From: Konrad Rzeszutek Wilk Subject: Re: [Xen-devel] Re: [PATCH 2/3] xen/granttable: Grant tables V2 implementation Date: Fri, 18 Nov 2011 13:05:00 -0500 Message-ID: <20111118180500.GA19469@phenom.dumpdata.com> References: <4EC3B62F.6080702@oracle.com> <1321451372-13596-1-git-send-email-annie.li@oracle.com> <4EC62FE5.2080608@oracle.com> <1321614166.3664.311.camel@zakaz.uk.xensource.com> <20111118135221.GC12433@phenom.dumpdata.com> <1321624845.3664.354.camel@zakaz.uk.xensource.com> <4EC6837F.1050201@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <4EC6837F.1050201@oracle.com> Sender: linux-kernel-owner@vger.kernel.org To: annie li Cc: Ian Campbell , "jeremy@goop.org" , "xen-devel@lists.xensource.com" , "kurt.hackel@oracle.com" , "linux-kernel@vger.kernel.org" , Paul Durrant List-Id: xen-devel@lists.xenproject.org > So we have two candidates: > xen_raw_printk + panic > panic + earlyprintk=xen > > If panic does work with earlyprintk, then the latter one is better. > Otherwise, there will be duplicated string printed out with > 'earlyprintk=xen'. The idea is just to have one function. Whichever prints the string and panics the machine. If 'panic' does this properly (and properly meaning it actually prints data when using the earlyprintk=xen as well as console=hvc0) printout system the we cuold just use 'panic' and not worry about it. But if it does not, then we (and by we I mean you) should provide a variant of panic() that prints the data properly using the earlprintk mechanism. Preferrabily to make it generic.