A missing or empty dma-ranges in a DT node implies a 1:1 mapping for dma
translations. In this specific case, the current behaviour is to zero out
the entire specifier so that the translation could be carried on as an
offset from zero. This includes address specifier that has flags (e.g.
PCI ranges).
Once the flags portion has been zeroed, the translation chain is broken
since the mapping functions will check the upcoming address specifier
against mismatching flags, always failing the 1:1 mapping and its entire
purpose of always succeeding.
Set to zero only the address portion while passing the flags through.
Fixes: dbbdee94734b ("of/address: Merge all of the bus translation code")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Andrea della Porta <andrea.porta@xxxxxxxx>
Tested-by: Herve Codina <herve.codina@xxxxxxxxxxx>
Link: https://lore.kernel.org/r/e51ae57874e58a9b349c35e2e877425ebc075d7a.1732441813.git.andrea.porta@xxxxxxxx
Signed-off-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
diff --git a/drivers/of/address.c b/drivers/of/address.c
index c5b925ac469f..5b7ee3ed5296 100644
--- a/drivers/of/address.c
+++ b/drivers/of/address.c
@@ -459,7 +459,8 @@ static int of_translate_one(const struct device_node *parent, const struct of_bu
}
if (ranges == NULL || rlen == 0) {
offset = of_read_number(addr, na);
- memset(addr, 0, pna * 4);
+ /* set address to zero, pass flags through */
+ memset(addr + pbus->flag_cells, 0, (pna - pbus->flag_cells) * 4);
pr_debug("empty ranges; 1:1 translation\n");
goto finish;
}
