Re: [PATCH 8/9] arm: initial drop

[Andrew Jones <drjones@redhat.com>]

On Thu, Oct 17, 2013 at 12:16:18PM +0200, Andrew Jones wrote:
> On Wed, Oct 16, 2013 at 06:06:35PM -0700, Christoffer Dall wrote:
> > > diff --git a/arm/boot.c b/arm/boot.c
> > > new file mode 100644
> > > index 0000000000000..375e8708a7c54
> > 
> > this file's indentation is also funny, you should really check your
> > editor configuration :)
> 
> Actually the editor prefers kernel style, I had to keep trying to
> force the other style on it, and now see below that I wasn't always
> successful. '++expected' should have had just a tab, not a
> tab+4spaces...
> 
> > > --- /dev/null
> > > +++ b/arm/boot.c
> > > @@ -0,0 +1,27 @@
> > > +#include "libcflat.h"
> > > +#include "arm/bootinfo.h"
> > > +
> > > +static bool info_check(u32 var, char *expected)
> > > +{
> > > +    char var_str[9];
> > > +    snprintf(var_str, 9, "%x", var);
> > > +    while (*expected == '0' || *expected == 'x')
> > > +	    ++expected;
> > > +    return !strcmp(var_str, expected);
> > > +}
> > > +
> > > +int main(int argc, char **argv)
> > > +{
> > > +    int ret = 0;
> > > +
> > > +    if (argc < 3) {
> > > +	printf("Not enough arguments. Can't test\n");
> > > +	return 1;
> > > +    }
> > > +
> > > +    if (!strcmp(argv[0], "info"))
> > > +	ret = !info_check(mem32.size, argv[1])
> > > +		|| !info_check(core.pagesize, argv[2]);
> > > +
> > > +    return ret;
> > > +}
> > 
> > I'm actually a little confused, when does this main get invoked and by
> > whom and what are we testing for here?
> 
> See cstart.S:start 'bl main' below for the who invokes. And
> arm/unittests.cfg, also below, for the (poorly documented) test case
> definition. You'll see we config the test to have 256G memory, and then
> here we confirm that we read the bootinfo correctly, i.e. it says 256G.

The above 'G's are of course 'M's. Who remapped my keyboard?

drew