Date: Fri, 3 May 2019 23:14:19 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: Garrett Wollman <wollman@csail.mit.edu> Cc: freebsd-net@freebsd.org Subject: Re: Disabling 9k jumbo clusters on mce(4) Message-ID: <20190503201419.GC2716@kib.kiev.ua> In-Reply-To: <23756.39015.553779.526064@khavrinen.csail.mit.edu> References: <23756.29338.729453.227593@khavrinen.csail.mit.edu> <20190503174715.GB2716@kib.kiev.ua> <23756.39015.553779.526064@khavrinen.csail.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, May 03, 2019 at 03:37:11PM -0400, Garrett Wollman wrote: > <<On Fri, 3 May 2019 20:47:15 +0300, Konstantin Belousov <kostikbel@gmail.com> said: > > > On Fri, May 03, 2019 at 12:55:54PM -0400, Garrett Wollman wrote: > >> Does anyone have an easy patch to keep mce(4) from trying to use 9k > >> jumbo mbuf clusters? I think I went down this road once before but > >> the fix wasn't as obvious as it is for the Intel drivers. (I assume > >> the hardware is not so broken that it requires packets to be stored in > >> contiguous physical memory.) > > mce(4) only uses 9k clusters when you configured jumbo size larger than 4k. > > See r336407. > > Yes, and the point is to make it not use them *ever*, no matter what > the MTU is. 9k clusters are Bad News; given the choice I would just > as soon rip out all the code for greater-than-page-size clusters, but > that would be way more divergence-from-upstream in my source tree than > I can reasonably manage. In fact I was wrong, I forgot that code, sorry. It does not use clusters at all, instead the revision I mentioned loads a chain of mbufs of the MCLBYTES size for receive. Jumbo size only defines the number of mbufs.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20190503201419.GC2716>