From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Dave Dodge" Subject: Re: ANNOUNCE: User-space System Device Enumation (uSDE) Date: Thu, 30 Oct 2003 21:29:26 -0500 Sender: linux-raid-owner@vger.kernel.org Message-ID: <200310310229.VAA30940@smarty.smart.net> References: <20031031005854.GC4906@kroah.com> Return-path: To: Greg KH Cc: "Guo, Min" , Steven Dake , Lars Marowsky-Bree , Mark Bellon , linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org, linux-hotplug-devel@lists.sourceforge.net, cgl_discussion@osdl.org, "Ling, Xiaofeng" List-Id: linux-raid.ids Greg KH writes: > On Thu, Oct 30, 2003 at 07:45:08PM -0500, David Dodge wrote: [...] > > Mainly I'm asking because I did try putting a hotplug script into an > > initramfs a few weeks ago (using -test7), and it didn't appear to be > > invoked for e.g. the VESA framebuffer. So I want to make sure this is a > > "future" capability and not something that should have worked :-) > > This is something that should have worked for you today, /sbin/hotplug > does get called during early boot, before init is started up. Okay, I'll keep working at it. Unfortunately my main test system has since had a major hardware failure and I haven't gotten a replacement set up yet. In the meantime, from a quick look at the kernel code I see this: - invocations of /sbin/hotplug normally go through call_usermodehelper. - kernel/kmod.c: call_usermodehelper does this prior to scheduling any work: if (!system_running) return -EBUSY; - init/main.c: system_running is not set non-zero until just prior to starting init. So it looks like calls to hotplug are dropped while compiled-in drivers are initializing. Or am I missing something obvious? I needed a one-line patch to be able to boot into initramfs; do I need another to enable call_usermodehelper earlier?