Re: [PATCH v3 1/5] arm64: vdso32: Introduce COMPAT_CC_IS_GCC

From: Vincenzo Frascino <vincenzo.frascino@arm.com>
To: Will Deacon <will@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, ard.biesheuvel@linaro.org,
	ndesaulniers@google.com, tglx@linutronix.de
Subject: Re: [PATCH v3 1/5] arm64: vdso32: Introduce COMPAT_CC_IS_GCC
Date: Tue, 1 Oct 2019 15:37:49 +0100	[thread overview]
Message-ID: <7558914c-fc2d-d05a-ccbe-76ef451670ae@arm.com> (raw)
In-Reply-To: <20191001142038.ptwyfbesfrz3kkoz@willie-the-truck>

On 10/1/19 3:20 PM, Will Deacon wrote:
> On Tue, Oct 01, 2019 at 03:20:35PM +0100, Vincenzo Frascino wrote:
>> On 10/1/19 2:27 PM, Catalin Marinas wrote:
>>> On Tue, Oct 01, 2019 at 02:14:23PM +0100, Will Deacon wrote:
>>>> On Thu, Sep 26, 2019 at 10:43:38PM +0100, Vincenzo Frascino wrote:
>>>>> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
>>>>> index 37c610963eee..0e5beb928af5 100644
>>>>> --- a/arch/arm64/Kconfig
>>>>> +++ b/arch/arm64/Kconfig
>>>>> @@ -110,7 +110,7 @@ config ARM64
>>>>>  	select GENERIC_STRNLEN_USER
>>>>>  	select GENERIC_TIME_VSYSCALL
>>>>>  	select GENERIC_GETTIMEOFDAY
>>>>> -	select GENERIC_COMPAT_VDSO if (!CPU_BIG_ENDIAN && COMPAT)
>>>>> +	select GENERIC_COMPAT_VDSO if (!CPU_BIG_ENDIAN && COMPAT && COMPATCC_IS_ARM_GCC)
>>>>>  	select HANDLE_DOMAIN_IRQ
>>>>>  	select HARDIRQS_SW_RESEND
>>>>>  	select HAVE_PCI
>>>>> @@ -313,6 +313,9 @@ config KASAN_SHADOW_OFFSET
>>>>>  	default 0xeffffff900000000 if ARM64_VA_BITS_36 && KASAN_SW_TAGS
>>>>>  	default 0xffffffffffffffff
>>>>>  
>>>>> +config COMPATCC_IS_ARM_GCC
>>>>> +	def_bool $(success,$(COMPATCC) --version | head -n 1 | grep -q "arm-.*-gcc")
>>>>
>>>> I've seen toolchains where the first part of the tuple is "armv7-", so they
>>>> won't get detected here. However, do we really need to detect this? If
>>>> somebody passes a duff compiler, then the build will fail in the same way as
>>>> if they passed it to CROSS_COMPILE=.
>>>
>>> Not sure what happens if we pass an aarch64 compiler. Can we end up with
>>> a 64-bit compat vDSO?
>>>
>>
>> I agree with Catalin here. The problem is not only when you pass and aarch64
>> toolchain but even an x86 and so on.
> 
> I disagree. What happens if you do:
> 
> $ make ARCH=arm64 CROSS_COMPILE=x86_64-linux-gnu-
> 
> on your x86 box?
>

The kernel compilation breaks as follows:

x86_64-linux-gnu-gcc: error: unrecognized command line option ‘-mlittle-endian’;
did you mean ‘-fconvert=little-endian’?
/data1/Projects/LinuxKernel/linux/scripts/Makefile.build:265: recipe for target
'scripts/mod/empty.o' failed
make[2]: *** [scripts/mod/empty.o] Error 1
/data1/Projects/LinuxKernel/linux/Makefile:1128: recipe for target 'prepare0' failed
make[1]: *** [prepare0] Error 2
make[1]: Leaving directory '/data1/Projects/LinuxKernel/linux-out'
Makefile:179: recipe for target 'sub-make' failed
make: *** [sub-make] Error 2

Similar issue in the compat vdso library compilation if I do (without the check):

$ make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu-
CROSS_COMPILE_COMPAT=x86_64-linux-gnu-

With this check the compilation completes correctly but the compat vdso does not
get built (unless my environment is playing me tricks ;) ).

>> If the problem is related to armv7- we can change the rule as "arm.*-gcc" which
>> should detect them as well. Do you know what is the triple that an armv7-
>> toolchain prints?
> 
> 'fraid not, since I don't have one to hand. I think you'd end up matching
> arm*-gcc, which is pretty broad.
> 

If we all agree I can extend this rule then.

> Will
> 

-- 
Regards,
Vincenzo