[2/2] 2.6.23-rc5: known regressions with patches
dwmw2 at infradead.org
Thu Sep 6 06:28:11 EDT 2007
On Thu, 2007-09-06 at 16:04 +0530, Satyam Sharma wrote:
> You shouldn't push this even for 2.6.24 ... I can't see why/how a runtime
> BUG() scores over erroring out at build-time itself. And if there is no
> codepath that leads to that BUG() at runtime, then what's the point of
> adding dead code ...
> So I wonder if what you're actually looking for is some kind of Kconfig
> dependencies that will *prevent* the kind of .config from being generated
> that Ingo ran into ?
I looked at that but decided against it. There's too much hand-holding
and arbitrary 'automatic' crap in the Kconfig crap already, and I
couldn't see a way to do it that didn't make that worse.
As long as we no longer break randconfig builds, it'll be fine. It's not
as if people _run_ those kernels, let alone actually exercise the code
path in question.
More information about the linux-mtd