Re: Back to BFS+TOI issue

On 13/03/12 08:09, Oleksandr Natalenko wrote:
> Great.
>
> Setting no_multithreaded_io to 1 did the trick: I've hibernated and
> resumed my laptop 4 times successively (3 w/LZF and 1 w/LZO) with no
> problems at all.
>
> Missing mutex somewhere?
I've always thought it sounded like an issue with the multithreading, 
but I've looked over the code repeatedly, fixed a couple of issues in 
error paths but still not nailed it. Since you can reliably reproduce it 
with BFS, perhaps it's time for me to give BFS a try and see if I can 
also reliably reproduce it. Perhaps it's also time to finally come up 
with a tool to allow analysis of an image post-resume-failure.

Regards,

Nigel