[Solved] [bug report] System unresponsive when using k3b
Posts: 959
I have reported this before with the earlier alphas, betas and rc1, and still no change with rc2
Of special note below is the way the IDE hard disk is reported as SATA. This is the same in windows so the Linux system is not mis-reporting the interface.
Basically, my laptop is pretty much unusable while I am burning a cd using the default settings and this has not changed since the release of SM3.4-3 earlier this year. Mouse movements are jerky and hesitant, typical of a system that is overloaded and any clicks can take 30-60 seconds to get a response, or sometimes I have to wait until burning is finished, same goes for any keyboard input, though this behavior is improved by setting the read-ahead using hdparm -c1 -d1 -a8 -u1 -k1 /dev/hdc as root. Read-ahead is not retained across reboots, which are necessary due to mine being a laptop.
When read-ahead is set, the first part of a burn has little or no impact, but as the drive climbs to 16X speeds, the system starts to struggle with many micro-pauses steadily getting worse until the burn is complete.
Burning a SM6rc2 iso image without read-ahead had an average speed of just over 10x, the same image with read-ahead was 14.5x, and I know my system is capable of much higher speeds.
Toshiba Tecra M30
Centrino 1.86Ghz 533Mhz fsb
GeForce Go 6200 64Mb PCIe
Intel915 Chipset
100Gb IDE HDD. (System has a converter or something similar that connects it as SATA)
1Gb DDR SODIMM RAM
Has anybody else had this problem?
[bug report] System unresponsive when using k3b
Posts: 2280
Mike did you read this thread http://www.mepis.org/node/10035
it hung in my mind cause it was so off the wall. It might not be the same problem but he was playing with an intel board , wont hurt to check it out . Probably wont help rich but stranger things have happened with electronics.
jim
Yes I did see that thread
Posts: 959
Yes I did see that thread and the reason I read it was my system has 2 misaligned resource pointers at every boot and like rgnglzrd who started that thread, mine changes address every time too, so does an IBM R51 I am working on right now, and it happens in both ubuntu and Mepis on both machines and neither of them have the via chipset. The IBM has an ati chipset and mine has the intel915.
Another thing I forgot to mention was my laptop simply will not permit dma for the optical drive, which I think, if it would, I would not have this problem, but to be fair, my optical drive is removeable, so there may be a reason for that. Still, we should be able to work with that.
Strange thing is, when I installed the latest Ubuntu onto my laptop, I had no issues with the optical drive, it worked flawlessly with no system slow-down, even though Ubuntu and Mepis use the same kernels (I still had no dma), but ubuntu is simply not for me. Their version of the file browser (even with kubuntu) with the lack of viewing options (namely no milticolumn) makes it totally useless to me because I'm all over my hard disk and network drives every day. I did not like their sudo for everything root, so it lived for no more than half a day on my system before I shot and killed it.
Long live Mepis - it's the best, even if it has issues on some hardware.
Mike P
unresponsive when runnig k3b
Posts: 123
So now I'm confused. 
I'm burning a disc now, with several other resource hungry programs running with no problems. I can't seem to find any conflicts with other programs but, I have discovered lots of programs I rarely if ever use.
Could my previous problems have been a conflict with a system program running in background? Don't you just love intermittent problems.
i know prelink.bin gets
Posts: 2299
i know prelink.bin gets fired off on some regular basis ( i think cron.weekly). Might've been the culprit
Newbie or not Newbie, there's always a question
The source of the problem has been revealed, but not fixed yet
Posts: 959
timkb4cq from Mepislovers responded to a secondary post on this same issue.
The problem seems to happen when PATA drives are used with an adapter on an Intel SATA controller. In many cases DMA cannot be turned on at all on these drives.
There is a new SATA Piix driver under development, but it's not really stable yet.
http://www.thinkwiki.org/wiki/Problems_with_SATA_and_Linux#No_DMA_on_DVD_drive
The link shows earlier model Intel chipsets, but mine is not listed, being their most recent, so I am uncertain if their suggestions would work for me. Will pass this info on to the devs.
Thanks for all your help guys
Mike P
Simple work-around
Posts: 959
After reporting the link from timkb4cq to the devs, Warren responded with a simple change to my grub bootloader, which I tested first, it worked beautifully and now I am returning to give thanks to the community and devs for their support.
Added hdc=noprobe to grub. I am burning a dvd right now and no system slow-down at all. I am pushing my laptop pretty hard too and not a single glitch or pause.
Thank you sooo much Warren and timkb4cq from Mepislovers
Mike P
RE: simple workaround - a question
Posts: 519
Exactly where in menu.lst does one place "hdc=noprobe" ???
--
GreyGeek
kernel params
Posts: 147
On 'kernel' line, like this:
kernel /boot/vmlinuz-2.6.15-25-386 root=/dev/sda7 nomce quiet vga=791 resume=/dev/sda5 hdc=noprobe
Excellent workaround! Well done!!! I'm still excited
Rgds,
-dim
Minor Correction
Posts: 5513
A minor correction for all. Dim, since you posted the kernel line to the forum, the forum software makes it appear as if the "hdc=noprobe" line is on a separate line. I'm sure you put it in as a single, continuous line. So for everyone the command line, starting with "kernel..." and ending with "hdc=noprobe" is one line, not two
Let me try posting it this way:
kernel /boot/vmlinuz-2.6.15-25-386 root=/dev/sda7 nomce quiet vga=791 resume=/dev/sda5 hdc=noprobe
(yup, just as I thought, one line)
Thanks again to m_pav for finding the fix
Jon
Minor Correction
Posts: 2280
Dont know whats up Jon but its rendering as one line on my rig.....
But i dont think it hurts to make sure all are clear on that point. Just a heads up.
jim
Thanks Jon for correction, I
Posts: 147
Thanks Jon for correction, I don't have a habit yet on BBCode to get more clean presentation...
To bring more light on this workaround: seems current H/W vendors are moving from old standard IDE interfaces to something new (or re-new?) and more performant, SCSI-like solutions. Of course, no problem happens on Windoze, but on Linux currently some mismatching going on associating a right (best) driver to device. For example, my SATA CD/DVD+RW is seen under MEPIS as IDE device and is still working, but with one small problem - there is no way to enable DMA on it! It means any access to device are going on a very low speed (1500 KB/s) and using fully(!) CPU time. So, on speed I've likely back to 10-years old laptop and due high CPU usage - your laptop became unusable during burning (or even DVD movie playing!)...
What is changing with "hdc=noprobe": hdc (CD/DVD) is no more seen as IDE but like a new SCSI (pseudo) device! So, no more '/dev/hdc' but '/dev/scd0', and managed by right 'ata_piix' driver (according my 'dmesg' output). BTW, there is no 'ata_piix' module in '/lib/modules/...', seems it's compiled withing MEPIS kernel.
Don't know if real fix will come soon, but current workaround makes life much more better!
Best regards!
-dim
system unresponsive when using k3b
Posts: 123
I have the same problem when burning cd's on both of these machines using mepis 6 rc1 with all upgrades. Older versions of mepis did not have this problem.
amd athlon 2800
nvidia 5200fx
80 gig hd
756 mb memory
amd athlon 1200+
on board video
40 gig hd
756 mb memory