[PATCH v5.10 v2] powerpc: Avoid nmi_enter/nmi_exit in real mode interrupt.

[Jinjie Ruan]

On 2024/8/13 18:47, Greg KH wrote:
> On Tue, Aug 06, 2024 at 07:16:16AM +0000, Jinjie Ruan wrote:
>> From: Mahesh Salgaonkar <mahesh at linux.ibm.com>
>>
>> [ Upstream commit 0db880fc865ffb522141ced4bfa66c12ab1fbb70 ]
>>
>> nmi_enter()/nmi_exit() touches per cpu variables which can lead to kernel
>> crash when invoked during real mode interrupt handling (e.g. early HMI/MCE
>> interrupt handler) if percpu allocation comes from vmalloc area.
>>
>> Early HMI/MCE handlers are called through DEFINE_INTERRUPT_HANDLER_NMI()
>> wrapper which invokes nmi_enter/nmi_exit calls. We don't see any issue when
>> percpu allocation is from the embedded first chunk. However with
>> CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK enabled there are chances where percpu
>> allocation can come from the vmalloc area.
>>
>> With kernel command line "percpu_alloc=page" we can force percpu allocation
>> to come from vmalloc area and can see kernel crash in machine_check_early:
>>
>> [    1.215714] NIP [c000000000e49eb4] rcu_nmi_enter+0x24/0x110
>> [    1.215717] LR [c0000000000461a0] machine_check_early+0xf0/0x2c0
>> [    1.215719] --- interrupt: 200
>> [    1.215720] [c000000fffd73180] [0000000000000000] 0x0 (unreliable)
>> [    1.215722] [c000000fffd731b0] [0000000000000000] 0x0
>> [    1.215724] [c000000fffd73210] [c000000000008364] machine_check_early_common+0x134/0x1f8
>>
>> Fix this by avoiding use of nmi_enter()/nmi_exit() in real mode if percpu
>> first chunk is not embedded.
>>
>> CVE-2024-42126
>> Cc: stable at vger.kernel.org#5.10.x
>> Cc: gregkh at linuxfoundation.org
>> Reviewed-by: Christophe Leroy <christophe.leroy at csgroup.eu>
>> Tested-by: Shirisha Ganta <shirisha at linux.ibm.com>
>> Signed-off-by: Mahesh Salgaonkar <mahesh at linux.ibm.com>
>> Signed-off-by: Michael Ellerman <mpe at ellerman.id.au>
>> Link: https://msgid.link/20240410043006.81577-1-mahesh@linux.ibm.com
>> [ Conflicts in arch/powerpc/include/asm/interrupt.h
>>   because machine_check_early() and machine_check_exception()
>>   has been refactored. ]
>> Signed-off-by: Jinjie Ruan <ruanjinjie at huawei.com>
>> ---
>> v2:
>> - Also fix for CONFIG_PPC_BOOK3S_64 not enabled.
>> - Add Upstream.
>> - Cc stable at vger.kernel.org.
> 
> You forgot a 5.15.y version, which is of course required if we were to
> take a 5.10.y version :(
> 
> Please resubmit both.

Sorry for forgetting the 5.15.y, I'll resend it sooner, thank you!

> 
> thanks,
> 
> greg k-h