Messages in this thread | | | Date | Wed, 20 Aug 2014 20:57:55 +0200 | From | Borislav Petkov <> | Subject | Re: amd_mce.c redundant if check? |
| |
On Wed, Aug 20, 2014 at 11:18:21AM -0600, Adam Duskett wrote: > I have recently come upon this section of code in > arch/x86/kernel/cpu/mcheck/mce_amd.c that seems to be a redundant > unnecessary if check. > > > From line 170 - 176: > > if (tr->set_lvt_off) { > if (lvt_off_valid(tr->b, tr->lvt_off, lo, hi)) { > /* set new lvt offset */ > hi &= ~MASK_LVTOFF_HI; > hi |= tr->lvt_off << 20; > } > } > > > This seems like it's not actually doing anything because it's setting > the same value that the bit-field already has to itself.
Are you sure?
Just try to find out what MASK_LVTOFF_HI is and then do the operations with a pen and paper to check whether something else actually happens.
-- Regards/Gruss, Boris.
Sent from a fat crate under my desk. Formatting is fine. --
| |