From: Emanoil Kotsev <deloptes@yahoo.com>
To: Jim Carter <jimc@math.ucla.edu>
Cc: linux-laptop@vger.kernel.org
Subject: Re: kernel strange behaviour
Date: Sat, 11 Oct 2008 03:06:45 -0700 (PDT) [thread overview]
Message-ID: <616547.15762.qm@web53205.mail.re2.yahoo.com> (raw)
Hello Jim,
thank you for your reply here my comments and findings
--- On Sat, 10/11/08, Jim Carter <jimc@math.ucla.edu> wrote:
>
> > I've recently installed a debian prebuild kernel
> 2.6.26.5 on my DELL D520
> > (Centrino Duo Core 2 with 2GB RAM). Before I was using
> 2.6.26 booting
> > from usb drive, but I was missing the ata support for
> my internal drive
> > that's why I had to update. After I installed the
> new kernel I'm not able
> > to watch tv and compile in the same time or record tv
> and watch in the
> > same time. I'm even not able to just record tv....
>
> That's strange. I have a Dell Inspiron 6400, which is
> basically the same
> machine but I think older: CPU is an Intel Core 2 Duo T5600
> @ 1.83GHz, and
My one is T7200 / Core 2 Due / 2GB RAM
> it has 1Gb RAM. As I'm writing this I am watching HDTV
> (looks like MPEG2
> containing mpgv video and a52 audio) captured by a
> HDHomeRun tuner
> ( http://www.silicondust.com/ ) and played by VLC
> ( http://www.videolan.org/vlc/ ) using the XVideo rendering
> interface.
> At the same time I'm compiling a source package.
In fact with the 2.6.26 version everything works fine. Did you use the kernel DVB drivers and what kind of card do you have? My one is Hauppauge HVR-900 with em28xx(_dvb)
I'm using mplayer and mencoder to do the job. The odd thing is that after I upgraded to 2.6.26.5 ( I patched the 2.6.26 kernel ) it stopped working.
Another thing I'm observing now is that if I compile SATA into the kernel it works more smoothly.
Does your machine has SATA - may be it has something to do with DMA or IRQ or I don't know what - I'm not so good in hardware
> Here's what the "top"
> command shows me:
>
> top - 22:07:39 up 5:47, 4 users, load average: 1.57,
> 0.94, 0.53
> Cpu0 : 74.5%us, 6.0%sy, 0.0%ni, 11.3%id, 1.3%wa,
> 2.6%hi, 4.3%si, 0.0%st
> Cpu1 : 1.6%us, 1.0%sy, 0.0%ni, 97.1%id, 0.0%wa,
> 0.0%hi, 0.2%si, 0.0%st
> Mem: 1034772k total, 863240k used, 171532k free,
> 46620k buffers
> Swap: 2104472k total, 0k used, 2104472k free,
> 443000k cached
>
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+
> COMMAND
> 20276 jimc 25 0 66960 37m 21m S 57 3.7
> 3:33.02 vlc
> 25830 root 25 0 16944 13m 4068 R 13 1.4
> 0:00.40 cc1
> 3094 root 15 0 111m 75m 48m S 10 7.5
> 4:41.71 X
> (etc.)
I have had a look at top (I have about 9-10 year linux experience, so I know how to basically track things down) but couldn't figure out why it's slowing down.
>
> So I'm using most (not all) of one core and the other
> is almost unused.
>
> In another trial a few months ago I tried an old 1.0 GHz
> Pentium III
> "Coppermine" with an analog TV tuner, and it
> could just barely compress
> (record) the video and play it back at the same time -- not
> satisfactory
> for a production solution but interesting as a demo. So
> applying to your
> case I can't figure why your machine is so overloaded.
I had a feeling that the machine switched to Pentium I :-)
>
>
> You aren't doing a parallel "make", are you?
What is parallel make? I tried to compile the kernel. I even tried to compile the kernel without watching tv and in about 10minutes everything was so slow that I couldn't scroll the console.
> With only 2 cores and a
> commitment to TV processing, parallel "make"
> would only make trouble.
