Re: [PATCH RESEND v5 2/2] dmaengine: dw-edma: Add changes to support Channel Separation

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] IRQ vectors allocated by pci_alloc_irq_vectors() are leaked on multiple error paths and during device removal.
--

commit ba0a3a6c27ea5625a5bca375aac32fcdc517816e
Author: Devendra K Verma <devverma@amd.com>

dmaengine: dw-edma: Add changes to support Channel Separation

This commit updates the DMA channel register base address calculation to
support a configurable channel register space separation. It modifies the
probe functions to pass the channel space size to the chip structure.

> diff --git a/drivers/dma/dw-edma/dw-edma-pcie.c b/drivers/dma/dw-edma/dw-edma-pcie.c
> index 07931b1817856..fab4faf68b774 100644
> --- a/drivers/dma/dw-edma/dw-edma-pcie.c
> +++ b/drivers/dma/dw-edma/dw-edma-pcie.c

[ ... ]

> @@ -491,6 +494,7 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
>  	chip->nr_irqs = nr_irqs;
>  	chip->ops = match->plat_ops;
>  	chip->cfg_non_ll = dma_data->cfg_non_ll;
> +	chip->ch_space_sz = dma_data->ch_space_sz;
>  
>  	chip->ll_wr_cnt = dma_data->wr_ch_cnt;
>  	chip->ll_rd_cnt = dma_data->rd_ch_cnt;

[Severity: High]
This is a pre-existing issue, but I noticed a potential resource leak in 
dw_edma_pcie_probe() regarding IRQ vector cleanup.

Does dw_edma_pcie_probe() leak IRQ vectors on error paths and during device
removal?

Earlier in dw_edma_pcie_probe(), vectors are allocated using
pci_alloc_irq_vectors(). However, multiple error paths after this point
return directly without calling pci_free_irq_vectors(). For example, if
pcim_iomap_table() fails:

    chip->reg_base = pcim_iomap_table(pdev)[dma_data->rg.bar];
    if (!chip->reg_base)
        return -ENOMEM;

Similar leaks appear to happen if allocating the logical link blocks fails,
or if dw_edma_probe() fails near the end of the function.

Furthermore, dw_edma_pcie_remove() does not appear to call
pci_free_irq_vectors(), which means these IRQ vectors are leaked when the
device is unbound or the module is unloaded. 

According to PCI subsystem guidelines, every error path after a successful
allocation must clean up the vectors to prevent IRQ resource exhaustion.
Could we add proper error labels to clean up these vectors?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260907101213.1917064-1-devverma@amd.com?part=2