Re: [PATCH] PCI: hisi: fix the hisi_pcie_cfg_read bug when size=4

[Bjorn Helgaas <helgaas@kernel.org>]

On Fri, Nov 27, 2015 at 12:58:00PM +0800, Dongdong Liu wrote:
> Current hisi_pcie_cfg_read code read the RC configuration space,
> this has a problem that the output parameter "*val" have not been
> assigned when size=4, so we fix the bug by "*val = reg_val" when size=4.
> 
> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com>

Wow, that's a pretty egregious bug.  Was the hisi driver ever tested
at all?  We use dword reads in pci_bus_read_dev_vendor_id() to
enumerate devices, so if those didn't return valid data, nothing at
all should have worked.

I tentatively put this on my for-linus branch for v4.4, but I wonder
whether it might be smarter to revert the driver completely until I
have more confidence in it.

What sort of testing have you done on this driver?

> ---
>  drivers/pci/host/pcie-hisi.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/host/pcie-hisi.c b/drivers/pci/host/pcie-hisi.c
> index 35457ec..da677b1 100644
> --- a/drivers/pci/host/pcie-hisi.c
> +++ b/drivers/pci/host/pcie-hisi.c
> @@ -61,7 +61,9 @@ static int hisi_pcie_cfg_read(struct pcie_port *pp, int where, int size,
>  		*val = *(u8 __force *) walker;
>  	else if (size == 2)
>  		*val = *(u16 __force *) walker;
> -	else if (size != 4)
> +	else if (size == 4)
> +		*val = reg_val;
> +	else
>  		return PCIBIOS_BAD_REGISTER_NUMBER;
>  
>  	return PCIBIOS_SUCCESSFUL;
> -- 
> 1.9.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html