2023-11-17 02:08:54

by Weili Qian

Subject: Re: [PATCH] crypto: hisilicon - Add check for pci_find_ext_capability

On 2023/11/9 10:13, Chen Ni wrote:
> Add check for pci_find_ext_capability() and return the error if it
> fails in order to transfer the error.
>
> Fixes: 6c6dd5802c2d ("crypto: hisilicon/qm - add controller reset interface")
> Signed-off-by: Chen Ni <[email protected]>
> ---
> drivers/crypto/hisilicon/qm.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
> index 18599f3634c3..adbab1286d4a 100644
> --- a/drivers/crypto/hisilicon/qm.c
> +++ b/drivers/crypto/hisilicon/qm.c
> @@ -3967,6 +3967,9 @@ static int qm_set_vf_mse(struct hisi_qm *qm, bool set)
> int i;
>
> pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_SRIOV);
> + if (!pos)
> + return -ENODEV;
> +

Thanks for your patch. The function qm_set_vf_mse() is called only after SRIOV
is enabled, so function pci_find_ext_capability() does not return 0. This check
makes no sense.

Thanks,
Weili
> pci_read_config_word(pdev, pos + PCI_SRIOV_CTRL, &sriov_ctrl);
> if (set)
> sriov_ctrl |= PCI_SRIOV_CTRL_MSE;
>

2023-11-17 10:37:17

by Herbert Xu

Subject: Re: [PATCH] crypto: hisilicon - Add check for pci_find_ext_capability

On Fri, Nov 17, 2023 at 10:07:00AM +0800, Weili Qian wrote:
>
> Thanks for your patch. The function qm_set_vf_mse() is called only after SRIOV
> is enabled, so function pci_find_ext_capability() does not return 0. This check
> makes no sense.

Perhaps we could add a comment instead?

Thanks,
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

2023-11-22 12:11:59

by Weili Qian

Subject: Re: [PATCH] crypto: hisilicon - Add check for pci_find_ext_capability

On 2023/11/17 18:36, Herbert Xu wrote:
> On Fri, Nov 17, 2023 at 10:07:00AM +0800, Weili Qian wrote:
>>
>> Thanks for your patch. The function qm_set_vf_mse() is called only after SRIOV
>> is enabled, so function pci_find_ext_capability() does not return 0. This check
>> makes no sense.
>
> Perhaps we could add a comment instead?
>
> Thanks,
>

Okay, I will add a comment for this.

Thanks,
Weili