From mboxrd@z Thu Jan 1 00:00:00 1970 From: mark.rutland@arm.com (Mark Rutland) Date: Thu, 21 Jan 2016 12:11:04 +0000 Subject: [PATCH 00/19] arm64 kexec kernel patches v13 In-Reply-To: <20160120025621.GC2999@dhcp-128-65.nay.redhat.com> References: <20160119123230.GB2904@dhcp-128-65.nay.redhat.com> <1453248900.21755.8.camel@infradead.org> <20160120025621.GC2999@dhcp-128-65.nay.redhat.com> Message-ID: <20160121121103.GE2581@leverpostej> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jan 20, 2016 at 10:56:21AM +0800, Dave Young wrote: > On 01/19/16 at 04:15pm, Geoff Levand wrote: > > On Tue, 2016-01-19 at 20:32 +0800, Dave Young wrote: > > > Geoff, another question about kexec-tools part is, can the kexec > > > -tools code > > > been written in kernel? We have the infrastructure for kexec_file_load. > > > > I see no technical reason why the arm64 kernel cannot support > > kexec_file_load. > > Cool, care to port it to kernel so that we have kexec_file_load only in arm64 > we do not need to support both kexec_load and kexec_file_load? I have reasons for wanting kexec_load, even if we have kexec_file_load. For example, being able to test modified DTBs during development, and being able to kexec to non-Linux OSs. I don't think we should drop kexec_load, as kexec_file_load is designed such that it cannot be used for these purposes, while kexec_load is more general. I appreciate that we will need kexec_file_load for Secure Boot, and it would be perfectly fine for your kdump userspace to require and always use kexec_file_load. Thanks, Mark.