On Fri, Nov 08, 2019 at 10:35:30AM +0200, Stefan Mavrodiev wrote:
>
> On 11/7/19 6:56 PM, David Gibson wrote:
> > On Tue, Nov 05, 2019 at 10:09:52AM +0200, Stefan Mavrodiev wrote:
> > > When libfdt applies overlay, which has aliases, they are added to
> > > the __symbols__ node. Since the overlay path is string, it ends with
> > > trailing zero. However, when calculating path length, this symbol should
> > > be excluded. The effect is, that the property is threated as integer
> > > instead of string.
> > >
> > > Here is a simple example. Using test/overlay_base:
> > > / {
> > > test-node {
> > > ...
> > > };
> > > __symbols__ {
> > > test = "/test-node";
> > > ...
> > > };
> > > };
> > >
> > > Compaling simple overlay with alias node:
> > > / {
> > > fragment@0 {
> > > target = <0xffffffff>;
> > > __overlay__ {
> > > foo {
> > > bar = "bar";
> > > phandle = <0x00000001>;
> > > };
> > > };
> > > };
> > > __symbols__ {
> > > foo = "/fragment@0/__overlay__/foo";
> > > };
> > > __fixups__ {
> > > test = "/fragment@0:target:0";
> > > };
> > > };
> > >
> > > After merge:
> > > / {
> > > ...
> > > __symbols__ {
> > > foo = <0x2f746573 0x742d6e6f 0x64652f66 0x6f6f0000>;
> > > ...
> > > };
> > > };
> > >
> > > You can see that the foo value is wrong. It's parsed as int, instead of string. Same
> > > test with the fix:
> > >
> > > / {
> > > ...
> > > __symbols__ {
> > > foo = "/test-node/foo";
> > > ...
> > > };
> > > };
> > >
> > > Signed-off-by: Stefan Mavrodiev <stefan@xxxxxxxxxx>
> > LGTM, but it'd be great if you could add a test case for this as well.
>
> Sure, I could do that. Should I do this by re-sending this series as v2 or
> as separate?
As a v2, please.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
Attachment:
signature.asc
Description: PGP signature