>
> Suggestion: use "top" to find out what is hogging
> your CPU. Sometimes you
> will find processes that count in the load average but
> which aren't really
> using much CPU. The symptoms suggest thrashing (too much
> continuously
> moving pages between swap and main RAM) -- look at how much
> swap is being
> used, zero in my case. The resident set (column headed
> RES) is the
> important figure for memory use, when thrashing is
> suspected.
thanks I don't understand most of this but I too have a suspission that there is some memory stuff going on as the CPU meter shows 1000Hz most the time and top does not indicate that some program is using that much CPU except gcc which would use as much from the band wide as possible.
In my case swap is also 0 used and I even have about 1GB RAM free
>
> I'm surprised that your distro's kernel was
> unsatisfactory. In the usual
> case that the disc driver (ata_piix in my case) is not
> hardwired in the
> kernel, you can use an initrd to preload it, and most
> distros will take
> care of building the initrd automatically and telling GRUB
> (the booter) to
> read it. I assume standard Debian does so, though I've
> never run Debian on
> an Intel box, only ARM handhelds. (I use OpenSuSE 10.3.)
> It's possible
> but unlikely that you made a bad choice when configuring
> your custom
> kernel. For example, if SMP (multiple processors) were
> turned off, that
> would not help at all.
Well, I recompiled the kernel after I figured out that I have the problem in fact I played about a week with it to findi out that if I compile ata_piix into the kernel things get better. But not much better with 2.6.26.5. I am using initrd my root and data partitions are encrypted. Debian kernel also uses initrd. In fact the debian kernel performs better than the one compiled by me and what is really odd I took the .config from the debian one, didn't change anything and build the same version 2.6.26.5
When I start xmms (mp3 player) it plays few seconds than stops then plays further ...
Crazy thing.
here are the results of top without compiling - just recording and playing
top - 11:29:50 up 16 min, 1 user, load average: 2.40, 1.96, 1.28
Tasks: 166 total, 1 running, 165 sleeping, 0 stopped, 0 zombie
Cpu(s): 66.4%us, 13.8%sy, 0.0%ni, 19.6%id, 0.2%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 2066920k total, 714280k used, 1352640k free, 17576k buffers
Swap: 1469908k total, 0k used, 1469908k free, 346852k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5737 emanoil 20 0 66732 15m 5872 S 84 0.8 5:11.91 mencoder
6019 emanoil 20 0 61256 26m 15m S 10 1.3 0:03.56 gmplayer
4154 root 20 0 312m 47m 5816 S 5 2.3 0:34.02 Xorg
5251 emanoil 20 0 33688 15m 11m S 2 0.8 0:07.16 konsole
5213 emanoil 20 0 22100 7148 5328 S 1 0.3 0:03.00 artsd
1345 root 15 -5 0 0 0 S 1 0.0 0:01.28 kcryptd
5183 emanoil 20 0 31676 11m 9200 S 1 0.6 0:00.94 kwin
5812 emanoil 20 0 2436 1152 852 R 1 0.1 0:00.68 top
1 root 20 0 1808 832 608 S 0 0.0 0:01.18 init
------------------
This is with compiling the kernel after about 5minutes
top - 11:40:15 up 26 min, 1 user, load average: 5.45, 4.21, 2.73
Tasks: 175 total, 3 running, 172 sleeping, 0 stopped, 0 zombie
Cpu(s): 91.1%us, 7.9%sy, 0.0%ni, 0.0%id, 0.0%wa, 0.9%hi, 0.1%si, 0.0%st
Mem: 2066920k total, 904816k used, 1162104k free, 42548k buffers
Swap: 1469908k total, 0k used, 1469908k free, 480544k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5737 emanoil 20 0 66628 15m 5872 S 82 0.8 13:18.44 mencoder
