Re: Firefox hangs in state 'kqread'
Jeroen Ruigrok/asmodai wrote:
> -On [20050304 22:02], Matthew Dillon (dillon@apollo.backplane.com)
> wrote:
>> If it recovers on its own then it probably isn't a bug in the kq
>> code. It kinda sounds like a DNS lookup issue. Both mozilla and
>> firefox have dns lookup issues if domain name being looked up has
>> DNS problems.
>
> No Matt, this is way different, at least what I am experiencing.
>
> Normally when it is blocking on the DNS lookups at least the app reacts
> normally, I can still open new tabs and whatnot.
>
> This time however when it blocks it locks up, cannot do anything and it
> appears the application has totally locked, ready to dump core.
> Switching virtual desktops shows the application in its 'outline' state
> with the graphics toolkit (I hope you know what I mean with this).
> And after a few seconds or a few minutes it will recover.
>
> Unless something with regard to resolving was changed to block harder
> than before.
>
I have no idea whether this might be related at all, but last week
when I was playing with dfly after bringing it up to -current I began
to get a lot of these from the console (/var/log/messages):
Mar 8 03:01:03 workstation kernel: [diagnostic] cache_resolve: had to recurse on usr
Mar 8 03:01:03 workstation kernel: [diagnostic] cache_resolve: had to recurse on share
Mar 8 03:01:03 workstation kernel: Warning: vnode 0xcabd1298 clean/cache_resolution race detected
Mar 8 03:01:03 workstation kernel: Warning: vnode 0xcabd1378 clean/cache_resolution race detected
Mar 8 03:01:03 workstation kernel: Warning: vnode 0xcabdc438 clean/cache_resolution race detected
Mar 8 03:01:03 workstation kernel: Warning: vnode 0xcabdc438 clean/cache_resolution race detected
Mar 8 03:01:03 workstation kernel: Warning: vnode 0xcac27198 clean/cache_resolution race detected
-Mike
討論串 (同標題文章)
完整討論串 (本文為第 9 之 27 篇)：