this forum has been dead for a while. did laclasse drop off of the face of the earth or something?
but anyway, it's been lifeless here, so i decided i'd start a thread with some questions/points i've been meaning to post for a while. [after i finished, i realized this message is pretty massive. have fun reading]
firstly, to anyone who uses the alsa intel8x0 driver (pretty much everyone):
do you have problems with random noise coming out of your speakers occasionally? i notice that sometimes, when i'm playing a song, it will turn all... 'staticy,' like you would hear when getting poor reception from a radio station. sometimes it lasts for just a couple seconds, and sometimes it lasts for several songs, which sucks. i haven't been able to discern anything that specifically causes this, but i have noticed that it happens on occasion when:
1. there is significant hard drive activity.
2. when the laptop fans spin up from rest.
3. when there is a change in activity; e.g. when I start up something cpu intensive when the laptop was idle before.
and, of course, sometimes it just happens randomly. (or it's triggered by something that's not easily detected). these are just the events i've been able to note off of the top of my head - sometimes they cause the static, sometimes they don't. there are no log messages at all.
at first glance, i would assume that it was audio latency or something, and i was going to apply some audio latency patches to my kernel. however, point number 2 - the fan starting up - does not fit this. the fans are controlled by temperature via the bios, not by acpi in linux, and would not cause enough of a system load for latency to become a problem.
which leads me to another conclusion - the only other thing i can think of that fits all three points is a power drain. all of those points require a shift in power - it takes power to spin up hard drives, run the processor at high loads, as well as to spin up those big fans on the bottom. so to compensate for this power spike, it's possible that a slight amount of power is taken from the rest of the laptop, before the spike is over and everything balances out. i'm sure everyone also knows that laptop sound cards aren't exactly the best pieces of hardware, and actually are pretty crappy. or maybe the alsa drivers are just sensitive like that.
although i'm not from sager, and i didn't design or build the laptop. so i could very well be completely wrong - it's just a guess.
i've only seen one or two posts about this in the gentoo forums, which leads me to believe it's not a very widespread problem. supposedly the oss kernel driver does not have this problem. however, i'm reluctant to switch, because the alsa driver has other nifty features - such as support for the ac'97 component, which allows for 3d sound spatialization. plus, i'm an audiophile who's sensitive to sound quality, and the alsa drivers usually just plain sound better.
i really want to get an echo indigo, but there are no alsa drivers for those, which also sucks.
i was just wondering if it's just a problem with my configuration and i'm the only one who has this, or if there are others...
[an update; something i thought of as i was typing the later part of this message. i tried decreasing the max timeslices to 10 ms. it helped a tiny bit - the static problem is _slightly_ less likely to occur. it still happens, though. and now, the audio is more likely to just plain skip and stutter under high load. that doesn't help at all. it suggests that it may be related to latency, but does not completely rule out the power theory]
[another update (... gives you an idea of how long it took to do this post. it's a little sad) - as i was sifting through dmesg to find the wireless info referenced later on in this post, i saw that something went retarded with my drives, and dma was disabled. that was causing the audio skipping and stuttering - easily fixed by reactivating dma. the static problem still persists, though]
... in the same vein as the static problem, does anyone have problems with the command line audio players? when i try to play something using ogg123 or mpg123, they play anywhere from 1 to 30 seconds of a song, and then freeze. this problem does not occur with xmms, though - i was thinking of the possibility of it being alsa output vs. oss output (xmms is using the alsa 0.9x output plugin), but then i remembered that ogg123 uses alsa 9 too. (i haven't had much time to speculate about this - i just discovered it about an hour ago.)
... another one related to alsa drivers - i have problems with the sound when i turn the pcm volume to max. the sound quality goes to hell. a lot of people seem to have this problem on the gentoo forums. the only way to fix this is to turn the pcm volume down to at least 70 (some people have to go as low as 40), and turn the volume on the speakers up. i was wondering if that was the case for everyone - is it an actual problem with the alsa driver, or just certain people's hardware?
