Date: Wed, 29 May 2019 14:52:00 -0600 From: Ian Lepore <ian@freebsd.org> To: Andriy Gapon <avg@FreeBSD.org>, Niclas Zeising <zeising@freebsd.org>, src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r348355 - head/sys/dev/iicbus Message-ID: <d1128088420c6e52721fb5df2280ca73096bf5c0.camel@freebsd.org> In-Reply-To: <def030c0-80a5-84ca-bb48-7009aa34e69c@FreeBSD.org> References: <201905290908.x4T98L89066643@repo.freebsd.org> <c3f1c60b-24b2-6098-501a-8cb81ef66d57@freebsd.org> <def030c0-80a5-84ca-bb48-7009aa34e69c@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 2019-05-29 at 15:12 +0300, Andriy Gapon wrote: > On 29/05/2019 14:54, Niclas Zeising wrote: > > On 2019-05-29 11:08, Andriy Gapon wrote: > > > Author: avg > > > Date: Wed May 29 09:08:20 2019 > > > New Revision: 348355 > > > URL: https://svnweb.freebsd.org/changeset/base/348355 > > > > > > Log: > > > revert r273728 and parts of r306589, iicbus no-stop by default feature > > > Since drm2 removal, there has not been any consumer of the feature in the > > > tree. I am also unaware of any out-of-tree consumer. > > > More importantly, the feature has been broken from the very start, both > > > before and after r306589, because the ivar was set on a device that does > > > not support it and it was read from another device that also does not > > > support it. > > > A bus-wide no-stop flag cannot be implemented as an ivar as iicbus > > > attaches as a child of various drivers. Implementing the ivar in each > > > and every I2C driver is just impractical. > > > If we ever want to implement this feature properly, then probably the > > > easiest way to do it would be via a flag in the softc of iicbus. > > > In fact, we might have to do that in the stable branches if we want to > > > fix the code for them. > > > Reported by: ian (long time ago) > > > MFC after: 1 month (maybe) > > > X-MFC-note: cannot just merge the change, must keep drm2 happy > > > > > > > Hi! > > Just a note, be aware that drm2 lives on in ports as drm-legacy-kmod. I haven't > > tested, but, from the description above I worry that it will affect the port. > > What do you think? > > Oh, I forgot about that one... > I think that it could be affected if it still uses FreeBSD iic code. > I guess I might have to revert the change. > > I don't think so, because I don't think this change ever worked. I'm not sure how anybody convinced themselves that it did. It attempts to retrieve ivars from a device that doesn't have them, so the net effect is that the nostop variable is initialized from stack garbage. Maybe whoever wrote and tested it was lucky enough to have that accidentally be consistently zero or non-zero, so their testing appeared to work. Looking at the drm2 code that is the only user of this, it appears that the nostop value is only used in the case where the driver falls back to using the builtin intel_iicbb_driver. That driver relies on iicbus_transfer_gen() which is where the nostop kludge was added. That's the fundamental problem in all of this: the right thing to do, IMO, would have been to implement the iicbus_transfer method directly in the intel_iicbb_driver (probably by just cut-and-pasting the code from iicconf.c then doing whatever is necessary to ignore stops). And we can still do that, pretty trivially, if necessary. -- Ian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?d1128088420c6e52721fb5df2280ca73096bf5c0.camel>