Finding a rogue IO source
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
I have an unused disk ( no mounted filesystems ) that I put to sleep ( hdparm -Y ) and every 10 minutes, the link is reset to process a request coming from a short lived process named "pool". I think it is a kernel work queue. I inserted a dump_stack() but it stops when it gets to the syscall and so doesn't trace back into what work item is running. How can I figure out what work item is doing this?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQEcBAEBAgAGBQJSQyVcAAoJEJrBOlT6nu75uqkH/j0zivSWfM5XsaDDJZ4gjb8i kvVlaidw9Wu2cv7lgVQB8Ml3eTmqx3ia9I72H9OWJcm324EJ10dD+3jiWAoT2T72 3uojSlbSrZ+0Yp+91EgKjc0/5D2Kz4vVI2rELHYt6ATA8rfLnA6Syrebr3zgY05m 41/pzBieT/lUBpQzWfkVxOtKk+Ma46uAmu1Ipo+G8wwFehbjT/LN57Bv3uFM4M7J sFaBAPCs8nNeAzR+Ud1Yhj6zNZnNct2bIm5oo7Lt7JPuBMzli8q+FVFLrvXOWU8M WXhkalBTYqlmNLXbNsJtYO2IfLFNlNsrYCoyS4IHKvHj8lSzlAKU/dO/OtbJFng= =/RfH
-----END PGP SIGNATURE-----
Hi!
Quoted 9 lines · Phillip Susi
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1I have an unused disk ( no mounted filesystems ) that I put to sleep ( hdparm -Y ) and every 10 minutes, the link is reset to process a request coming from a short lived process named "pool". I think it is a kernel work queue. I inserted a
dump_stack()but it stops when it gets to the syscall and so doesn't trace back into what work item is running. How can I figure out what work item is doing this?
Maybe you could try something like this:printk(KERN_ERR "wakeup %d", current->pid);
-Michi
Signature · 2 lines
programing a layer 3+4 network protocol for mesh networks see http://michaelblizek.twilightparadox.com