> On an FSGSBASE-enabled system, I think we need to provide deterministic,
> documented, tested behavior. I can think of three plausible choices:
> 1a. modify_ldt() immediately updates FSBASE and GSBASE all threads that
> reference the modified selector.
> 1b. modify_ldt() immediatley updates FSBASE and GSBASE on all threads that
> reference the LDT.
> 2. modify_ldt() leaves FSBASE and GSBASE alone on all threads.
> (2) is trivial to implement, whereas (1a) and (1b) are a bit nasty to
> implement when FSGSBASE is on.
> The tricky bit is that 32-bit kernels can't do (2), so, if we want
> modify_ldt() to behave the same on 32-bit and 64-bit kernels, we're stuck
> with (1).
While implementing (1) is still unclear for context switch, here is one idea
for (1b):
- thread struct has new entry for ldt pointer that last seen
- modify_ldt happens
- ldtr upated for active threads via IPI
- for inactive threads being scheduled in, ldtr updated before __switch_to
- in __switch_to, read ldtr by sldt and compare the new ldt pointer
sldt is ucode that likely takes only a couple cycles
- mostly matched given modify_ldt is rare
- unmatched, don't write gsbase if gs indicating LDT
> (I think we can implement (2) with acceptable performance on 64-bit
> non-FSGSBASE kernels if we wanted to.)
Nonetheless, with Andy's argument for (1), (2) might be straightforward
assuming that user code already followed the legacy around modify_ldt in 64bit.
