[patch V4 36/37] x86/smpboot: Support parallel startup of secondary CPUs

Jeffrey Hugo quic_jhugo at quicinc.com
Fri May 19 10:44:58 PDT 2023


On 5/19/2023 10:57 AM, Andrew Cooper wrote:
> On 19/05/2023 5:28 pm, Jeffrey Hugo wrote:
>>    DESCEND objtool
>>    INSTALL libsubcmd_headers
>>    CALL    scripts/checksyscalls.sh
>>    AS      arch/x86/kernel/head_64.o
>> arch/x86/kernel/head_64.S: Assembler messages:
>> arch/x86/kernel/head_64.S:261: Error: missing ')'
>> arch/x86/kernel/head_64.S:261: Error: junk `UL<<10)' after expression
>>    CC      arch/x86/kernel/head64.o
>>    CC      arch/x86/kernel/ebda.o
>>    CC      arch/x86/kernel/platform-quirks.o
>> scripts/Makefile.build:374: recipe for target
>> 'arch/x86/kernel/head_64.o' failed
>> make[3]: *** [arch/x86/kernel/head_64.o] Error 1
>> make[3]: *** Waiting for unfinished jobs....
>> scripts/Makefile.build:494: recipe for target 'arch/x86/kernel' failed
>> make[2]: *** [arch/x86/kernel] Error 2
>> scripts/Makefile.build:494: recipe for target 'arch/x86' failed
>> make[1]: *** [arch/x86] Error 2
>> make[1]: *** Waiting for unfinished jobs....
>> Makefile:2026: recipe for target '.' failed
>> make: *** [.] Error 2
>>
>> This is with GCC 5.4.0, if it matters.
>>
>> Reverting this change allows the build to move forward, although I
>> also need to revert "x86/smpboot/64: Implement
>> arch_cpuhp_init_parallel_bringup() and enable it" for the build to
>> fully succeed.
>>
>> I'm not familiar with this code, and nothing obvious stands out to me.
>> What can I do to help root cause this?
> 
> Can you try:
> 
> -#define XAPIC_ENABLE    (1UL << 11)
> -#define X2APIC_ENABLE    (1UL << 10)
> +#define XAPIC_ENABLE    BIT(11)
> +#define X2APIC_ENABLE    BIT(10)
> 
> The UL suffix isn't understood by older binutils, and this patch adds
> the first use of these constants in assembly.

Ah, makes sense.

Your suggested change works for me.  No more compile error.

I assume you will be following up with a patch to address this.  Feel 
free to add the following tags as you see fit:

Reported-by: Jeffrey Hugo <quic_jhugo at quicinc.com>
Tested-by: Jeffrey Hugo <quic_jhugo at quicinc.com>

-Jeff



More information about the linux-riscv mailing list