Re: [PATCH 1/6] symbols: add minimal self-test

[Jan Beulich <jbeulich@xxxxxxxx>]

On 13.03.2025 16:44, Andrew Cooper wrote:
> On 13/03/2025 3:39 pm, Jan Beulich wrote:
>> On 13.03.2025 16:35, Andrew Cooper wrote:
>>> On 13/03/2025 1:52 pm, Jan Beulich wrote:
>>>> ... before making changes to the involved logic.
>>>>
>>>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>>>> ---
>>>> With this FAST_SYMBOL_LOOKUP may make sense to permit enabling even
>>>> when LIVEPATCH=n. Thoughts? (In this case "symbols: centralize and re-
>>>> arrange $(all_symbols) calculation" would want pulling ahead.)
>>>>
>>>> --- a/xen/common/symbols.c
>>>> +++ b/xen/common/symbols.c
>>>> @@ -260,6 +260,41 @@ unsigned long symbols_lookup_by_name(con
>>>>      return 0;
>>>>  }
>>>>  
>>>> +#ifdef CONFIG_SELF_TESTS
>>>> +
>>>> +static void __init test_lookup(unsigned long addr, const char *expected)
>>>> +{
>>>> +    char buf[KSYM_NAME_LEN + 1];
>>>> +    const char *name, *symname;
>>>> +    unsigned long size, offs;
>>>> +
>>>> +    name = symbols_lookup(addr, &size, &offs, buf);
>>>> +    if ( !name )
>>>> +        panic("%s: address not found\n", expected);
>>>> +    if ( offs )
>>>> +        panic("%s: non-zero offset (%#lx) unexpected\n", expected, offs);
>>>> +
>>>> +    /* Cope with static symbols, where varying file names/paths may be 
>>>> used. */
>>>> +    symname = strchr(name, '#');
>>>> +    symname = symname ? symname + 1 : name;
>>>> +    if ( strcmp(symname, expected) )
>>>> +        panic("%s: unexpected symbol name: '%s'\n", expected, symname);
>>>> +
>>>> +    offs = symbols_lookup_by_name(name);
>>>> +    if ( offs != addr )
>>>> +        panic("%s: address %#lx unexpected; wanted %#lx\n",
>>>> +              expected, offs, addr);
>>>> +}
>>>> +
>>>> +static void __init __constructor test_symbols(void)
>>>> +{
>>>> +    /* Be sure to only try this for cf_check functions. */
>>> I'm very happy to see the take-up of SELF_TESTs.  Although I probably
>>> ought to tie it into a Kconfig option to make the errors non-fatal,
>>> which I've been meaning to do for a bit.
>>>
>>> One question though.  cf_check is an x86-ism, even if it leaks out into
>>> common code.
>>>
>>> I think you mean "functions emitted into the final image"?  If so, I
>>> don't think this is relevant then, because ...
>>>
>>>> +    test_lookup((unsigned long)dump_execstate, "dump_execstate");
>>>> +    test_lookup((unsigned long)test_symbols, __func__);
>>> ... taking the function address here forces it to be emitted even if it
>>> would otherwise have been inlined.
>> No, I really mean cf_check. If we took the address of a non-cf_check
>> function, the special gcc13 build's checking would trigger, aiui.
> 
> It's GCC-11 sadly.  cf_check is part of the function type, and triggers
> when a function type check would be relevant.  Just casing to an integer
> won't trigger it, I don't think.

Is there a way to double check? I'd be happy to drop that comment (and
use some other, maybe less random function), but I don't have a compiler
available that includes that patch.

Jan