[OpenAFS-devel] Re: pthreading the bosserver

[Benjamin Kaduk]

On Wed, 4 Sep 2013, Benjamin Kaduk wrote:

> On Fri, 9 Aug 2013, Benjamin Kaduk wrote:
>
>> On Fri, 9 Aug 2013, Benjamin Kaduk wrote:
>> 
>>> Maybe, I should just try to implement my idea and we will see if it works 
>>> or not.
>> 
>> A sketch of what I had in mind is up at 
>> https://github.com/kaduk/openafs/commits/chas-bozo ; it passes minimal 
>> smoke-testing on my dev box (running a ptserver that doesn't have a prdb).
>> It does not have the extra "shutting down" flag value that jhutz mentioned, 
>> for now.
>> 
>> At 72 insertions/75 deletions, it doesn't really have a "smaller code" 
>> endorsement going for it ... what do people think?
>
> I'm assembling a changeset merging the still-useful bits of my patchset with 
> Chas's heavy lifting on the pthreading front, at 
> https://github.com/kaduk/openafs/commits/newbozo .  I changed the commit-IDs 
> for everything past the first commit (which was verified by buildbot 
> already), so as to not cause confusion on gerrit.  (I also abandoned my 
> gerrit changes for my old patchset.)

I ended up tweaking a number of things, so I left the 'newbozo' branch 
with separate patchsets to indicate what changed, and have a 
'newbozo-squashed' branch which is the squashed version of the same 
patches (modulo the last two, which came in after I started rebasing onto 
master).  Rebasing onto master revealed a couple other issues which are 
fixed in gerrit but not on my github branch.

I did not attempt to track attribution for all the bits I changed; 
hopefully I did not add any bugs to Chas's work -- I do not want to sully 
his good name.

> Once I have squahsed, reviewed and tested things, I plan to push it to gerrit 
> for general review.

I reused the 'pthread-bos' label; it's changesets 10284-10294.

-Ben