Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: GCC bug ? Re: [PATCH v2 10/10] powerpc/32s: Implement Kernel
 Userspace Access Protection
To:     Michael Ellerman <mpe@ellerman.id.au>,
        Benjamin Herrenschmidt <benh@kernel.crashing.org>,
        Paul Mackerras <paulus@samba.org>, ruscur@russell.cc,
        Segher Boessenkool <segher@kernel.crashing.org>
Cc:     linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org
References: <cover.1552292207.git.christophe.leroy@c-s.fr>
 <a2847248a92cb1641b1740fa121c5a30593ae662.1552292207.git.christophe.leroy@c-s.fr>
 <87ftqfu7j1.fsf@concordia.ellerman.id.au>
From:   Christophe Leroy <christophe.leroy@c-s.fr>
Message-ID: <a008a182-f1db-073c-7d38-27bfd1fd8676@c-s.fr>
Date:   Tue, 21 Jan 2020 17:22:32 +0000
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.7.0
MIME-Version: 1.0
In-Reply-To: <87ftqfu7j1.fsf@concordia.ellerman.id.au>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On 04/18/2019 06:55 AM, Michael Ellerman wrote:
> Christophe Leroy <christophe.leroy@c-s.fr> writes:
>> diff --git a/arch/powerpc/include/asm/book3s/32/kup.h b/arch/powerpc/include/asm/book3s/32/kup.h
>> index 5f97c742ca71..b3560b2de435 100644
>> --- a/arch/powerpc/include/asm/book3s/32/kup.h
>> +++ b/arch/powerpc/include/asm/book3s/32/kup.h
>> @@ -37,6 +37,113 @@
> ...
>> +
>> +static inline void allow_user_access(void __user *to, const void __user *from, u32 size)
>> +{
>> +	u32 addr = (__force u32)to;
>> +	u32 end = min(addr + size, TASK_SIZE);
>> +
>> +	if (!addr || addr >= TASK_SIZE || !size)
>> +		return;
>> +
>> +	current->thread.kuap = (addr & 0xf0000000) | ((((end - 1) >> 28) + 1) & 0xf);
>> +	kuap_update_sr(mfsrin(addr) & ~SR_KS, addr, end);	/* Clear Ks */
>> +}
> 
> When rebasing on my v6 I changed the above to:
> 
> static inline void allow_user_access(void __user *to, const void __user *from, u32 size)
> {
> 	u32 addr, end;
> 
> 	if (__builtin_constant_p(to) && to == NULL)
> 		return;

Look like the above doesn't work: gcc bug ?

#define NULL (void*)0

static inline int f1(void *to)
{
	if (__builtin_constant_p(to) && to == NULL)
		return 3;
	return 5;
}

int g1(void)
{
	return f1(NULL);
}

static inline int f2(int x)
{
	if (__builtin_constant_p(x) && x == 0)
		return 7;
	return 9;
}

int g2(void)
{
	return f2(0);
}


toto.o:     file format elf32-powerpc


Disassembly of section .text:

00000000 <g1>:
    0:	38 60 00 05 	li      r3,5
    4:	4e 80 00 20 	blr

00000008 <g2>:
    8:	38 60 00 07 	li      r3,7
    c:	4e 80 00 20 	blr


It works for the int const, but not for the pointer const:

g1() should return 3, not 5. GCC bug ?

Christophe