[v3,0/4] EDAC/mc: Fixes for mci device removal
mbox series

Message ID 20200212120340.4764-1-rrichter@marvell.com
Headers show
Series
  • EDAC/mc: Fixes for mci device removal
Related show

Message

Robert Richter Feb. 12, 2020, 12:03 p.m. UTC
This series is a split of

 [v2] EDAC/mc: Fix use-after-free and memleaks during device removal
 https://lore.kernel.org/patchwork/patch/1190002/

in smaller patches to better handle stable backports.

Patch #1 helps to ease backports of patch #2 to 5.3 and earlier
kernels, see patch descriptions of both for details.

Patch #2 is the smallest possible fix to address use-after-free and
memleak issues aimed for backports.

Patch #3 is another small patch that is split off for stable. This
fixes code already commited to stable trees.

Patch #4 is the all remaining code containing the rework of the mci
device removal. Resulting code is the same as in v2 except for the
changes outlined for v3:

v3:
 * split patch into smaller pieces to ease backports,
 * removed edac_remove_sysfs_mci_device() in edac_mc_free(), at this
   point the mci device is always unregistered

Robert Richter (4):
  Revert parts of "EDAC/mc_sysfs: Make debug messages consistent"
  EDAC/mc: Fix use-after-free and memleaks during device removal
  EDAC/sysfs: Remove csrow objects on errors
  EDAC/mc: Change mci device removal to use put_device()

 drivers/edac/edac_mc.c       |  20 +++----
 drivers/edac/edac_mc_sysfs.c | 100 +++++++++++++----------------------
 drivers/edac/edac_module.h   |   1 -
 3 files changed, 48 insertions(+), 73 deletions(-)

Comments

Borislav Petkov Feb. 13, 2020, 11:05 a.m. UTC | #1
On Wed, Feb 12, 2020 at 01:03:36PM +0100, Robert Richter wrote:
> This series is a split of
> 
>  [v2] EDAC/mc: Fix use-after-free and memleaks during device removal
>  https://lore.kernel.org/patchwork/patch/1190002/
> 
> in smaller patches to better handle stable backports.
> 
> Patch #1 helps to ease backports of patch #2 to 5.3 and earlier
> kernels, see patch descriptions of both for details.
> 
> Patch #2 is the smallest possible fix to address use-after-free and
> memleak issues aimed for backports.
> 
> Patch #3 is another small patch that is split off for stable. This
> fixes code already commited to stable trees.
> 
> Patch #4 is the all remaining code containing the rework of the mci
> device removal. Resulting code is the same as in v2 except for the
> changes outlined for v3:
> 
> v3:
>  * split patch into smaller pieces to ease backports,
>  * removed edac_remove_sysfs_mci_device() in edac_mc_free(), at this
>    point the mci device is always unregistered
> 
> Robert Richter (4):
>   Revert parts of "EDAC/mc_sysfs: Make debug messages consistent"
>   EDAC/mc: Fix use-after-free and memleaks during device removal
>   EDAC/sysfs: Remove csrow objects on errors
>   EDAC/mc: Change mci device removal to use put_device()
> 
>  drivers/edac/edac_mc.c       |  20 +++----
>  drivers/edac/edac_mc_sysfs.c | 100 +++++++++++++----------------------
>  drivers/edac/edac_module.h   |   1 -
>  3 files changed, 48 insertions(+), 73 deletions(-)

Thanks, first three (1+2 squashed) pushed here:

https://git.kernel.org/pub/scm/linux/kernel/git/ras/ras.git/log/?h=edac-urgent

@John: if you could run them one more time on your machines, that would
be great!

Thx.
John Garry Feb. 13, 2020, 11:10 a.m. UTC | #2
On 13/02/2020 11:05, Borislav Petkov wrote:
>> Robert Richter (4):
>>    Revert parts of "EDAC/mc_sysfs: Make debug messages consistent"
>>    EDAC/mc: Fix use-after-free and memleaks during device removal
>>    EDAC/sysfs: Remove csrow objects on errors
>>    EDAC/mc: Change mci device removal to use put_device()
>>
>>   drivers/edac/edac_mc.c       |  20 +++----
>>   drivers/edac/edac_mc_sysfs.c | 100 +++++++++++++----------------------
>>   drivers/edac/edac_module.h   |   1 -
>>   3 files changed, 48 insertions(+), 73 deletions(-)
> Thanks, first three (1+2 squashed) pushed here:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/ras/ras.git/log/?h=edac-urgent
> 
> @John: if you could run them one more time on your machines, that would
> be great!

ok, give me a little while and I'll test that branch

Cheers

> 
> Thx.
> 
> -- Regards/Gruss, Boris. 
> https://people.kernel.org/tglx/notes-about-netiquette .
John Garry Feb. 13, 2020, 12:08 p.m. UTC | #3
On 13/02/2020 11:10, John Garry wrote:
> On 13/02/2020 11:05, Borislav Petkov wrote:
>>> Robert Richter (4):
>>>    Revert parts of "EDAC/mc_sysfs: Make debug messages consistent"
>>>    EDAC/mc: Fix use-after-free and memleaks during device removal
>>>    EDAC/sysfs: Remove csrow objects on errors
>>>    EDAC/mc: Change mci device removal to use put_device()
>>>
>>>   drivers/edac/edac_mc.c       |  20 +++----
>>>   drivers/edac/edac_mc_sysfs.c | 100 +++++++++++++----------------------
>>>   drivers/edac/edac_module.h   |   1 -
>>>   3 files changed, 48 insertions(+), 73 deletions(-)
>> Thanks, first three (1+2 squashed) pushed here:
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/ras/ras.git/log/?h=edac-urgent 
>>
>>
>> @John: if you could run them one more time on your machines, that would
>> be great!
> 
> ok, give me a little while and I'll test that branch
> 

Yeah, it looks ok - I'm just booting and running a kmemleak scan

Cheers

> 
>>
>> Thx.
>>
>> -- Regards/Gruss, Boris. 
>> https://people.kernel.org/tglx/notes-about-netiquette .
> 
> .
Borislav Petkov Feb. 13, 2020, 12:28 p.m. UTC | #4
On Thu, Feb 13, 2020 at 12:08:59PM +0000, John Garry wrote:
> Yeah, it looks ok - I'm just booting and running a kmemleak scan

Thanks, I'll add your Tested-by: