[llvm-dev] Altering the return address , for a function with multiple return paths

[John McCall via llvm-dev]

On 24 Jul 2019, at 18:24, Philip Reames wrote:
> On 7/24/19 7:42 AM, John McCall wrote:
>>
>> On 24 Jul 2019, at 1:46, Philip Reames wrote:
>>
>>     We've also got some cases where returning a value in a flag might
>>     be useful.  Our typical use case is we have a "rare, but not
>>     *that* rare* slowpath which sometimes needs to run after a call
>>     from a runtime function.  Our other compiler(s) - which use hand
>>     rolled assembly for all of these bits - return the "take-rare" 
>> bit
>>     in ZF, and branch on that after the call.  For our LLVM based
>>     system, we just materialize the value into $rax and branch on
>>     that.  That naive scheme has been surprisingly not bad 
>> performance
>>     wise.
>>
>>     * The "not *that* rare" part is needed to avoid having 
>> exceptional
>>     unwinding be the right answer.
>>
>>     If we were to support something like this, you'd really want to 
>> be
>>     able to define individual flags in the callee's calling 
>> convention
>>     clobber/preserve lists.  It's really common to have a helper
>>     routine which sets say ZF, but leaves others unchanged.  Or to
>>     have a function which sets ZF, clobbers OF, and preserves all
>>     others.  But if we were going to do that, we'd quickly realize
>>     that the x86 backend doesn't track individual flags at all, and
>>     thus conclude it probably wasn't worth it begin with.  :)
>>
>> I'm intrigued what functions would usefully /preserve/ flags. That's 
>> quite difficult on x86, since it means you have to spill flags if you 
>> do any basic arithmetic at all (unless you do it all with |lea|). 
>> Maybe on a target like PPC with really fleshed-out support for flags 
>> and conditionality?
>>
> My use case would be hand written assembly stubs for particular 
> fastpath operations.  These aren't compiler generated, and the hour 
> of human time to line of code ratio is quite high.  Having said that, 
> while I have practical examples where returning flags have been used, 
> I don't, off the top of my head, have a concrete example of needing to 
> preserve flags (except DF).  So, take that part more as a "if we're 
> there anyways, the ideal would be...". On a practical level, return 
> via flags with full clobber (except DF) would probably be more than 
> sufficient.

Ah, assembly stubs make perfect sense.

John.