On Wed, Aug 26, 2020 at 10:47 AM Thomas Gleixner tglx@linutronix.de wrote:
Andy,
On Wed, Aug 26 2020 at 09:13, Andy Lutomirski wrote:
On Wed, Aug 26, 2020 at 7:27 AM Thomas Gleixner tglx@linutronix.de wrote:
The below nasty hack cures it, but I hate it with a passion. I'll look deeper for a sane variant.
Fundamentally, the way we overload orig_ax is problematic. I have a half-written series to improve it, but my series is broken. I think it's fixable, though.
First is this patch to use some __csh bits to indicate the entry type. As far as I know, this patch is correct:
https://git.kernel.org/pub/scm/linux/kernel/git/luto/linux.git/commit/?h=x86...
Yes, that looks about right.
Then I wrote this incorrect patch:
https://git.kernel.org/pub/scm/linux/kernel/git/luto/linux.git/commit/?h=x86...
That one is wrong because the orig_ax wreckage seems to have leaked into user ABI -- user programs think that orig_ax has certain semantics on user-visible entries.
Yes, orig_ax is pretty much user ABI for a very long time.
But I think that the problem in this thread could be fixed quite nicely by the first patch, plus a new CS_ENTRY_IRQ and allocating eight bits of __csh to store the vector. Then we could read out the vector.
That works. Alternatively I can just store the vector in the irq descriptor itself. That's trivial enough and can be done completely in C independent of the stuff above.
The latter sounds quite sensible to me. It does seem vaguely ridiculous to be trying to fish the vector out of pt_regs in the APIC code.
--Andy