linux-kernel - Re: [RfC PATCH] fbdev: list all memory bars as conflicting apertures

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 14 Mar 2019 11:37:13 +0100
From:   Daniel Vetter <daniel@...ll.ch>
To:     Gerd Hoffmann <kraxel@...hat.com>
Cc:     dri-devel@...ts.freedesktop.org,
        Daniel Vetter <daniel.vetter@...ll.ch>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        "open list:FRAMEBUFFER LAYER" <linux-fbdev@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [RfC PATCH] fbdev: list all memory bars as conflicting apertures

On Wed, Mar 13, 2019 at 12:07:41PM +0100, Gerd Hoffmann wrote:
> Simply add all pci memory bars to struct apertures_struct in
> remove_conflicting_pci_framebuffers().  That allows to drop
> the res_id parameter.
> 
> TODO: actually remove the res_id parameter.

I think best to do that in a cleanup patch afterwards.
> 
> Signed-off-by: Gerd Hoffmann <kraxel@...hat.com>
> ---
>  drivers/video/fbdev/core/fbmem.c | 22 ++++++++++++++++++----
>  1 file changed, 18 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/video/fbdev/core/fbmem.c b/drivers/video/fbdev/core/fbmem.c
> index cb43a2258c51..e8223c231de2 100644
> --- a/drivers/video/fbdev/core/fbmem.c
> +++ b/drivers/video/fbdev/core/fbmem.c
> @@ -1879,14 +1879,28 @@ int remove_conflicting_pci_framebuffers(struct pci_dev *pdev, int res_id, const
>  {
>  	struct apertures_struct *ap;
>  	bool primary = false;
> -	int err;
> +	int err, idx, cnt, bar;
>  
> -	ap = alloc_apertures(1);
> +	for (cnt = 0, bar = 0; bar < PCI_ROM_RESOURCE; bar++) {
> +		if (!(pci_resource_flags(pdev, bar) & IORESOURCE_MEM))
> +			continue;
> +		cnt++;
> +	}
> +
> +	ap = alloc_apertures(cnt);
>  	if (!ap)
>  		return -ENOMEM;
>  
> -	ap->ranges[0].base = pci_resource_start(pdev, res_id);
> -	ap->ranges[0].size = pci_resource_len(pdev, res_id);
> +	for (idx = 0, bar = 0; bar < PCI_ROM_RESOURCE; bar++) {
> +		if (!(pci_resource_flags(pdev, bar) & IORESOURCE_MEM))
> +			continue;
> +		ap->ranges[idx].base = pci_resource_start(pdev, bar);
> +		ap->ranges[idx].size = pci_resource_len(pdev, bar);
> +		pci_info(pdev, "%s: bar %d: 0x%lx -> 0x%lx\n", __func__, bar,
> +			 (unsigned long)pci_resource_start(pdev, bar),
> +			 (unsigned long)pci_resource_end(pdev, bar));
> +		idx++;
> +	}

lgtm. I think this is ok to merge as-is already ...

Reviewed-by: Daniel Vetter <daniel.vetter@...ll.ch>

Maybe if you want to be paranoid in the conversion add a check that we're
not skipping the pci bar requested in res_id. Then we could test this for
a kernel and remove the parameter later on.
-Daniel


>  #ifdef CONFIG_X86
>  	primary = pdev->resource[PCI_ROM_RESOURCE].flags &
>  					IORESOURCE_ROM_SHADOW;
> -- 
> 2.18.1
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch