linux-kernel - Re: [PATCH 2/2] amd64-agp: Probe unknown AGP devices the right way

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] [day] [month] [year] [list]

Message-ID: <1273972177.2564.68.camel@localhost>
Date:	Sun, 16 May 2010 02:09:37 +0100
From:	Ben Hutchings <ben@...adent.org.uk>
To:	David Airlie <airlied@...ux.ie>
Cc:	FUJITA Tomonori <fujita.tomonori@....ntt.co.jp>,
	Greg Kroah-Hartman <gregkh@...e.de>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] amd64-agp: Probe unknown AGP devices the right way

On Wed, 2010-03-24 at 03:36 +0000, Ben Hutchings wrote:
> The current initialisation code probes 'unsupported' AGP devices
> simply by calling its own probe function.  It does not lock these
> devices or even check whether another driver is already bound to
> them.
> 
> We must use the device core to manage this.  So if the specific
> device id table didn't match anything and agp_try_unsupported=1,
> switch the device id table and call driver_attach() again.

Can you please apply or nak this?

Ben.

> Signed-off-by: Ben Hutchings <ben@...adent.org.uk>
> ---
> This is untested as I have no suitable hardware.
> 
> Greg, please confirm that the use of driver_attach() is sane.
> 
> Ben.
> 
>  drivers/char/agp/amd64-agp.c |   27 +++++++++++++++------------
>  1 files changed, 15 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c
> index fd50ead..93f56d6 100644
> --- a/drivers/char/agp/amd64-agp.c
> +++ b/drivers/char/agp/amd64-agp.c
> @@ -499,6 +499,10 @@ static int __devinit agp_amd64_probe(struct pci_dev *pdev,
>  	u8 cap_ptr;
>  	int err;
>  
> +	/* The Highlander principle */
> +	if (agp_bridges_found)
> +		return -ENODEV;
> +
>  	cap_ptr = pci_find_capability(pdev, PCI_CAP_ID_AGP);
>  	if (!cap_ptr)
>  		return -ENODEV;
> @@ -562,6 +566,8 @@ static void __devexit agp_amd64_remove(struct pci_dev *pdev)
>  			   amd64_aperture_sizes[bridge->aperture_size_idx].size);
>  	agp_remove_bridge(bridge);
>  	agp_put_bridge(bridge);
> +
> +	agp_bridges_found--;
>  }
>  
>  #ifdef CONFIG_PM
> @@ -709,6 +715,11 @@ static struct pci_device_id agp_amd64_pci_table[] = {
>  
>  MODULE_DEVICE_TABLE(pci, agp_amd64_pci_table);
>  
> +static DEFINE_PCI_DEVICE_TABLE(agp_amd64_pci_promisc_table) = {
> +	{ PCI_DEVICE_CLASS(0, 0) },
> +	{ }
> +};
> +
>  static struct pci_driver agp_amd64_pci_driver = {
>  	.name		= "agpgart-amd64",
>  	.id_table	= agp_amd64_pci_table,
> @@ -734,7 +745,6 @@ int __init agp_amd64_init(void)
>  		return err;
>  
>  	if (agp_bridges_found == 0) {
> -		struct pci_dev *dev;
>  		if (!agp_try_unsupported && !agp_try_unsupported_boot) {
>  			printk(KERN_INFO PFX "No supported AGP bridge found.\n");
>  #ifdef MODULE
> @@ -750,17 +760,10 @@ int __init agp_amd64_init(void)
>  			return -ENODEV;
>  
>  		/* Look for any AGP bridge */
> -		dev = NULL;
> -		err = -ENODEV;
> -		for_each_pci_dev(dev) {
> -			if (!pci_find_capability(dev, PCI_CAP_ID_AGP))
> -				continue;
> -			/* Only one bridge supported right now */
> -			if (agp_amd64_probe(dev, NULL) == 0) {
> -				err = 0;
> -				break;
> -			}
> -		}
> +		agp_amd64_pci_driver.id_table = agp_amd64_pci_promisc_table;
> +		err = driver_attach(&agp_amd64_pci_driver.driver);
> +		if (err == 0 && agp_bridges_found == 0)
> +			err = -ENODEV;
>  	}
>  	return err;
>  }

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Download attachment "signature.asc" of type "application/pgp-signature" (829 bytes)