Messages in this thread | | | Date | Sun, 10 Aug 2014 11:16:01 +0400 | From | Cyrill Gorcunov <> | Subject | Re: [RFC PATCH 1/5] proc: intoduce proc_inode->pid_entry and is_tgid_pid_entry() |
| |
On Sat, Aug 09, 2014 at 04:28:50PM +0200, Oleg Nesterov wrote: > > > > Hi Oleg, sorry for not responding to previous emails, will try to review this > > things tomorrow (from a glance looks quite good!). Btw, this moment strike my > > eyes -- why don't we use kmem_cache_zalloc here but do assign nils again and > > again, maybe worth to address as well? > > Probably because this doesn't make sense to nullify proc_inode->vfs_inode, > it will be reinitialized at least by inode_init_always() anyway.
Yeah, thanks. I managed to fogeth that we call inode_init_always after inode get allocated. As to series -- looks like a good start, and I think Eric is a way more confident in this area than me so i've nothing to add up to his comments.
| |