Re: [PATCH drivers/misc 0/4] lkdtm: Various clean ups

[Richard Weinberger]

----- UrsprÃngliche Mail -----
> Von: "Kees Cook" <keescook@xxxxxxxxxxxx>
> An: "Randy Dunlap" <rdunlap@xxxxxxxxxxxxx>
> CC: "Richard Weinberger" <richard.weinberger@xxxxxxxxx>, "richard" <richard@xxxxxx>, "Greg Kroah-Hartman"
> <gregkh@xxxxxxxxxxxxxxxxxxx>, "Prasad Sodagudi" <psodagud@xxxxxxxxxxxxxx>, "Sami Tolvanen" <samitolvanen@xxxxxxxxxx>,
> "Amit Daniel Kachhap" <amit.kachhap@xxxxxxx>, "linux-kselftest" <linux-kselftest@xxxxxxxxxxxxxxx>, "clang-built-linux"
> <clang-built-linux@xxxxxxxxxxxxxxxx>, "linux-kernel" <linux-kernel@xxxxxxxxxxxxxxx>
> Gesendet: Donnerstag, 25. Juni 2020 08:06:18
> Betreff: Re: [PATCH drivers/misc 0/4] lkdtm: Various clean ups

> On Wed, Jun 24, 2020 at 06:45:47PM -0700, Randy Dunlap wrote:
>> Looks like lkdtm/bugs.c needs to get/use arch/x86/include/asm/processor.h
>> but it actually uses arch/x86/um/asm/processor*.h, which does not have the
>> needed structs etc.
> 
> Should I just test for !UML in bugs.c? (This is all for the
> lkdtm_DOUBLE_FAULT() test.) I already do those kinds of checks for the
> lkdtm_UNSET_SMEP() test. e.g.:

Just had a look. Yes, this sounds good to me. UML has CONFIG_X86_32=y but no GDT. :-)

Thanks,
//richard