[PATCH u-boot-marvell v2] arm: mvebu: turris_omnia: Fixup SATA or PCIe nodes at runtime in DT blob

[Stefan Roese]

On 1/10/22 10:06, Pali Rohár wrote:
> On Monday 10 January 2022 08:43:17 Stefan Roese wrote:
>>> +	/* If mSATA card is not present, disable SATA DT node */
>>> +	if (!mode_sata) {
>>> +		fdt_for_each_node_by_compatible (node, blob, -1,
>>> +						 "marvell,armada-380-ahci") {
>>
>> Why are adding the space before the "(" here? This does not seem the
>> common and preferred coding style AFAICT.
> 
> Hello Stefan! Just to note that I'm adding this space too when writing
> for-loops. E.g.:
> 
> for (i = 0; i < 10; i++) {
> ...
> }
> 
> or:
> 
> fdt_for_each_node_by_compatible (node, blob, -1, compatible) {
> ...
> }

Right. But we already have multiple of these for () helpers in U-Boot
and Linux. And AFACIT all are used without this space:

[stefan at ryzen u-boot (master)]$ git grep for_each drivers/
drivers/block/blkcache.c:       list_for_each_entry(node, &block_cache, lh)
drivers/block/blkcache.c:       list_for_each_safe(entry, n, &block_cache) {
drivers/button/button-adc.c:    dev_for_each_subnode(node, dev->parent) {
drivers/button/button-adc.c:    dev_for_each_subnode(node, parent) {
drivers/button/button-gpio.c:   dev_for_each_subnode(node, parent) {
drivers/clk/clk-uclass.c:       list_for_each_entry(child_dev, 
&clk->dev->child_head, sibling_node) {
drivers/clk/clk_stm32mp1.c:     ofnode_for_each_subnode(subnode, node) {
drivers/core/device-remove.c:   list_for_each_entry_safe(pos, n, 
&dev->child_head, sibling_node) {
drivers/core/device-remove.c:   list_for_each_entry_safe(pos, n, 
&dev->child_head, sibling_node) {
drivers/core/device.c:  list_for_each_entry_safe(pos, n, 
&dev->parent->child_head,
drivers/core/device.c:  list_for_each_entry(uc, gd->uclass_root, 
sibling_node) {

So I would prefer to also use this new helper without the extra space as
well.

Thanks,
Stefan