From mboxrd@z Thu Jan 1 00:00:00 1970 From: Harald Hoyer Subject: Re: [dracut] rd.shell won't work if "root=" is missing Date: Fri, 15 Apr 2016 12:04:23 +0200 Message-ID: <5710BCA7.30204@redhat.com> References: Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: initramfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" To: tammo-v8BkjJTOJzf1qYPpFx2fzhvVK+yQ3ZXh@public.gmane.org, initramfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org On 05.01.2016 10:11, tammo-v8BkjJTOJzf1qYPpFx2fzhvVK+yQ3ZXh@public.gmane.org wrote: > Hi, > > i just stumbled upon a detail in dracut that i found surprising and appears as a small bug to me. > > If you boot a dracut image without giving a "root=" parameter (or e.g. a misspelled argument to that parameter) the system will _not_ drop you into a shell, even if rd.shell is set. IMHO this is not the expected behavior, as the manual says that rd.shell will drop you into a shell if rd.shell is set and no root is found. So i would suggest to change the "die()" function in the dracut-lib.sh and add a test for rd.shell in line 460 (getargbool 0 "rd.debug=" && emergency_shell) too. The system will already drop you in a shell if you have a "rd.debug" parameter set, but i think that rd.shell (and probably "rdshell" too) alone should be sufficient, as this is more convinient and less surprising. > > I could prepare a patch, if you prefer that. > > If this is not some kind of bug, you may call it a "wishlist item" and/or change the documentation accordingly. > > kr > Tammo > yes... changed with https://github.com/dracutdevs/dracut/commit/9d74747f926d038f2e149d70a9c8f955861ab354