[Lldb-commits] [PATCH] D55718: [ARC] Basic support in gdb-remote process plugin

Tue Mar 26 12:26:12 PDT 2019

labath added a comment.

In D55718#1443415 <https://reviews.llvm.org/D55718#1443415>, @jasonmolenda wrote:

> In D55718#1442965 <https://reviews.llvm.org/D55718#1442965>, @clayborg wrote:
>
> > It would be really nice to get the GDB remote server for ARC so that it vends the correct regs to begin with, but if that isn't possible I guess we need to do what we are doing here. I would really like to not see more architectures have to add code to ProcessGDBRemote.cpp at all.  If that isn't possible this solution will need to be used.
>
>
> Unrelated to this patchset, but just last week I was working on something and thinking about how our system of "the RSP stub is the canonical source of register knowledge" is probably not the best architecture.  IMO the remote stub should teach us (1) the name of each register, (2) the numbers it will use for each register, (3) the size in bits of each register, and (4) the offset into the g/G packet of that register.
>
> Everything else about registers should come from lldb in either architecture-specific definitions (rax has 64 bits, and it has a slice register eax of 32 bits, an ax of 16 bits etc; it is printed as Uint64/32/16), or ABI definitions (rdi is arg1 in the posix x86_64 ABI; rsp is eh_frame regnum 7, dwarf regnum 7).  Coordination between the RSP stub and lldb should be done by register name, confirmed by register size (the remote stub says it has a 64-bit r0 and I only know about a 32-bit r0, we have a problem).
>
> Back in 2015 I added some code in r247121 that would fill in eh_frame and dwarf register numbers from the ABI plugin if the remote stub didn't provide them in ProcessGDBRemote's AugmentRegisterInfoViaABI.  It might be better for ARC to hardcode these generic register additions some place like that.  The change in 247121 was a step in the direction of having lldb be the canonical source of information about registers, but only doing the easy first bit.

+1000

This sounds really great to me.

Repository:
  rLLDB LLDB

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D55718/new/

https://reviews.llvm.org/D55718