ok, next is to anyone who uses the ck patches on their kernel (if anyone):
have you tried the latest patchset, 2.4.21-ck3? (http://members.optusnet.com.au/ckolivas/kernel/) it causes problems with swap - when the swap tries to activate on boot, the computer hard locks and i have to use the power button. i talked to con about this, and he suggested removing the 1010 and 1011 patches, ck autoregulating vm and swap prefetching, respectively. that does fix the problem - however, i would like to have swap prefetching; it's a pretty nifty feature, especially with a large amount of ram like 1 gig. i tried running without swap, which normally should not be a problem; however i got random freezes without it. that's clearly not a good thing, so it was time to reactivate the swap partition. anyone using ck have problems activating swap? (i.e. my config or not?)
on that note, my problem is listed as a bug on con's page. hooray, i'm mentioned! oh wait, that's not a good thing...
on the note of the wireless minipci card - it turns out i was _completely_ wrong about this. once again, it was me assuming and jumping to conclusions, something i'm actually incredibly adept at...
it does _not_ run through pcmcia like i thought and mentioned before. at all. i just enabled wireless support (non-hamradio) in the kernel under network devices, and modularized the hermes 802.11b chipset (orinoco/prism2/symbol) option, as well as both options under it - hermes in plx9052 based pci adapter support (which probably isn't even necessary), and prism 2.5 pci 802.11b adaptor support (which is necessary - the card is listed by lscpi -v as a harris semiconductor prism 2.5 wavelan chipset). i didn't even have to tell any modules to autoload - on boot, the comp detected the card automatically and loaded the appropriate modules. here is the pertinent part of dmesg:
hermes.c: 4 Dec 2002 David Gibson <hermes@gibson.dropbear.id.au>
orinoco.c 0.13b (David Gibson <hermes@gibson.dropbear.id.au> and others)
orinoco_pci.c 0.13b (David Gibson <hermes@gibson.dropbear.id.au> & Jean Tourrilhes <jt@hpl.hp.com>)
Detected Orinoco/Prism2 PCI device at 02:03.0, mem:0x80000000 to 0x80000FFF -> 0xfa1b3000, irq:16
Reset done......................<lots more dots, cut to prevent ugly line wrapping in message>;
Clear Reset..............................<same here>;
pci_cor : reg = 0x0 - 90E1 - 8EED
eth1: Station identity 001f:0005:0001:0003
eth1: Looks like an Intersil firmware version 1.3.5
eth1: Ad-hoc demo mode supported
eth1: IEEE standard IBSS ad-hoc mode supported
eth1: WEP supported, 104-bit key
eth1: MAC address 00:90:4B:0A:52:06
eth1: Station name "Prism I"
eth1: ready
and i know it's not pcmcia, because i have no pcmcia/cardbus support in the kernel, and pcmcia-cs is not installed. i was really wrong about that.
while i have no wireless network to test this on, the hard part - getting the kernel to detect and activate the card - is over. connecting to a network using wireless-tools is a piece of cake from here.
i don't know how many people have this working already, or if anyone needed this at all. but there have been a few posts about getting wireless up, so it's here if anyone still needs it.
another one for 8887 (888x) users:
does anyone know how to change the agp aperture with this laptop in linux? i can tell from posts around here that the 56xxes can. but the 8887 bios is a gui piece of **** that can't do anything useful. linux always detects the aperture as 64M. i know for a fact that the aperture has more of an effect in linux than it does in windows (and laclasse's posts around here seem to agree). and with 1 gig of ram, using only 64 megs of it... is a waste. especially when the vram is twice as large as the aperture. it would be more effective the other way around.
i've seen posts in the other areas of sagerforums with info about using programs in windows to do it, but i'm not brave enough to screw around with chipset registers in linux by myself (without a program, that is).
this is the last one, i assure you.
i don't know if you've heard about the usb 2.0 crap that was recently revealed.
quick summary: people were complaining about computers not having usb2.0. so a while ago, the usb forum pulled a fast one - they renamed usb1.1 to usb2.0, and left usb2.0 as usb2.0. so now there are two different versions of usb2. the former usb1.1 is now called usb2.0 'full-speed,' while the real usb2.0 is called usb2.0 'hi-speed.' so a lot of people got laptops from certain companies (sony is one example) that claimed to have usb2.0, but really only had usb1.1. the laptop manufacturers didn't even know about this.
so i was wondering if the usb2.0 in our sagers is the real usb2.0. the kernel seems to load the usb2 ehci-hcd hub driver, which suggests that it's real usb2, but i was wondering if anyone had actually tested a usb 2 device - i don't have any.
