[quagga-dev 3272] Re: [PATCH] non-blocking I/O from client daemons tozebra

Andrew J. Schorr aschorr at telemetry-investments.com
Thu Apr 21 15:07:37 BST 2005

On Thu, Apr 21, 2005 at 03:03:55PM +0100, Paul Jakma wrote:
> On Thu, 21 Apr 2005, Andrew J. Schorr wrote:
> It's right on the limit, it's interpreted as a bit flag in places. 
> It's a number in struct thread alright, but not in other places (ie 
> 'show thread cpu <filter>' - see the filter), I actually got quite 
> confused initially as to why my new thread type wasnt working, till i 
> realised I had put THREAD_BACKGROUND in as 9, instead of replacing 
> the THREAD_QUEUE type i had in.
> changing it to short will save future confusion, and I doubt it'll 
> hurt anywhere.

Hmmm, perhaps I'm confused, but it seems to me that the only change required
is to increase the size of the struct cpu_thread_history "types" field from
unsigned char to unsigned short.  That value is constructed as a bitmask
from (1 << thread->type).  So I don't think thread->type
itself needs to be unsigned short.  But perhaps I'm missing something...


More information about the Quagga-dev mailing list