| Summary: | A trivial fix for a potential crash | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | jr <j-r> |
| Component: | Drivers/DRI/nouveau | Assignee: | Nouveau Project <nouveau> |
| Status: | RESOLVED FIXED | QA Contact: | Nouveau Project <nouveau> |
| Severity: | trivial | ||
| Priority: | medium | ||
| Version: | git | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: | Candidate patch | ||
Great patch. Please mail it out to nouveau@lists.freedesktop.org, these things are usually just done over email when there's no real bug to keep track of. One small thing is I normally use "nv50/ir" prefix when modifying stuff in codegen. I'm away from computers where I can check this out for another week or so, but I'll be able to take a detailed look then. [And your other bug too.] Hrmph.... one small difference is that mkOp() appears to set the instruction with ->fixed = 1. I'm adding that in just in case it mattered in this case... Converter::insertConvergenceOps seems to do this too. Now pushed. In the future, you can just mail patches to nouveau@lists.freedesktop.org and/or mesa-dev@lists.freedesktop.org. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
Created attachment 116340 [details] Candidate patch OP_JOIN are static casted to FlowInstruction, but handleTXL creates OP_JOIN that is no FlowInstruction. It probably doesn't matter outside debug builds, since only the IR printer seems to access the additional FlowInstruction fields.