Hi, On Mon, 2009-03-30 at 09:57 -0400, Ed Cashin wrote: > Thanks for pointing out your motivations. And please feel free to > post patches after reading the HACKING file. Even if a patch doesn't > go right in, it can still be a help to folks by being distributed in > the "contrib" directory. Some existing features started out that way > and migrated into the vblade proper by popular demand.
I am also looking at kvblade. I had to make some minor adjustments to get it to work with a more recent kernel. Those crazy kernel guys love re-naming things. As soon as I have a chance to dig through git and find out where things changed, I'll send a patch so that make just works for old and new. I don't subscribe to LKML, its just too much noise, I just read the changelogs for every rc. > > I think the idea of making vbladed check for a config file is > interesting. > I will send a patch for that, its trivial. Only a few entries so that nice and logger options can be controlled. Exploiting vblade's rudimentary options for access control also comes to mind, but that is not best implemented in a default configuration. In short, I'll soon send a patch for the above, and a few more that can live in the contrib/ directory. For instance, its probably a good idea for a user space target to use mlockall(). That would help to ensure benchmarks from a 'cold start' to be consistent. vblade should also have an option to put it out of the OOM killers reach on start up, depending on how you use it. Heavy use of DM on a storage server + disk sleep can cause interesting things to happen, especially with the new LVM2 liblvm about to emerge in beta. Cheers, --Tim ------------------------------------------------------------------------------ _______________________________________________ Aoetools-discuss mailing list Aoetools-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/aoetools-discuss
