do absolutely anything you want with it except spamming and cracking.
It is a legal requirement that any cell phone, unlocked or not, must be able to place emergency calls and any network, subscriber or not, must put the call through. This same requirement applies to any phone--you must be able to call 911 from any phone. For example, my phone will accept 911, *911, and 0711 from any of the locked states it is in.
In fact, some charity drives have been organized that collect older phones that don't have a subscription anymore and give them to domestic violence victims for use in emergencies (since they can call 911).
I always do emerge -au world instead; I was just showing emerge -u world for simplicity. I never do it blind like that. (I also don't run mission-critical servers, yet;-)
Isn't it nice having your OS manage all your software? I'd hate to have to go back to checking for updates on each program separately like on Windows (or usually not updating them). Linux distribution writers have really done something awesome with things like apt-get, portage, and the like...
Those aren't XHTML-valid sarcasm tags: you must write them in lowercase: <sarcasm><!-- insert witty comment here --></sarcasm>
I do see your point though... I was just saying that if they're going to have a secret, undocumented, unofficial "policy" of requiring ID *sometimes*, they should offer a search (which arguably protects security *better*) as an alternative.
A choice between ID or a search seems appropriate--either they have to know who you are so you can be held accountable for what you do, or they have to make sure you won't be able to do anything you would need to be held accountable for.
Usually if I'm going to set up a Linux server, I have to hand pick each component with ones I know are supported.
I would hope if you were building a server for any OS you would be hand-picking the components to make sure they're decent anyway. Besides, if I'm building a system, Windows or Linux, I pick hardware that I know is stable and well-supported by whatever OS I'm using.
Driver support for wifi is kinda there (with ndiswrapper)
Not true! There are relatively stable (both are past 1.0) drivers for both the 802.11b and 802.11g wireless cards used in Centrino systems, and a decent fraction of PCMCIA cards are supported as well. Besides, probably more than 1/2 or 2/3 of laptops are Centrino now, solely because the Pentium-M is better than the Pentium 4-M and manufacturers are adding wireless anyway, so they use the bundle.
I'm having a pretty good time with my Linux laptop. It's an Asus M2400Ne running Gentoo. Of the things you listed, 3d graphics, DVD playing, ACPI, wireless, power management, and extra buttons work, and suspend doesn't. (I don't have Bluetooth.)
ACPI has been both a blessing and a curse. Aside from the fact that Microsoft got there first and created a non-standard for Bluetooth (their compiler makes horribly erroneous code), it has standardized one of the things that could have been the undoing of Linux on laptops--if all those features were implemented in vendor-specific proprietary systems, there's little chance driver writers could have kept up.
SpeedStep, in particular, works wonderfully, and gives me almost 5 hours of battery life with light use.
This would defeat part of the purpose of dual-core chips. By sharing the clock generator and distribution circuits between the cores, it saves power. If the clocks are different, they have to be separate and will use more power. (IANACD--I am not a chip designer.)
I understand this, too... the broadcaster is sending out an almost infinite number of "copies" of their show---each stream of bouncing electrons that forms a wave is another copy. If I take this copy, take it out of the air, put it on magnetic tape (or a hard drive) and then later show it again, how is this invalidating a copyright? Personally I don't think broadcasters (cable is a different story) have a leg to stand on. They are dropping flyers from the sky, and complaining if you put one in your pocket to read it later.
It could be a company with an entire exchange. One time a guy from Earthlink called me back (I was rearranging phone cords to fix DSL interference, so I had to hang up) and it only showed six digits of the phone number, i.e. the area code and exchange. Either he used a "generic" outside line and they don't want people calling them back, or they don't give out extensions by policy. (I actually know that for a fact--if you were talking to someone, they're not allowed to forward your call back to them without their permission... quite aggravating sometimes.)
Actually, it doesn't require a black and a color cartridge, it only requires two cartridges. One of their troubleshooting techniques is to swap the cartridges. If the same cartridge (black or color) is having trouble, it's the cartridge itself. If the other one is having trouble, it's farther up the chain, either in the printer or on the computer. In other words, you can put a black cartridge in it if that's cheaper.
