[PATCH] amd64-agp: fix crash at second module load

[PATCH] amd64-agp: fix crash at second module load

[Florian Mickler]

The module forgot to sometimes unregister some resources.

This fixes Bug #22882.

[Patch updated to 2.6.38-rc3 by Randy Dunlap.]
Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Florian Mickler <florian@mickler.org>
---
 drivers/char/agp/amd64-agp.c |    9 +++++++--
 1 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c
index 9252e85..780498d 100644
--- a/drivers/char/agp/amd64-agp.c
+++ b/drivers/char/agp/amd64-agp.c
@@ -773,18 +773,23 @@ int __init agp_amd64_init(void)
 #else
 			printk(KERN_INFO PFX "You can boot with agp=try_unsupported\n");
 #endif
+			pci_unregister_driver(&agp_amd64_pci_driver);
 			return -ENODEV;
 		}
 
 		/* First check that we have at least one AMD64 NB */
-		if (!pci_dev_present(amd_nb_misc_ids))
+		if (!pci_dev_present(amd_nb_misc_ids)) {
+			pci_unregister_driver(&agp_amd64_pci_driver);
 			return -ENODEV;
+		}
 
 		/* Look for any AGP bridge */
 		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)
+		if (err == 0 && agp_bridges_found == 0) {
+			pci_unregister_driver(&agp_amd64_pci_driver);
 			err = -ENODEV;
+		}
 	}
 	return err;
 }
-- 
1.7.4.rc3

Re: [PATCH] amd64-agp: fix crash at second module load

[Florian Mickler]

Hi Andrew,
I sent this to Dave Airlie because that is what MAINTAINERS says, but
that is probably the wrong receipient? 

Regards,
Flo

On Mon,  7 Feb 2011 23:29:31 +0100
Florian Mickler <florian@mickler.org> wrote:

> The module forgot to sometimes unregister some resources.
> 
> This fixes Bug #22882.
> 
> [Patch updated to 2.6.38-rc3 by Randy Dunlap.]
> Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
> Signed-off-by: Florian Mickler <florian@mickler.org>
> ---
>  drivers/char/agp/amd64-agp.c |    9 +++++++--
>  1 files changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c
> index 9252e85..780498d 100644
> --- a/drivers/char/agp/amd64-agp.c
> +++ b/drivers/char/agp/amd64-agp.c
> @@ -773,18 +773,23 @@ int __init agp_amd64_init(void)
>  #else
>  			printk(KERN_INFO PFX "You can boot with agp=try_unsupported\n");
>  #endif
> +			pci_unregister_driver(&agp_amd64_pci_driver);
>  			return -ENODEV;
>  		}
>  
>  		/* First check that we have at least one AMD64 NB */
> -		if (!pci_dev_present(amd_nb_misc_ids))
> +		if (!pci_dev_present(amd_nb_misc_ids)) {
> +			pci_unregister_driver(&agp_amd64_pci_driver);
>  			return -ENODEV;
> +		}
>  
>  		/* Look for any AGP bridge */
>  		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)
> +		if (err == 0 && agp_bridges_found == 0) {
> +			pci_unregister_driver(&agp_amd64_pci_driver);
>  			err = -ENODEV;
> +		}
>  	}
>  	return err;
>  }

Re: [PATCH] amd64-agp: fix crash at second module load

[Dave Airlie]

On Wed, Feb 23, 2011 at 6:26 PM, Florian Mickler <florian@mickler.org> wrote:
> Hi Andrew,
> I sent this to Dave Airlie because that is what MAINTAINERS says, but
> that is probably the wrong receipient?
>

No I just get agp blindspots, esp for the 64-bit ones since they
interact with x86.

I'll push it to Linus tomorrow.

Dave.

> Regards,
> Flo
>
> On Mon,  7 Feb 2011 23:29:31 +0100
> Florian Mickler <florian@mickler.org> wrote:
>
>> The module forgot to sometimes unregister some resources.
>>
>> This fixes Bug #22882.
>>
>> [Patch updated to 2.6.38-rc3 by Randy Dunlap.]
>> Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
>> Signed-off-by: Florian Mickler <florian@mickler.org>
>> ---
>>  drivers/char/agp/amd64-agp.c |    9 +++++++--
>>  1 files changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c
>> index 9252e85..780498d 100644
>> --- a/drivers/char/agp/amd64-agp.c
>> +++ b/drivers/char/agp/amd64-agp.c
>> @@ -773,18 +773,23 @@ int __init agp_amd64_init(void)
>>  #else
>>                       printk(KERN_INFO PFX "You can boot with agp=try_unsupported\n");
>>  #endif
>> +                     pci_unregister_driver(&agp_amd64_pci_driver);
>>                       return -ENODEV;
>>               }
>>
>>               /* First check that we have at least one AMD64 NB */
>> -             if (!pci_dev_present(amd_nb_misc_ids))
>> +             if (!pci_dev_present(amd_nb_misc_ids)) {
>> +                     pci_unregister_driver(&agp_amd64_pci_driver);
>>                       return -ENODEV;
>> +             }
>>
>>               /* Look for any AGP bridge */
>>               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)
>> +             if (err == 0 && agp_bridges_found == 0) {
>> +                     pci_unregister_driver(&agp_amd64_pci_driver);
>>                       err = -ENODEV;
>> +             }
>>       }
>>       return err;
>>  }
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: [PATCH] amd64-agp: fix crash at second module load

[Florian Mickler]

On Wed, 23 Feb 2011 18:28:21 +1000
Dave Airlie <airlied@gmail.com> wrote:

> On Wed, Feb 23, 2011 at 6:26 PM, Florian Mickler <florian@mickler.org> wrote:
> > Hi Andrew,
> > I sent this to Dave Airlie because that is what MAINTAINERS says, but
> > that is probably the wrong receipient?
> >
> 
> No I just get agp blindspots, esp for the 64-bit ones since they
> interact with x86.
> 
> I'll push it to Linus tomorrow.
> 
> Dave.

Ah ok, I wasn't shure, because there were no commits from you over
there. 

Thanks,
Flo