Comment 12 for bug 927515

Regression or not, it's still an annoying bug.

I'm attaching performance data from my older computer, where this is a nuisance. In the two minutes I let Thunderbird run, it fetched less than 1000 emails - which means just a few emails per second (which is not good IMO!).

The latest comment indicates that this would be eCryptfs related - for the record, I'm also running eCryptfs. Seeing the top perf results, it looks like they are crypto related, but I don't think this is the only cause. The fact that it also seems to spend a significant amount of time in pthread_mutex_lock/unlock also looks like a programming error to me.

 14,78% thunderbird-bin [kernel.kallsyms] [k] aes_encrypt
  3,65% thunderbird-bin libfreebl3.so [.] rijndael_decryptBlock128
  3,56% thunderbird-bin libmozsqlite3.so [.] zeroPage
  3,50% thunderbird-bin [kernel.kallsyms] [k] aes_decrypt
  2,56% thunderbird-bin libpthread-2.15.so [.] pthread_mutex_lock
  1,84% thunderbird-bin libpthread-2.15.so [.] __pthread_mutex_unlock_usercnt
  1,80% thunderbird-bin [kernel.kallsyms] [k] _raw_spin_unlock_irqrestore
  1,75% thunderbird-bin [kernel.kallsyms] [k] __copy_from_user_ll_nozero
  1,71% thunderbird-bin libfreebl3.so [.] shaCompress