I finally gave up on my Lexmark 5700 and made my dad buy a laser printer after it coughed up halfway through the third black cartridge with an error. (It's really obscure--the single, useless, information-free LED blinks 12 times in a row, and if you look it up on their website, it means "cartridge overcurrent condition"--in other words, their proprietary chips shorted out. I figure it's a conspiracy, and they are designed to break somewhere between 2/3 and empty, further increasing Lexmark's revenue...)
I have a Pentium-M laptop and I will explain how SpeedStep works in practice.
SpeedStep itself allows the CPU to change frequency (and voltage along with it) to save power. For example, my 1.7 GHz Pentium-M can run at 600, 800, 1000, 1200, 1400, 1500, or 1700 MHz. This is usually controlled by turning up the speed when the CPU usage is high and turning it down when the usage is low. Thus, if you are playing Doom, it will run as fast as it needs to to keep up. These "performance" states are described as "P-states" in the ACPI spec (they run from P0 [full speed] to P7 [slowest speed] on my computer)
There is a second, older type of throttling. that is just called throttling, and described as "T-states" in the ACPI spec. On older Pentium III's, there are two states, T0 (on all the time) and T1 (off half the time). On my Pentium-M, it goes from T0 (on all the time) to T7 (on 1/8 of the time) in 1/8 time increments. This throttling is much less efficient (it has to start and stop the CPU constantly, and still runs the clock at 1.7 GHz when it's on) but is used for a different purpose, as you will see.
ACPI "thermal zones" are objects that consist of a temperature, and "trip points" (temperatures) that trigger "active" and "passive" cooling. Active cooling is set at a lower temperature, and is linked to a fan object that the OS should turn on. Passive cooling is set at a higher temperature, and is linked to the processor object. When the temperature passes the "passive" threshold, the CPU is throttled using T-states.
I actually have a bunch of data about CPU speed (P-state) and throttling (T-state) versus temperature and power usage, and I can tell you that both types of throttling save battery power and run cooler. However, P-states are much more efficient. If you take a 1.7 GHz processor and run it at P0 and T7 (1.7 GHz on 1/8 of the time = about 215 MHz) it runs almost 20 degrees hotter and uses up about 5 Watts (the lowest usage I recorded was about 12.5 Watts, so that's a large fraction) more than running at P7 and T0 (600 MHz on all the time). It's also 1/3 the speed. So basically, P-states are much more efficient, but T-states are what is tied to cooling, probably because they existed first.
The unfortunate problem here is that P-states are much more efficient, but traditionally P-states are tied to usage and T-states are tied to temperature. It is often suggested to use T-states once you are in the bottom P-state (i.e. go 1700*8/8, 1500*8/8,... 800*8/8, 600*8/8, 600*7/8... 600*2/8, 600*1/8) but frankly that doesn't save much power, and does hurt the responsiveness of the computer. (It needs a certain minimum speed to be able to speed itself back up in time to not look laggy...)
The best thing to do if your processor thermal-throttles itself is to 1. cool it better (perhaps attach an air conditioner to the side of the case?), 2. turn down the speed--voluntary throttling may sound like a waste, but it keeps the temp down better than letting the OS throttle it, and it gives better performance, or 3. get a Pentium-M.
I haven't actually been able to compare my Pentium-M to a Pentium 4, since I avoid those like the plague (1.5 hours battery life? how about 5?), but I can say that subjectively, it's quite snappy (thanks to 2 MB L2 cache and Linux's good disk caching) and doesn't show its slower clock except in raw processing work. (If anyone wants me to time a kernel build, email me and I'll do it, you pick the version and.config...)
Yeah. I've always admired Speakeasy for that...too bad I'm stuck with Earthlink. Speakeasy's a bit expensive though.
The advantage of this is that firmware upgrades could make it possible to turn a huge pile of Linksys routers into 1. recurring profit for Linksys, 2. extra money for router owners (this is why they would want to upgrade their routers), 3. biggest hotspot network ever. It would really rock.
My only problem with Netshare is that their services are quite expensive already. I would get their service and do it if their DSL wasn't $100 for the top speed. Phone companies are--for obvious reasons--much cheaper... it's quite unfortunate, because small, real ISP's usually put a ton more work into providing good, useful service and just an honest and fair service, instead of Earthlink's "best efforts", "temporary problems", "network outages", and the 1000's of other excuses they use to not fix your problem.
