On Thu, May 23, 2013 at 10:22:33PM +0300, Xenia Ragiadakou wrote: > On 05/23/2013 09:29 PM, Dan Carpenter wrote: > >On Thu, May 23, 2013 at 08:51:57PM +0300, Xenia Ragiadakou wrote: > >>This patch replaces the calls to alloc_skb with calls > >>to dev_alloc_skb. > >> > >>dev_alloc_skb has GFP_ATOMIC priority so the > >>replacement does not change the code semantics. > >> > >I don't think this change is correct. What problem are you trying > >to solve? Perhaps we can help you. Also the reason for the change > >should be described in the changelog. > > > >regards, > >dan carpenter > > > > Well, i was reading a tutorial on linux networking where it > was suggested to use dev_alloc_skb() instead of alloc_skb() > in drivers. > > As I told you, I am not at all experienced. No worries. > Where is appropriate to use dev_alloc_skb()? > Can you give me an insight? I don't know much about networking. The problems with this patch are: 1) No reason to apply it in the changelog. 2) The kfree_skb() wasn't updated to match. After sending my email I saw that you fixed this in a later patch and I sent a grumpier than necessary email. Sorry. Splitting that up into two patches breaks git bisect so it's not allowed. Anyways, it's a newbie mistake. No worries again. 3) The comments on dev_alloc_skb() say it's obsolete and we should use netdev_alloc_skb() instead. I don't actually know when it's appropriate to use one or the other honestly... It could be that you're right about this here. I honestly don't know, so that's why it has to be explained to me in the changelog. regards, dan carpenter _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
