linux-kernel - Re: [PATCH v3 4/5] powerpc/fault: Avoid heavy search_exception

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: <d66f9706-9e36-5b92-5a87-90ebd05587e9@csgroup.eu>
Date:   Tue, 8 Dec 2020 15:26:14 +0100
From:   Christophe Leroy <christophe.leroy@...roup.eu>
To:     "Aneesh Kumar K.V" <aneesh.kumar@...ux.ibm.com>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Michael Ellerman <mpe@...erman.id.au>, npiggin@...il.com
Cc:     linux-kernel@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH v3 4/5] powerpc/fault: Avoid heavy
 search_exception_tables() verification



Le 08/12/2020 à 14:00, Aneesh Kumar K.V a écrit :
> On 12/8/20 2:07 PM, Christophe Leroy wrote:
>> search_exception_tables() is an heavy operation, we have to avoid it.
>> When KUAP is selected, we'll know the fault has been blocked by KUAP.
>> Otherwise, it behaves just as if the address was already in the TLBs
>> and no fault was generated.
>>
>> Signed-off-by: Christophe Leroy <christophe.leroy@...roup.eu>
>> Reviewed-by: Nicholas Piggin <npiggin@...il.com>
>> ---
>> v3: rebased
>> v2: Squashed with the preceeding patch which was re-ordering tests that get removed in this patch.
>> ---
>>   arch/powerpc/mm/fault.c | 23 +++++++----------------
>>   1 file changed, 7 insertions(+), 16 deletions(-)
>>
>> diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
>> index 3fcd34c28e10..1770b41e4730 100644
>> --- a/arch/powerpc/mm/fault.c
>> +++ b/arch/powerpc/mm/fault.c
>> @@ -210,28 +210,19 @@ static bool bad_kernel_fault(struct pt_regs *regs, unsigned long error_code,
>>           return true;
>>       }
>> -    if (!is_exec && address < TASK_SIZE && (error_code & (DSISR_PROTFAULT | DSISR_KEYFAULT)) &&
>> -        !search_exception_tables(regs->nip)) {
>> -        pr_crit_ratelimited("Kernel attempted to access user page (%lx) - exploit attempt? (uid: 
>> %d)\n",
>> -                    address,
>> -                    from_kuid(&init_user_ns, current_uid()));
>> -    }
>> -
>>       // Kernel fault on kernel address is bad
>>       if (address >= TASK_SIZE)
>>           return true;
>> -    // Fault on user outside of certain regions (eg. copy_tofrom_user()) is bad
>> -    if (!search_exception_tables(regs->nip))
>> -        return true;
>> -
>> -    // Read/write fault in a valid region (the exception table search passed
>> -    // above), but blocked by KUAP is bad, it can never succeed.
>> -    if (bad_kuap_fault(regs, address, is_write))
>> +    // Read/write fault blocked by KUAP is bad, it can never succeed.
>> +    if (bad_kuap_fault(regs, address, is_write)) {
>> +        pr_crit_ratelimited("Kernel attempted to %s user page (%lx) - exploit attempt? (uid: %d)\n",
>> +                    is_write ? "write" : "read", address,
>> +                    from_kuid(&init_user_ns, current_uid()));
>>           return true;
>> +    }
> 
> 
> Should we update bad_kuap_fault to check for !is_kernel_addr() and error_code & (DSISIR_PROT_FAULT | 
> DSISIR_KEYFAULT). I am wondering whether we can take another fault w.r.t kernel address/user address 
> and end up reporting that as KUAP fault?

Just before this we do:

	if (address >= TASK_SIZE)
		return true;

About the error code, I don't know. Can we take a fault that is not a DSISR_PROT_FAULT |
  DSISR_KEYFAULT and that is not a KUAP fault ?

Previously (before this patch), the error code was taken into account for the call to 
search_exception_tables(), but has never been for the bad_kuap_fault().

> 
>> -    // What's left? Kernel fault on user in well defined regions (extable
>> -    // matched), and allowed by KUAP in the faulting context.
>> +    // What's left? Kernel fault on user and allowed by KUAP in the faulting context.
>>       return false;
>>   }
>>
> 
> 
> -aneesh

Christophe