mbox series

[SRU,N,0/1] nfsd: initialise nfsd_info.mutex early (LP: 2070358)

Message ID 20240712184214.251614-1-patricia.domingues@canonical.com
Headers show
Series nfsd: initialise nfsd_info.mutex early (LP: 2070358) | expand

Message

patricia.domingues@canonical.com July 12, 2024, 6:42 p.m. UTC
BugLink: https://bugs.launchpad.net/bugs/2070358

SRU Justification: 

[Impact]
 * When the sosreport command is executed, a kernel OOPS happens and the system is crashing,
  depending on the configuration (but default) the system/LPAR is rebooting.
 
[Fix]
 * e0011bca603c101f2a3c007bdb77f7006fa78fb1 e0011bca603c "nfsd: initialise nfsd_info.mutex early"
 
[Test Case]
 * Have a Ubuntu Server 24.04 LTS installation on ppc64el.
 * one option is only running sosreport on the system - and
 the crash is seen when the sosreport is starting to capture dump
 * second option (without sosreport) is:
 * CONFIG_NFSD=m (or y) must be set
 * mount nfsd if not already, using "$ mount -t nfsd nfsd /proc/fs/nfsd" command
 * The kernel oops will happen and the logs will show:
   ...
   BUG: Kernel NULL pointer dereference on read at 0x00000000
   Faulting instruction address: 0xc0000000016ff114
   Oops: Kernel access of bad area, sig: 11 [#1]
   ...
 * On a system with that kernel that incl. the above patch
   no oops will occur and the sosreport command will execute normally.
 
[Regression Potential]
* There is a certain risk of a regression, with any code modification,
  and here because the mutex handling in nfsd is modified.
 
* But the changes are pretty traceable.
 
* On top the commit is already upstream reviewed and accepted.
 
* The modifications were done by the NFSD maintainer and also tested by IBM.
 
[Other]
* The fix/commit got upstream accepted with kernel v6.10-rc7,
  hence Oracular is not affected.


NeilBrown (1):
  nfsd: initialise nfsd_info.mutex early.

 fs/nfsd/nfsctl.c | 2 ++
 fs/nfsd/nfssvc.c | 1 -
 2 files changed, 2 insertions(+), 1 deletion(-)

Comments

Kevin Becker July 12, 2024, 7:40 p.m. UTC | #1
On Fri, Jul 12, 2024 at 2:43 PM <patricia.domingues@canonical.com> wrote:
>
> BugLink: https://bugs.launchpad.net/bugs/2070358
>
> SRU Justification:
>
> [Impact]
>  * When the sosreport command is executed, a kernel OOPS happens and the system is crashing,
>   depending on the configuration (but default) the system/LPAR is rebooting.
>
> [Fix]
>  * e0011bca603c101f2a3c007bdb77f7006fa78fb1 e0011bca603c "nfsd: initialise nfsd_info.mutex early"
>
> [Test Case]
>  * Have a Ubuntu Server 24.04 LTS installation on ppc64el.
>  * one option is only running sosreport on the system - and
>  the crash is seen when the sosreport is starting to capture dump
>  * second option (without sosreport) is:
>  * CONFIG_NFSD=m (or y) must be set
>  * mount nfsd if not already, using "$ mount -t nfsd nfsd /proc/fs/nfsd" command
>  * The kernel oops will happen and the logs will show:
>    ...
>    BUG: Kernel NULL pointer dereference on read at 0x00000000
>    Faulting instruction address: 0xc0000000016ff114
>    Oops: Kernel access of bad area, sig: 11 [#1]
>    ...
>  * On a system with that kernel that incl. the above patch
>    no oops will occur and the sosreport command will execute normally.
>
> [Regression Potential]
> * There is a certain risk of a regression, with any code modification,
>   and here because the mutex handling in nfsd is modified.
>
> * But the changes are pretty traceable.
>
> * On top the commit is already upstream reviewed and accepted.
>
> * The modifications were done by the NFSD maintainer and also tested by IBM.
>
> [Other]
> * The fix/commit got upstream accepted with kernel v6.10-rc7,
>   hence Oracular is not affected.
>
>
> NeilBrown (1):
>   nfsd: initialise nfsd_info.mutex early.
>
>  fs/nfsd/nfsctl.c | 2 ++
>  fs/nfsd/nfssvc.c | 1 -
>  2 files changed, 2 insertions(+), 1 deletion(-)
>
> --

