Re: [Qemu-devel] [PATCH V8 1/4] tests/migration: Convert x86 boot block compilation script into Makefile

[Andrew Jones]

On Mon, Sep 03, 2018 at 01:45:33PM +0200, Juan Quintela wrote:
> Andrew Jones <address@hidden> wrote:
> > On Sat, Sep 01, 2018 at 01:11:12AM -0400, Wei Huang wrote:
> >> The x86 boot block header currently is generated with a shell script.
> >> To better support other CPUs (e.g. aarch64), we convert the script
> >> into Makefile. This allows us to 1) support cross-compilation easily,
> >> and 2) avoid creating a script file for every architecture.
> >> 
> >> Note that, in the new design, the cross compiler prefix can be specified by
> >> setting the CROSS_PREFIX in "make" command. Also to allow gcc pre-processor
> >> to include the C-style file correctly, it also renames the
> >> x86-a-b-bootblock.s file extension from .s to .S.
> >> + * tests/migration/rebuild-x86-bootblock.sh to update,
> >> + * and then remember to send both in your patch submission.
> >> + */
> >> +endef
> >> +export __note
> >> +
> >> +.PHONY: all clean
> >> +all: x86-a-b-bootblock.h
> >> +
> >> +x86-a-b-bootblock.h: x86.bootsect
> >> +  echo "$$__note" > header.tmp
> >> +  xxd -i $< | sed -e 's/.*int.*//' >> header.tmp
> >> +  mv header.tmp $@
> >
> > The shell script this Makefile is replacing used mktemp
> > for a randomly named tmp dir. Shouldn't we continue to
> > use random names?
> 
> We shouldn't be trying to create the file twice in parallel, no?

Yeah, you're right. This should be safe.

> 
> >
> >> +
> >> +x86.bootsect: x86.boot
> >> +  dd if=$< of=$@ bs=256 count=2 skip=124
> >> +
> >> +x86.boot: x86.o
> >> +  $(CROSS_PREFIX)objcopy -O binary $< $@
> >> +
> >> +x86.o: x86-a-b-bootblock.S
> >> +  $(CROSS_PREFIX)gcc -m32 -march=i486 -c $< -o $@
> >> +
> >> +clean:
> >> +  @rm -rf *.boot *.o *.bootsect
> >
> > We don't want to remove the generated header file when cleaning?
> 
> No, because we only generated the header file when running native, have
> the right crosscompilers, etc.  In general we only recompile it when we
> change the test file, otherwise we store the generated ones.
> 
> Idea is:
> We have a c file that does what we want.
> We compile it (when we have a propper compiler) and store it on the
> tree.
> We compile it for each test.
> 
> This is more "like firmware" than normal "qemu" source code.
>

Right. I momentarily forget that when replying to this patch.

Thanks,
drew