Re: [PATCH v1 1/1] x86/fred/signal: Prevent single-step upon ERETU completion

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 22/05/2025 6:33 pm, Xin Li wrote:
> On 5/22/2025 10:20 AM, kernel test robot wrote:
>> Hi Xin,
>>
>> kernel test robot noticed the following build errors:
>>
>> [auto build test ERROR on 6a7c3c2606105a41dde81002c0037420bc1ddf00]
>>
>> url:   
>> https://github.com/intel-lab-lkp/linux/commits/Xin-Li-Intel/x86-fred-signal-Prevent-single-step-upon-ERETU-completion/20250522-140954
>> base:   6a7c3c2606105a41dde81002c0037420bc1ddf00
>> patch link:   
>> https://lore.kernel.org/r/20250522060549.2882444-1-xin%40zytor.com
>> patch subject: [PATCH v1 1/1] x86/fred/signal: Prevent single-step
>> upon ERETU completion
>> config: i386-buildonly-randconfig-003-20250522
>> (https://download.01.org/0day-ci/archive/20250523/202505230141.4YBHhrPI-lkp@intel.com/config)
>> compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
>> reproduce (this is a W=1 build):
>> (https://download.01.org/0day-ci/archive/20250523/202505230141.4YBHhrPI-lkp@intel.com/reproduce)
>>
>> If you fix the issue in a separate patch/commit (i.e. not just a new
>> version of
>> the same patch/commit), kindly add following tags
>> | Reported-by: kernel test robot <lkp@intel.com>
>> | Closes:
>> https://lore.kernel.org/oe-kbuild-all/202505230141.4YBHhrPI-lkp@intel.com/
>>
>> All errors (new ones prefixed by >>):
>>
>>     In file included from arch/x86/kernel/signal_32.c:32:
>>     arch/x86/include/asm/sighandling.h: In function
>> 'prevent_single_step_upon_eretu':
>>>> arch/x86/include/asm/sighandling.h:44:21: error: 'struct pt_regs'
>>>> has no member named 'fred_ss'
>>        44 |                 regs->fred_ss.swevent = 0;
>>           |                     ^~
>>
>
> Hmm, this statement is under IS_ENABLED(CONFIG_X86_FRED), which should
> be a compile time FALSE with i386.  Why it is still being compiled?

Because what the compiler is seeing is:

if (0 && ...)
    regs->fred_ss.swevent = 0;

and the bad field name is found at parse time, while the whole
expression is only discarded during optimisation.

The one thing you can't IS_ENABLED() around is conditional fields.  That
needs to be full #ifdef.

~Andrew