Hi Caleb,
On Tue, 12 Nov 2024 at 22:21, Caleb Connolly caleb.connolly@linaro.org wrote:
Under some conditions it's possible to hit the null condition here like when running with OF_LIVE and using the ofnode API before initr_of_live() is called. There is very little null checking for this in the FDT framework, so returning null here can result in weird null pointer conditions.
Instead let's return the control FDT in the fallback case, this is usually what the user was expecting.
Signed-off-by: Caleb Connolly caleb.connolly@linaro.org
drivers/core/ofnode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/core/ofnode.c b/drivers/core/ofnode.c index 950895e72a99..7a7f25fc537c 100644 --- a/drivers/core/ofnode.c +++ b/drivers/core/ofnode.c @@ -152,9 +152,9 @@ void *ofnode_lookup_fdt(ofnode node) uint i = OFTREE_TREE_ID(node.of_offset);
if (i >= oftree_count) { log_debug("Invalid tree ID %x\n", i);
return NULL;
return (void *)gd->fdt_blob; } return oftree_list[i]; } else {-- 2.47.0
Eek I really don't like that, since it will silently return an unexpected value. I think we should panic. Do you know what code path gets you here?
More work is needed in livetree to sort of the rough edges, sadly...
Regards, Simon