all right, i'm done. really. i promise.
but anyway, it's been lifeless here, so i decided i'd start a thread with some questions/points i've been meaning to post for a while. [after i finished, i realized this message is pretty massive. have fun reading]
firstly, to anyone who uses the alsa intel8x0 driver (pretty much everyone):
do you have problems with random noise coming out of your speakers occasionally? i notice that sometimes, when i'm playing a song, it will turn all... 'staticy,' like you would hear when getting poor reception from a radio station. sometimes it lasts for just a couple seconds, and sometimes it lasts for several songs, which sucks. i haven't been able to discern anything that specifically causes this, but i have noticed that it happens on occasion when:
1. there is significant hard drive activity.
2. when the laptop fans spin up from rest.
3. when there is a change in activity; e.g. when I start up something cpu intensive when the laptop was idle before.
and, of course, sometimes it just happens randomly. (or it's triggered by something that's not easily detected). these are just the events i've been able to note off of the top of my head - sometimes they cause the static, sometimes they don't. there are no log messages at all.
at first glance, i would assume that it was audio latency or something, and i was going to apply some audio latency patches to my kernel. however, point number 2 - the fan starting up - does not fit this. the fans are controlled by temperature via the bios, not by acpi in linux, and would not cause enough of a system load for latency to become a problem.
which leads me to another conclusion - the only other thing i can think of that fits all three points is a power drain. all of those points require a shift in power - it takes power to spin up hard drives, run the processor at high loads, as well as to spin up those big fans on the bottom. so to compensate for this power spike, it's possible that a slight amount of power is taken from the rest of the laptop, before the spike is over and everything balances out. i'm sure everyone also knows that laptop sound cards aren't exactly the best pieces of hardware, and actually are pretty crappy. or maybe the alsa drivers are just sensitive like that.
although i'm not from sager, and i didn't design or build the laptop. so i could very well be completely wrong - it's just a guess.
i've only seen one or two posts about this in the gentoo forums, which leads me to believe it's not a very widespread problem. supposedly the oss kernel driver does not have this problem. however, i'm reluctant to switch, because the alsa driver has other nifty features - such as support for the ac'97 component, which allows for 3d sound spatialization. plus, i'm an audiophile who's sensitive to sound quality, and the alsa drivers usually just plain sound better.
i really want to get an echo indigo, but there are no alsa drivers for those, which also sucks.
i was just wondering if it's just a problem with my configuration and i'm the only one who has this, or if there are others...
[an update; something i thought of as i was typing the later part of this message. i tried decreasing the max timeslices to 10 ms. it helped a tiny bit - the static problem is _slightly_ less likely to occur. it still happens, though. and now, the audio is more likely to just plain skip and stutter under high load. that doesn't help at all. it suggests that it may be related to latency, but does not completely rule out the power theory]
[another update (... gives you an idea of how long it took to do this post. it's a little sad) - as i was sifting through dmesg to find the wireless info referenced later on in this post, i saw that something went retarded with my drives, and dma was disabled. that was causing the audio skipping and stuttering - easily fixed by reactivating dma. the static problem still persists, though]
... in the same vein as the static problem, does anyone have problems with the command line audio players? when i try to play something using ogg123 or mpg123, they play anywhere from 1 to 30 seconds of a song, and then freeze. this problem does not occur with xmms, though - i was thinking of the possibility of it being alsa output vs. oss output (xmms is using the alsa 0.9x output plugin), but then i remembered that ogg123 uses alsa 9 too. (i haven't had much time to speculate about this - i just discovered it about an hour ago.)
... another one related to alsa drivers - i have problems with the sound when i turn the pcm volume to max. the sound quality goes to hell. a lot of people seem to have this problem on the gentoo forums. the only way to fix this is to turn the pcm volume down to at least 70 (some people have to go as low as 40), and turn the volume on the speakers up. i was wondering if that was the case for everyone - is it an actual problem with the alsa driver, or just certain people's hardware?
