On Thu, 16 Jan 2025 at 09:26, David Woodhouse <dw...@infradead.org> wrote: > > On Thu, 2024-07-18 at 09:20 +0200, Stefano Garzarella wrote: > > If a file imported from Linux is touched, emit a warning and suggest > > using scripts/update-linux-headers.sh. > > > > Also check that updating imported files from Linux are not mixed with > > other changes, in which case emit an error. > > > > Signed-off-by: Stefano Garzarella <sgarz...@redhat.com> > > This now seems to complain about a commit which simply adds a file to > update-linux-headers.sh and simultaneously adds the newly-imported > file. Such as commit aa274c33c39e7de981dc195abe60e1a246c9d248. > > Was that the intent? Do we really need to have *three* separate > commits, one for update-linux-headers.sh, one to import the new header, > and then the third to add the QEMU code which actually uses the new > header? > > $ scripts/checkpatch.pl > 0001-update-linux-headers-import-linux-kvm_para.h-header.patch > WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? > #36: > new file mode 100644 > > WARNING: added, moved or deleted file(s) imported from Linux, are you using > scripts/update-linux-headers.sh? > #41: FILE: include/standard-headers/linux/kvm_para.h:1: > +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ > > ERROR: headers imported from Linux should be self-contained in a patch with > no other changes > #99: FILE: scripts/update-linux-headers.sh:63: > -e 'linux/kernel' \
Just to state for the record the conclusion of our conversation on IRC; yes, the error is correct. Header-imports from Linux should be in a patch of their own which has no changes except the results of running update-linux-headers.sh (and ideally we would not have let aa274c33c39e7 through code review). So, yes, you have a series of three patches: * change update-linux-headers.sh * automated-change-only commit resulting from running the script * the change you want to make that requires the new header thanks -- PMM