Re: Need help tracing regular write activity in 5 s interval

[Padraig Brady]

Ah I see this is a FAQ:

http://marc.theaimsgroup.com/?l=linux-kernel&m=101279800921627&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=99099294313922&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=98381407016589&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=97661027509987&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=96125325519621&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=95061999803667&w=2 *
http://marc.theaimsgroup.com/?l=linux-kernel&m=94240877002500&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=94036109819436&w=2

*here be dragons

Padraig.

Padraig Brady wrote:
> I'm sure it will :-)
> 
> However this it just masking the "problem", and I don't
> think it's "buggy CDROM drives" as I've tried 3 different
> machines with the following drives:
> 
> SAMSUNG DVD-ROM SD-612
> TOSHIBA DVD-ROM SD-C2402
> CREATIVE CD5233E
> 
> and they all show the same problem. I.E. logs filling with
> "VFS: Disk change detected on device ide1(22,0)".
> 
> magicdev essentially does:
> 
> while (1) {
>     cd = open ("/dev/cdrom", O_RDONLY|O_NONBLOCK);
>     if (ioctl (cd, CDROM_DRIVE_STATUS, CDSL_CURRENT) == CDS_DISC_OK) {
>         /* do stuff */
>     }
>     close(cd);
>     sleep(2);
> }
> 
> Note, it's the open() that causes the check_media_changed(),
> so why does this always return true? Is there a way you
> can turn it it off? echoing [01] to /proc/sys/dev/cdrom/check_media
> made no difference.
> 
> Also related, why does the LED flash on every ATA command?
> Is this controlled by the drive or ide controller?
> Are you telling me that windows would flash the LED every so often
> to automount CDs?
> 
> thanks,
> Padraig.
> 
> Erik Andersen wrote:
> 
>> On Tue Jun 04, 2002 at 12:21:20PM +0100, Padraig Brady wrote:
>>
>>> As an aside, Nautilus (1.0.4) does stuff every 2 seconds
>>> (checking is there a CD inserted) that causes the disk LED to flash.
>>> The same action also causes the kernel (2.4.13) to fill up the ring
>>> buffer with: "VFS: Disk change detected on device ide1(22,0)".
>>
>>
>>
>> This should fix the symptom...
>>
>> --- linux/fs/block_dev.c.orig    Tue Jun  4 06:03:44 2002
>> +++ linux/fs/block_dev.c    Tue Jun  4 06:03:44 2002
>> @@ -582,8 +582,11 @@
>>      if (!bdops->check_media_change(dev))
>>          return 0;
>>  
>> +    #if 0
>> +    /* Polling buggy CD-ROM drives can fill the logs.  Make it 
>> shutup. */
>>      printk(KERN_DEBUG "VFS: Disk change detected on device %s\n",
>>          bdevname(dev));
>> +    #endif
>>  
>>      sb = get_super(dev);
>>      if (sb && invalidate_inodes(sb))
>>  -Erik
>>
>> -- 
>> Erik B. Andersen             http://codepoet-consulting.com/
>> --This message was written using 73% post-consumer electrons--