| Message ID | 20210913182550.264165-3-maz@kernel.org |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
| Series | PCI: Add support for Apple M1 | expand |
| Context | Check | Description |
|---|---|---|
| robh/checkpatch | success |
On Mon, Sep 13, 2021 at 1:26 PM Marc Zyngier <maz@kernel.org> wrote: > > of_irq_parse_raw() has a baked assumption that if a node has an > interrupt-controller property, it cannot possibly also have an > interrupt-map property (the latter being ignored). > > This seems to be an odd behaviour, and there are no reason why s/are/is/ > we should avoid supporting this use case. This is specially > useful when a PCI root port acts as an interrupt controller for > PCI endpoints, such as this: > > pcie0: pcie@690000000 { > [...] > port00: pci@0,0 { > device_type = "pci"; > [...] > #address-cells = <3>; > > interrupt-controller; > #interrupt-cells = <1>; > > interrupt-map-mask = <0 0 0 7>; > interrupt-map = <0 0 0 1 &port00 0 0 0 0>, > <0 0 0 2 &port00 0 0 0 1>, > <0 0 0 3 &port00 0 0 0 2>, > <0 0 0 4 &port00 0 0 0 3>; > }; > }; > > Handle it by detecting that we have an interrupt-map early in the > parsing, and special case the situation where the phandle in the > interrupt map refers to the current node (which is the interesting > case here). > > Signed-off-by: Marc Zyngier <maz@kernel.org> > --- > drivers/of/irq.c | 17 ++++++++++++----- > 1 file changed, 12 insertions(+), 5 deletions(-) Reviewed-by: Rob Herring <robh@kernel.org>
diff --git a/drivers/of/irq.c b/drivers/of/irq.c index 352e14b007e7..32be5a03951f 100644 --- a/drivers/of/irq.c +++ b/drivers/of/irq.c @@ -156,10 +156,14 @@ int of_irq_parse_raw(const __be32 *addr, struct of_phandle_args *out_irq) /* Now start the actual "proper" walk of the interrupt tree */ while (ipar != NULL) { - /* Now check if cursor is an interrupt-controller and if it is - * then we are done + /* + * Now check if cursor is an interrupt-controller and + * if it is then we are done, unless there is an + * interrupt-map which takes precedence. */ - if (of_property_read_bool(ipar, "interrupt-controller")) { + imap = of_get_property(ipar, "interrupt-map", &imaplen); + if (imap == NULL && + of_property_read_bool(ipar, "interrupt-controller")) { pr_debug(" -> got it !\n"); return 0; } @@ -173,8 +177,6 @@ int of_irq_parse_raw(const __be32 *addr, struct of_phandle_args *out_irq) goto fail; } - /* Now look for an interrupt-map */ - imap = of_get_property(ipar, "interrupt-map", &imaplen); /* No interrupt map, check for an interrupt parent */ if (imap == NULL) { pr_debug(" -> no map, getting parent\n"); @@ -255,6 +257,11 @@ int of_irq_parse_raw(const __be32 *addr, struct of_phandle_args *out_irq) out_irq->args_count = intsize = newintsize; addrsize = newaddrsize; + if (ipar == newpar) { + pr_debug("%pOF interrupt-map entry to self\n", ipar); + return 0; + } + skiplevel: /* Iterate again with new parent */ out_irq->np = newpar;
of_irq_parse_raw() has a baked assumption that if a node has an interrupt-controller property, it cannot possibly also have an interrupt-map property (the latter being ignored). This seems to be an odd behaviour, and there are no reason why we should avoid supporting this use case. This is specially useful when a PCI root port acts as an interrupt controller for PCI endpoints, such as this: pcie0: pcie@690000000 { [...] port00: pci@0,0 { device_type = "pci"; [...] #address-cells = <3>; interrupt-controller; #interrupt-cells = <1>; interrupt-map-mask = <0 0 0 7>; interrupt-map = <0 0 0 1 &port00 0 0 0 0>, <0 0 0 2 &port00 0 0 0 1>, <0 0 0 3 &port00 0 0 0 2>, <0 0 0 4 &port00 0 0 0 3>; }; }; Handle it by detecting that we have an interrupt-map early in the parsing, and special case the situation where the phandle in the interrupt map refers to the current node (which is the interesting case here). Signed-off-by: Marc Zyngier <maz@kernel.org> --- drivers/of/irq.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-)