ok, next is to anyone who uses the ck patches on their kernel (if anyone):
have you tried the latest patchset, 2.4.21-ck3? (http://members.optusnet.com.au/ckolivas/kernel/) it causes problems with swap - when the swap tries to activate on boot, the computer hard locks and i have to use the power button. i talked to con about this, and he suggested removing the 1010 and 1011 patches, ck autoregulating vm and swap prefetching, respectively. that does fix the problem - however, i would like to have swap prefetching; it's a pretty nifty feature, especially with a large amount of ram like 1 gig. i tried running without swap, which normally should not be a problem; however i got random freezes without it. that's clearly not a good thing, so it was time to reactivate the swap partition. anyone using ck have problems activating swap? (i.e. my config or not?)
on that note, my problem is listed as a bug on con's page. hooray, i'm mentioned! oh wait, that's not a good thing...
on the note of the wireless minipci card - it turns out i was _completely_ wrong about this. once again, it was me assuming and jumping to conclusions, something i'm actually incredibly adept at...
it does _not_ run through pcmcia like i thought and mentioned before. at all. i just enabled wireless support (non-hamradio) in the kernel under network devices, and modularized the hermes 802.11b chipset (orinoco/prism2/symbol) option, as well as both options under it - hermes in plx9052 based pci adapter support (which probably isn't even necessary), and prism 2.5 pci 802.11b adaptor support (which is necessary - the card is listed by lscpi -v as a harris semiconductor prism 2.5 wavelan chipset). i didn't even have to tell any modules to autoload - on boot, the comp detected the card automatically and loaded the appropriate modules. here is the pertinent part of dmesg:
hermes.c: 4 Dec 2002 David Gibson <hermes@gibson.dropbear.id.au>
orinoco.c 0.13b (David Gibson <hermes@gibson.dropbear.id.au> and others)
orinoco_pci.c 0.13b (David Gibson <hermes@gibson.dropbear.id.au> & Jean Tourrilhes <jt@hpl.hp.com>)
Detected Orinoco/Prism2 PCI device at 02:03.0, mem:0x80000000 to 0x80000FFF -> 0xfa1b3000, irq:16
Reset done......................<lots more dots, cut to prevent ugly line wrapping in message>;
Clear Reset..............................<same here>;
pci_cor : reg = 0x0 - 90E1 - 8EED
eth1: Station identity 001f:0005:0001:0003
eth1: Looks like an Intersil firmware version 1.3.5
eth1: Ad-hoc demo mode supported
eth1: IEEE standard IBSS ad-hoc mode supported
eth1: WEP supported, 104-bit key
eth1: MAC address 00:90:4B:0A:52:06
eth1: Station name "Prism I"
eth1: ready
and i know it's not pcmcia, because i have no pcmcia/cardbus support in the kernel, and pcmcia-cs is not installed. i was really wrong about that.
while i have no wireless network to test this on, the hard part - getting the kernel to detect and activate the card - is over. connecting to a network using wireless-tools is a piece of cake from here.
i don't know how many people have this working already, or if anyone needed this at all. but there have been a few posts about getting wireless up, so it's here if anyone still needs it.
another one for 8887 (888x) users:
does anyone know how to change the agp aperture with this laptop in linux? i can tell from posts around here that the 56xxes can. but the 8887 bios is a gui piece of **** that can't do anything useful. linux always detects the aperture as 64M. i know for a fact that the aperture has more of an effect in linux than it does in windows (and laclasse's posts around here seem to agree). and with 1 gig of ram, using only 64 megs of it... is a waste. especially when the vram is twice as large as the aperture. it would be more effective the other way around.
i've seen posts in the other areas of sagerforums with info about using programs in windows to do it, but i'm not brave enough to screw around with chipset registers in linux by myself (without a program, that is).
this is the last one, i assure you.
i don't know if you've heard about the usb 2.0 crap that was recently revealed.
quick summary: people were complaining about computers not having usb2.0. so a while ago, the usb forum pulled a fast one - they renamed usb1.1 to usb2.0, and left usb2.0 as usb2.0. so now there are two different versions of usb2. the former usb1.1 is now called usb2.0 'full-speed,' while the real usb2.0 is called usb2.0 'hi-speed.' so a lot of people got laptops from certain companies (sony is one example) that claimed to have usb2.0, but really only had usb1.1. the laptop manufacturers didn't even know about this.
so i was wondering if the usb2.0 in our sagers is the real usb2.0. the kernel seems to load the usb2 ehci-hcd hub driver, which suggests that it's real usb2, but i was wondering if anyone had actually tested a usb 2 device - i don't have any.
all right, i'm done. really. i promise.
