Article View: gnu.gcc.help
Article #53507GCC internals - register allocation, v1.40
From: Howard Chu
Date: Wed, 06 Nov 1991 19:50
Date: Wed, 06 Nov 1991 19:50
12 lines
1330 bytes
1330 bytes
I've been working on modifying the 68k description to always use base-register + index for data references. I have pretty much succeeded, but something is not right in the register allocation/reload pass, there is always one new pseudo-register that didn't make it into a hard register. E.g., the local-register dump says it used 150 registers, and after the global/reload pass I see a pseudo-reg number 60 (for example) changed into #151. In the final assembly pass constrain_operands says this doesn't match its register constraint, so the compiler aborts. Has anyone seen this happen before, and does anyone know why it's happening? -- -- Howard Chu @ Jet Propulsion Laboratory, Pasadena, CA
Technical Details
Message-ID:
<1991Nov6.195053.9048@elroy.jpl.nasa.gov>Path:
arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!elroy.jpl.nasa.gov!hanauma.jpl.nasa.gov!hycShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 11318c,17062c4159962e44,start X-Google-Attributes: gid11318c,public X-Google-ArrivalTime: 1991-11-06 12:02:08 PST Path: arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!elroy.jpl.nasa.gov!hanauma.jpl.nasa.gov!hyc From: hyc@hanauma.jpl.nasa.gov (Howard Chu) Newsgroups: gnu.gcc.help Subject: GCC internals - register allocation, v1.40 Message-ID: <1991Nov6.195053.9048@elroy.jpl.nasa.gov> Date: 6 Nov 91 19:50:53 GMT Sender: news@elroy.jpl.nasa.gov (Usenet) Organization: SAR Systems Development and Processing, JPL Lines: 12 Nntp-Posting-Host: hanauma.jpl.nasa.gov