Re: [net-next v3 3/3] net: ethernet: mtk_eth_soc: change code to skip first IRQ completely

[Frank Wunderlich]

Am 15. Juni 2025 21:49:43 MESZ schrieb Daniel Golle <daniel at makrotopia.org>:
>On Sun, Jun 15, 2025 at 05:03:18PM +0200, Frank Wunderlich wrote:
>> From: Frank Wunderlich <frank-w at public-files.de>
>> 
>> On SoCs without MTK_SHARED_INT capability (mt7621 + mt7628) the first
>> IRQ (eth->irq[0]) was read but never used. Skip reading it now too.
>> 
>> Signed-off-by: Frank Wunderlich <frank-w at public-files.de>
>> ---
>>  drivers/net/ethernet/mediatek/mtk_eth_soc.c | 11 ++++++++---
>>  drivers/net/ethernet/mediatek/mtk_eth_soc.h |  4 ++--
>>  2 files changed, 10 insertions(+), 5 deletions(-)
>> 
>> diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.c b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> index 9aec67c9c6d7..4d7de282b940 100644
>> --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> @@ -3346,10 +3346,15 @@ static int mtk_get_irqs(struct platform_device *pdev, struct mtk_eth *eth)
>>  		return 0;
>>  
>>  	for (i = 0; i < MTK_ETH_IRQ_MAX; i++) {
>> -		if (MTK_HAS_CAPS(eth->soc->caps, MTK_SHARED_INT) && i > 0)
>> -			eth->irq[i] = eth->irq[MTK_ETH_IRQ_SHARED];
>> +		if (MTK_HAS_CAPS(eth->soc->caps, MTK_SHARED_INT)) {
>> +			if (i == 0)
>> +				eth->irq[MTK_ETH_IRQ_SHARED] = platform_get_irq(pdev, i);
>> +			else
>> +				eth->irq[i] = eth->irq[MTK_ETH_IRQ_SHARED];
>> +		} else if (i < 2)  //skip the 1st and 4th IRQ on !MTK_SHARED_INT
>
>Please use conformant comment style, ie. do not use '//' but always use
>'/* ... */' instead, on a dedicated line.

Ok,not sure if the comment should stay :)
It was a comment for me and i forgot to remove it.

>> +			eth->irq[i] = platform_get_irq(pdev, i + 1);
>>  		else
>> -			eth->irq[i] = platform_get_irq(pdev, i);
>> +			continue;
>>  
>>  		if (eth->irq[i] < 0) {
>>  			dev_err(&pdev->dev, "no IRQ%d resource found\n", i);
>> diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.h b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
>> index 6b1208d05f79..ff2ae3c80179 100644
>> --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
>> +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
>> @@ -643,8 +643,8 @@
>>  #define MTK_MAC_FSM(x)		(0x1010C + ((x) * 0x100))
>>  
>>  #define MTK_ETH_IRQ_SHARED	0
>> -#define MTK_ETH_IRQ_TX		1
>> -#define MTK_ETH_IRQ_RX		2
>> +#define MTK_ETH_IRQ_TX		0
>> +#define MTK_ETH_IRQ_RX		1
>>  #define MTK_ETH_IRQ_MAX		3
>
>Shouldn't MAX be 1 now?

No as we still need to loop over the first 3 irqs
on the older devicetrees. Or was the MAX
intented for anything else (e.g. array size for
storing the sw irqs - not changed this yet,but
imho thi should be 2 now if SHARED_INT does
not need 3 entries)?


regards Frank