Acked-by: Kevin Becker <kevin.becker@canonical.com>
Manuel Diewald July 15, 2024, 12:12 p.m. UTC | #2
On Fri, Jul 12, 2024 at 03:42:13PM -0300, patricia.domingues@canonical.com wrote:
> BugLink: https://bugs.launchpad.net/bugs/2070358
> 
> SRU Justification: 
> 
> [Impact]
>  * When the sosreport command is executed, a kernel OOPS happens and the system is crashing,
>   depending on the configuration (but default) the system/LPAR is rebooting.
>  
> [Fix]
>  * e0011bca603c101f2a3c007bdb77f7006fa78fb1 e0011bca603c "nfsd: initialise nfsd_info.mutex early"
>  
> [Test Case]
>  * Have a Ubuntu Server 24.04 LTS installation on ppc64el.
>  * one option is only running sosreport on the system - and
>  the crash is seen when the sosreport is starting to capture dump
>  * second option (without sosreport) is:
>  * CONFIG_NFSD=m (or y) must be set
>  * mount nfsd if not already, using "$ mount -t nfsd nfsd /proc/fs/nfsd" command
>  * The kernel oops will happen and the logs will show:
>    ...
>    BUG: Kernel NULL pointer dereference on read at 0x00000000
>    Faulting instruction address: 0xc0000000016ff114
>    Oops: Kernel access of bad area, sig: 11 [#1]
>    ...
>  * On a system with that kernel that incl. the above patch
>    no oops will occur and the sosreport command will execute normally.
>  
> [Regression Potential]
> * There is a certain risk of a regression, with any code modification,
>   and here because the mutex handling in nfsd is modified.
>  
> * But the changes are pretty traceable.
>  
> * On top the commit is already upstream reviewed and accepted.
>  
> * The modifications were done by the NFSD maintainer and also tested by IBM.
>  
> [Other]
> * The fix/commit got upstream accepted with kernel v6.10-rc7,
>   hence Oracular is not affected.
> 
> 
> NeilBrown (1):
>   nfsd: initialise nfsd_info.mutex early.
> 
>  fs/nfsd/nfsctl.c | 2 ++
>  fs/nfsd/nfssvc.c | 1 -
>  2 files changed, 2 insertions(+), 1 deletion(-)
> 
> -- 
> 2.25.1
> 
> -- 
> kernel-team mailing list
> kernel-team@lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/kernel-team

Acked-by: Manuel Diewald <manuel.diewald@canonical.com>
Stefan Bader July 19, 2024, 9:27 a.m. UTC | #3
On 12.07.24 20:42, patricia.domingues@canonical.com wrote:
> BugLink: https://bugs.launchpad.net/bugs/2070358
> 
> SRU Justification:
> 
> [Impact]
>   * When the sosreport command is executed, a kernel OOPS happens and the system is crashing,
>    depending on the configuration (but default) the system/LPAR is rebooting.
>   
> [Fix]
>   * e0011bca603c101f2a3c007bdb77f7006fa78fb1 e0011bca603c "nfsd: initialise nfsd_info.mutex early"
>   
> [Test Case]
>   * Have a Ubuntu Server 24.04 LTS installation on ppc64el.
>   * one option is only running sosreport on the system - and
>   the crash is seen when the sosreport is starting to capture dump
>   * second option (without sosreport) is:
>   * CONFIG_NFSD=m (or y) must be set
>   * mount nfsd if not already, using "$ mount -t nfsd nfsd /proc/fs/nfsd" command
>   * The kernel oops will happen and the logs will show:
>     ...
>     BUG: Kernel NULL pointer dereference on read at 0x00000000
>     Faulting instruction address: 0xc0000000016ff114
>     Oops: Kernel access of bad area, sig: 11 [#1]
>     ...
>   * On a system with that kernel that incl. the above patch
>     no oops will occur and the sosreport command will execute normally.
>   
> [Regression Potential]
> * There is a certain risk of a regression, with any code modification,
>    and here because the mutex handling in nfsd is modified.
>   
> * But the changes are pretty traceable.
>   
> * On top the commit is already upstream reviewed and accepted.
>   
> * The modifications were done by the NFSD maintainer and also tested by IBM.
>   
> [Other]
> * The fix/commit got upstream accepted with kernel v6.10-rc7,
>    hence Oracular is not affected.
> 
> 
> NeilBrown (1):
>    nfsd: initialise nfsd_info.mutex early.
> 
>   fs/nfsd/nfsctl.c | 2 ++
>   fs/nfsd/nfssvc.c | 1 -
>   2 files changed, 2 insertions(+), 1 deletion(-)
> 

Applied to noble:linux/master-next. Thanks.

-Stefan