linux-kernel - Re: [PATCH] arm64: Don't call NULL in do_compat_alignment

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <c1dc28a9-7a36-4303-a8eb-0e227d866c37@arm.com>
Date: Tue, 1 Apr 2025 13:17:53 +0530
From: Anshuman Khandual <anshuman.khandual@....com>
To: Angelos Oikonomopoulos <angelos@...lia.com>,
 linux-arm-kernel@...ts.infradead.org
Cc: catalin.marinas@....com, will@...nel.org, linux-kernel@...r.kernel.org,
 kernel-dev@...lia.com
Subject: Re: [PATCH] arm64: Don't call NULL in do_compat_alignment_fixup



On 4/1/25 12:28, Angelos Oikonomopoulos wrote:
> On Tue Apr 1, 2025 at 8:05 AM CEST, Anshuman Khandual wrote:
>> On 3/31/25 14:24, Angelos Oikonomopoulos wrote:
>>> do_alignment_t32_to_handler only fixes up alignment faults for specific
>>> instructions; it returns NULL otherwise. When that's the case, signal to
>>> the caller that it needs to proceed with the regular alignment fault
>>> handling (i.e. SIGBUS).
>>>
>>> Signed-off-by: Angelos Oikonomopoulos <angelos@...lia.com>
>>> ---
>>>  arch/arm64/kernel/compat_alignment.c | 2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/arch/arm64/kernel/compat_alignment.c b/arch/arm64/kernel/compat_alignment.c
>>> index deff21bfa680..b68e1d328d4c 100644
>>> --- a/arch/arm64/kernel/compat_alignment.c
>>> +++ b/arch/arm64/kernel/compat_alignment.c
>>> @@ -368,6 +368,8 @@ int do_compat_alignment_fixup(unsigned long addr, struct pt_regs *regs)
>>>  		return 1;
>>>  	}
>>>  
>>> +	if (!handler)
>>> +		return 1;
>>
>> do_alignment_t32_to_handler() could return NULL, returning 1 seems to be
>> the right thing to do here and consistent. Otherwise does this cause a
>> kernel crash during subsequent call into handler() ?
> 
> Yes. We call a NULL pointer so we Oops.

Then the commit message should have the kernel Oops splash dump and also
might need to have Fixes: and CC: stable tags etc ?

Also wondering if handler return value should be checked inside the switch
block just after do_alignment_t32_to_handler() assignment.

	handler = do_alignment_t32_to_handler()
	if (!handler)
		return 1