OK, that work. I was doing an assert(c->aux) on close, but that was miserably failing because we get more close() than open(). Dropped that, and now they seem to be matching. Enough for me now.
On Sun, Nov 29, 2015 at 7:45 AM, barret rhoden <[email protected]> wrote: > On 2015-11-29 at 7:30 'Davide Libenzi' via Akaros wrote: > > Kind of a mess, yeah. That makes std Unix VFS look like Christmas 😀 > > OK, so, to create my ->aux context, don't I have to hook into > > ->open()? attach() and walk() instead? > > Ah, your case is actually a lot easier than the walk/attach case. I > got sidetracked a bit. =) > > All chans that have been opened by your open() will also be closed by > your close(). > > In your open, be sure to set c->flags |= COPEN. In your close(), check > for COPEN, and if set, then you can clean up c->aux. > > Your chan is also unique and will remain so once you set COPEN. For > instance, we'll panic if someone tries to devclone() a COPEN'd chan. > (It's a 9ns bug if that happens, so the user shouldn't be able to > trigger it). > > Barret > > -- > You received this message because you are subscribed to the Google Groups > "Akaros" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Akaros" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/d/optout.