I've often wondered if Linksys could include a coupon with their wireless routers that allowed you to become a hotspot on their network. You would sign up, it would turn your router into a captive portal, you would be able to add your own computers, and other people would be charged a monthly fee and authenticated against Linksys's servers. Then you'd get a share of the profits. Given how widespread their routers are, this could be a good situation for the router owners, the users, and Linksys.
Remember - skype's program makes at least 50 random connections to other computers in the distributed network, and any one of these could be used to route voice traffic.
Um... Skype uses those only for directory listing and routing messages. Voice traffic, IIRC, goes directly to another IP. Maybe it can route through other nodes, but I didn't think it could.
Slashdot Mirror
I agree somewhat. I think that if the monolithic kernel is working for now, we should keep it--there's no reason to rewrite something that works. A little look into any possible performance gains vs. the cost of added complexity might help. However, I wholeheartedly agree with you about the binary module idea. Eventually, perhaps by 2.8, there should be a solid set of API's that have stabilized, and there should be a way to load modules with a more abstract interface that only works with those API's. It is a pain to have to use a specific kernel with specific config options just to make use of a binary kernel module. (I'd like to point out, however, that between major versions of Windows there is almost never binary driver compatibility, or at the very least problems can show up.)
There are a few reasons USB keys don't make sense:
However, a USB key does have the advantage that it is more "live"--CD's would be re-burnt at shutdown while USB keys would be edited in real time. (Because of this, the CD would have a longer lifetime, and also allow you to revert to where you started by rebooting and not saving the home directory.)
I'm looking forward to trying this--it looks really cool.
This is a really great idea, to see how secure Linux is.
Some people have suggested that we test vanilla Linux distributions out of the box. I think this is a dumb idea, and I'll explain why: in any OS, there has to be a balance between security and convenience. Many features that a Linux desktop requires, for example (such as low-level hardware access) would be a security risk if exploited or messed with remotely. In a desktop Linux OS, the balance will be set to a convenient desktop that requires authentication for more important system stuff, while a server OS would have a stricter balance, requiring su or sudo for almost anything. Desktops are also less exposed to the world than servers, and don't run mission-critical services.
This will be fun to see. I'm guessing the server will succumb to some exploit discovered while it is running, that could be patched. What would be interesting is to calculate the percentage of time that exploits (root or otherwise, depending on what you're concerned about) are known against an OS before a patch is available. This is where M$'s patch-once-a-month policy would hurt them a lot.
They wouldn't be obvious enough to do this. There's probably an explanation, like they were re-organizing their pages or something. This also isn't as bad as other sites do--feeding the search engines an entirely different page--but is only adding keywords to the title. (Neither page had a keyword meta tag, oddly enough.)
Yes. The Jetway Magic Twin (review at AnandTech) is a small form factor box running XP that allows you to have two users at once. (It's designed for businesses with users who aren't using nearly as much CPU power as they have, so it can be shared. I think it's some kind of hack on Terminal Service that makes the second monitor a terminal or something.)
That's what I love about Gentoo... it adapts quickly, but never so quickly that it's broken. I'm running it on my laptop and I love it.
Among Intel's chips, the one I'm really drooling over and looking forward to is the dual-core Pentium-M coming out in Q1 2006. It's really going to be awesome--667 MHz FSB, DDR2 (of course), and the graphics chipset that comes in Centrino bundles is getting better too.
Do you have Speakeasy?
It is a legal requirement that any cell phone, unlocked or not, must be able to place emergency calls and any network, subscriber or not, must put the call through. This same requirement applies to any phone--you must be able to call 911 from any phone. For example, my phone will accept 911, *911, and 0711 from any of the locked states it is in.
In fact, some charity drives have been organized that collect older phones that don't have a subscription anymore and give them to domestic violence victims for use in emergencies (since they can call 911).
I always do emerge -au world instead; I was just showing emerge -u world for simplicity. I never do it blind like that. (I also don't run mission-critical servers, yet ;-)
Or, for the Gentoo users,
Isn't it nice having your OS manage all your software? I'd hate to have to go back to checking for updates on each program separately like on Windows (or usually not updating them). Linux distribution writers have really done something awesome with things like apt-get, portage, and the like...
