[Lldb-commits] [PATCH] D116211: With firmware debug sess, if gdb stub knows the UUID/address of binary, try to load it

Jason Molenda via Phabricator via lldb-commits lldb-commits at lists.llvm.org
Thu Dec 23 01:11:51 PST 2021

jasonmolenda created this revision.
jasonmolenda added a reviewer: JDevlieghere.
jasonmolenda added a project: LLDB.
jasonmolenda requested review of this revision.

We have a gdb stub that can get the UUID and address/slide of the firmware running on a device, and I'm adding support for sending that information up to lldb, so lldb can try to find & load the binary at the correct address/offset.

I'd discussed this once in the past, long ago, trying to decide how we would handle this - a new packet? - and I think it was Greg who suggested qHostInfo could get a couple of new key-value pairs to do it.  I think that idea is the right one, but it makes more sense in qProcessInfo to me, so that's where I've implemented it.

I haven't thought up a way to test this artificially yet.  While writing the patch, I hardcoded the new fields in debugserver's qProcessInfo reply, and I removed the darwin dynamic loader plugins in lldb, so they wouldn't get control and wipe the image list I manually loaded, hah.

  rG LLVM Github Monorepo



-------------- next part --------------
A non-text attachment was scrubbed...
Name: D116211.395995.patch
Type: text/x-patch
Size: 7813 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20211223/da0fe59f/attachment.bin>

More information about the lldb-commits mailing list