From mboxrd@z Thu Jan 1 00:00:00 1970 From: andy@warmcat.com (Andy Green) Date: Wed, 24 Mar 2010 13:22:13 +0000 Subject: Device probe order (i2c regulator vs. platform device) In-Reply-To: <20100324131131.GB26453@rakim.wolfsonmicro.main> References: <4BA9D93E.50705@warmcat.com> <005501cacb42$c5bbcde0$513369a0$%szyprowski@samsung.com> <4BAA0AA6.80902@warmcat.com> <20100324131131.GB26453@rakim.wolfsonmicro.main> Message-ID: <4BAA1205.6020309@warmcat.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 03/24/10 13:11, Somebody in the thread at some point said: > On Wed, Mar 24, 2010 at 12:50:46PM +0000, Andy Green wrote: >> On 03/24/10 11:11, Somebody in the thread at some point said: >> >> Hi - >> >>> Right, this is a really common case. Maybe this callback could be added >>> to the regulator framework as well? > > Looks like there's somethin up with the quoting configuration in your Must be bugs in rawhide thunderbird. >> Right, it's a generic issue with needing to delay other device >> registration until something else has completed probe, not just PMU. >> PMU is just the most common parent. > >> A callback is needed after probe() returned without error somehow. > > I rather suspect that the approach you suggest would have some usability > issues, especially on PCs which don't by and by large have custom board > code at all. Something more data driven, or handled more in drivers, is > probably in order if this does get done. Anyway, this should probably > be discussed elsewhere. Fine, at the moment there are no arrangements for it and you have to go hack callbacks in case by case, so anything would be better. I agree it's not arm-specific so this isn't the place, but it'll be sad if it just disappears off the radar. -Andy