Those aren't XHTML-valid sarcasm tags: you must write them in lowercase: <sarcasm><!-- insert witty comment here --></sarcasm>
I do see your point though... I was just saying that if they're going to have a secret, undocumented, unofficial "policy" of requiring ID *sometimes*, they should offer a search (which arguably protects security *better*) as an alternative.
A choice between ID or a search seems appropriate--either they have to know who you are so you can be held accountable for what you do, or they have to make sure you won't be able to do anything you would need to be held accountable for.
I would hope if you were building a server for any OS you would be hand-picking the components to make sure they're decent anyway. Besides, if I'm building a system, Windows or Linux, I pick hardware that I know is stable and well-supported by whatever OS I'm using.
Not true! There are relatively stable (both are past 1.0) drivers for both the 802.11b and 802.11g wireless cards used in Centrino systems, and a decent fraction of PCMCIA cards are supported as well. Besides, probably more than 1/2 or 2/3 of laptops are Centrino now, solely because the Pentium-M is better than the Pentium 4-M and manufacturers are adding wireless anyway, so they use the bundle.
I'm having a pretty good time with my Linux laptop. It's an Asus M2400Ne running Gentoo. Of the things you listed, 3d graphics, DVD playing, ACPI, wireless, power management, and extra buttons work, and suspend doesn't. (I don't have Bluetooth.)
ACPI has been both a blessing and a curse. Aside from the fact that Microsoft got there first and created a non-standard for Bluetooth (their compiler makes horribly erroneous code), it has standardized one of the things that could have been the undoing of Linux on laptops--if all those features were implemented in vendor-specific proprietary systems, there's little chance driver writers could have kept up.
SpeedStep, in particular, works wonderfully, and gives me almost 5 hours of battery life with light use.
This would defeat part of the purpose of dual-core chips. By sharing the clock generator and distribution circuits between the cores, it saves power. If the clocks are different, they have to be separate and will use more power. (IANACD--I am not a chip designer.)
I understand this, too... the broadcaster is sending out an almost infinite number of "copies" of their show---each stream of bouncing electrons that forms a wave is another copy. If I take this copy, take it out of the air, put it on magnetic tape (or a hard drive) and then later show it again, how is this invalidating a copyright? Personally I don't think broadcasters (cable is a different story) have a leg to stand on. They are dropping flyers from the sky, and complaining if you put one in your pocket to read it later.
It could be a company with an entire exchange. One time a guy from Earthlink called me back (I was rearranging phone cords to fix DSL interference, so I had to hang up) and it only showed six digits of the phone number, i.e. the area code and exchange. Either he used a "generic" outside line and they don't want people calling them back, or they don't give out extensions by policy. (I actually know that for a fact--if you were talking to someone, they're not allowed to forward your call back to them without their permission... quite aggravating sometimes.)
...or an Athlon64, yeah... how cool do those run anyway?
I know they also have Athlon64-M processors... are those any good?
Is there any way to overclock a 1.7 GHz chip to 2.0 GHz, or is it entirely locked?
Actually, it doesn't require a black and a color cartridge, it only requires two cartridges. One of their troubleshooting techniques is to swap the cartridges. If the same cartridge (black or color) is having trouble, it's the cartridge itself. If the other one is having trouble, it's farther up the chain, either in the printer or on the computer. In other words, you can put a black cartridge in it if that's cheaper.
I finally gave up on my Lexmark 5700 and made my dad buy a laser printer after it coughed up halfway through the third black cartridge with an error. (It's really obscure--the single, useless, information-free LED blinks 12 times in a row, and if you look it up on their website, it means "cartridge overcurrent condition"--in other words, their proprietary chips shorted out. I figure it's a conspiracy, and they are designed to break somewhere between 2/3 and empty, further increasing Lexmark's revenue...)
I have a Pentium-M laptop and I will explain how SpeedStep works in practice.