6019 emanoil 20 0 64072 29m 15m R 46 1.5 1:45.38 gmplayer
15287 emanoil 20 0 14200 8800 2172 R 29 0.4 0:01.68 cc1
4154 root 20 0 312m 47m 5816 S 17 2.4 1:16.74 Xorg
5265 emanoil 20 0 69252 36m 11m S 5 1.8 0:13.61 skype
5251 emanoil 20 0 36820 18m 11m S 4 0.9 0:18.92 konsole
15270 emanoil 20 0 2440 1160 852 R 3 0.1 0:00.92 top
1345 root 15 -5 0 0 0 S 2 0.0 0:04.96 kcryptd
4501 haldaemo 20 0 6364 4348 3564 S 2 0.2 0:01.12 hald
5483 emanoil 20 0 34344 14m 11m S 2 0.7 0:03.50 konsole
49 root 15 -5 0 0 0 S 1 0.0 0:00.16 kblockd/0
142 root 15 -5 0 0 0 S 1 0.0 0:00.86 ata/0
2581 root 18 -2 1740 472 396 S 1 0.0 0:01.06 ifplugd
4008 syslog 20 0 1936 644 512 S 1 0.0 0:00.32 syslogd
4061 root 20 0 1872 536 444 S 1 0.0 0:01.38 dd
4487 root 20 0 2020 768 668 S 1 0.0 0:00.06 dhcdbd
4649 root 20 0 3420 1144 988 S 1 0.1 0:00.56 hald-addon-stor
5213 emanoil 20 0 21992 7040 5220 S 1 0.3 0:03.62 artsd
5217 emanoil 20 0 24528 10m 9056 S 1 0.5 0:01.08 kcpuload
5245 emanoil 20 0 47172 12m 10m S 1 0.6 0:00.81 skim
5287 emanoil 20 0 36476 10m 7336 S 1 0.5 0:00.16 knotify
15289 root 20 0 1772 500 424 S 1 0.0 0:00.02 hal-system-kill
5463 emanoil 20 0 144m 43m 22m S 0 2.2 0:11.37 firefox-bin
1 root 20 0 1808 832 608 S 0 0.0 0:01.18 init
-------------------
This is with 2.6.26 (not 2.6.26.5) compiling recording and watching
top - 12:03:21 up 15 min, 1 user, load average: 4.29, 3.41, 1.82
Tasks: 173 total, 3 running, 170 sleeping, 0 stopped, 0 zombie
Cpu(s): 94.7%us, 4.7%sy, 0.0%ni, 0.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 2066748k total, 785604k used, 1281144k free, 27836k buffers
Swap: 1469908k total, 0k used, 1469908k free, 377500k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
6962 emanoil 20 0 68264 16m 5880 S 93 0.8 8:33.07 mencoder
6992 emanoil 20 0 61592 26m 15m S 14 1.3 1:43.88 gmplayer
5461 root 20 0 317m 52m 5708 S 6 2.6 0:51.16 Xorg
13594 emanoil 20 0 12996 6928 2156 R 3 0.3 0:00.10 cc1
6624 emanoil 20 0 35968 16m 11m S 2 0.8 0:16.95 konsole
2221 root 15 -5 0 0 0 S 1 0.0 0:03.49 kcryptd
6489 emanoil 20 0 67252 34m 11m S 1 1.7 0:12.17 skype
5802 haldaemo 20 0 6504 4392 3568 S 0 0.2 0:01.05 hald
6953 emanoil 20 0 34296 14m 11m S 0 0.7 0:02.50 konsole
12693 emanoil 20 0 2440 1156 852 R 0 0.1 0:00.76 top
13334 emanoil 20 0 3584 1332 788 S 0 0.1 0:00.07 make
1 root 20 0 2844 1688 544 S 0 0.1 0:01.09 init
I will try the programs suggested by you - somehow I don't like the vlc player
Can you suggest something (some kernel option) to debug?
regards
next reply other threads:[~2008-10-11 10:06 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-10-11 10:06 Emanoil Kotsev [this message]
-- strict thread matches above, loose matches on Subject: below --
2008-10-10 9:42 kernel strange behaviour Emanoil Kotsev
2008-10-11 6:03 ` Jim Carter
2008-10-11 14:53 ` Thorny
2008-10-11 17:16 ` Emanoil Kotsev
2008-10-12 10:48 ` Thorny
2008-10-13 5:55 ` Emanoil Kotsev
2008-10-13 12:09 ` Thorny
2008-10-13 22:05 ` Emanoil Kotsev
2008-10-14 10:43 ` Thorny
2008-10-14 22:59 ` Emanoil Kotsev
2008-10-15 22:48 ` Emanoil Kotsev
2008-10-13 17:09 ` Thorny
2008-10-13 21:11 ` Emanoil Kotsev
2008-10-14 10:43 ` Thorny
2008-10-14 22:55 ` Emanoil Kotsev
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=616547.15762.qm@web53205.mail.re2.yahoo.com \
--to=deloptes@yahoo.com \
--cc=jimc@math.ucla.edu \
--cc=linux-laptop@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).