On 9/17/24 23:44, Sergey Shtylyov wrote: [...] >>> From: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx> >>> Date: Tue, 17 Sep 2024 13:43:24 +0200 >>> >>> Add a label so that a bit of exception handling can be better reused >> >> s/exception/error/. >> >>> in a subsequent if branch of this function implementation. >>> >>> This issue was detected by using the Coccinelle software. >>> >>> Signed-off-by: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx> >>> --- >>> drivers/ata/pata_octeon_cf.c | 9 ++++----- >>> 1 file changed, 4 insertions(+), 5 deletions(-) >>> >>> diff --git a/drivers/ata/pata_octeon_cf.c b/drivers/ata/pata_octeon_cf.c >>> index 0bb9607e7348..62289f6aef95 100644 >>> --- a/drivers/ata/pata_octeon_cf.c >>> +++ b/drivers/ata/pata_octeon_cf.c >>> @@ -848,14 +848,13 @@ static int octeon_cf_probe(struct platform_device *pdev) >>> struct resource *res_dma; >>> int i; >>> res_dma = platform_get_resource(dma_dev, IORESOURCE_MEM, 0); >>> - if (!res_dma) { >>> - put_device(&dma_dev->dev); >>> - of_node_put(dma_node); >>> - return -EINVAL; >>> - } >>> + if (!res_dma) >>> + goto put_device; >>> + >>> cf_port->dma_base = (u64)devm_ioremap(&pdev->dev, res_dma->start, >>> resource_size(res_dma)); >>> if (!cf_port->dma_base) { >>> +put_device: >> >> Ugh... :-/ >> Please use the new-fangled *devm_platform_ioremap_resource() instead of those Ugh, s/*//... >> old-fashioned APIs. > > Another option is to place this error patch somewhere at the end s/patch/path/, of course... :-) > of this function but I think I prefer the new APIs. [...] MBR, Sergey