SpeedStep itself allows the CPU to change frequency (and voltage along with it) to save power. For example, my 1.7 GHz Pentium-M can run at 600, 800, 1000, 1200, 1400, 1500, or 1700 MHz. This is usually controlled by turning up the speed when the CPU usage is high and turning it down when the usage is low. Thus, if you are playing Doom, it will run as fast as it needs to to keep up. These "performance" states are described as "P-states" in the ACPI spec (they run from P0 [full speed] to P7 [slowest speed] on my computer)
There is a second, older type of throttling. that is just called throttling, and described as "T-states" in the ACPI spec. On older Pentium III's, there are two states, T0 (on all the time) and T1 (off half the time). On my Pentium-M, it goes from T0 (on all the time) to T7 (on 1/8 of the time) in 1/8 time increments. This throttling is much less efficient (it has to start and stop the CPU constantly, and still runs the clock at 1.7 GHz when it's on) but is used for a different purpose, as you will see.
ACPI "thermal zones" are objects that consist of a temperature, and "trip points" (temperatures) that trigger "active" and "passive" cooling. Active cooling is set at a lower temperature, and is linked to a fan object that the OS should turn on. Passive cooling is set at a higher temperature, and is linked to the processor object. When the temperature passes the "passive" threshold, the CPU is throttled using T-states.
I actually have a bunch of data about CPU speed (P-state) and throttling (T-state) versus temperature and power usage, and I can tell you that both types of throttling save battery power and run cooler. However, P-states are much more efficient. If you take a 1.7 GHz processor and run it at P0 and T7 (1.7 GHz on 1/8 of the time = about 215 MHz) it runs almost 20 degrees hotter and uses up about 5 Watts (the lowest usage I recorded was about 12.5 Watts, so that's a large fraction) more than running at P7 and T0 (600 MHz on all the time). It's also 1/3 the speed. So basically, P-states are much more efficient, but T-states are what is tied to cooling, probably because they existed first.
The unfortunate problem here is that P-states are much more efficient, but traditionally P-states are tied to usage and T-states are tied to temperature. It is often suggested to use T-states once you are in the bottom P-state (i.e. go 1700*8/8, 1500*8/8, ... 800*8/8, 600*8/8, 600*7/8 ... 600*2/8, 600*1/8) but frankly that doesn't save much power, and does hurt the responsiveness of the computer. (It needs a certain minimum speed to be able to speed itself back up in time to not look laggy...)
The best thing to do if your processor thermal-throttles itself is to 1. cool it better (perhaps attach an air conditioner to the side of the case?), 2. turn down the speed--voluntary throttling may sound like a waste, but it keeps the temp down better than letting the OS throttle it, and it gives better performance, or 3. get a Pentium-M.
I haven't actually been able to compare my Pentium-M to a Pentium 4, since I avoid those like the plague (1.5 hours battery life? how about 5?), but I can say that subjectively, it's quite snappy (thanks to 2 MB L2 cache and Linux's good disk caching) and doesn't show its slower clock except in raw processing work. (If anyone wants me to time a kernel build, email me and I'll do it, you pick the version and .config...)
Yeah. I've always admired Speakeasy for that...too bad I'm stuck with Earthlink. Speakeasy's a bit expensive though.
The advantage of this is that firmware upgrades could make it possible to turn a huge pile of Linksys routers into 1. recurring profit for Linksys, 2. extra money for router owners (this is why they would want to upgrade their routers), 3. biggest hotspot network ever. It would really rock.
My only problem with Netshare is that their services are quite expensive already. I would get their service and do it if their DSL wasn't $100 for the top speed. Phone companies are--for obvious reasons--much cheaper... it's quite unfortunate, because small, real ISP's usually put a ton more work into providing good, useful service and just an honest and fair service, instead of Earthlink's "best efforts", "temporary problems", "network outages", and the 1000's of other excuses they use to not fix your problem.
Good point.
I've often wondered if Linksys could include a coupon with their wireless routers that allowed you to become a hotspot on their network. You would sign up, it would turn your router into a captive portal, you would be able to add your own computers, and other people would be charged a monthly fee and authenticated against Linksys's servers. Then you'd get a share of the profits. Given how widespread their routers are, this could be a good situation for the router owners, the users, and Linksys.
Um... Skype uses those only for directory listing and routing messages. Voice traffic, IIRC, goes directly to another IP. Maybe it can route through other nodes, but I didn't think it could.