Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | remove sparcv9 code. | 2004-02-02 | 1 | -8/+2 | |
| | | | | ok marc@ | ||||
* | - save and restore %fsr, too. | 2003-02-14 | 1 | -1/+15 | |
| | | | | - better function descriptions | ||||
* | priviledged -> privileged from David Krause via Henning Brauer | 2003-02-10 | 1 | -2/+2 | |
| | |||||
* | save fpu context (needs optimization) | 2003-01-26 | 1 | -1/+48 | |
| | |||||
* | sigh, do it right. | 2001-12-18 | 1 | -2/+2 | |
| | |||||
* | Fix func decl. in comment. | 2001-12-18 | 1 | -2/+5 | |
| | |||||
* | Software invoked traps should use the ST_ versions of the trap number | 2001-12-18 | 1 | -4/+4 | |
| | | | | | | | | | | | | with the high bit of the trap number stripped per the V8 arch manual. The stack alignment code was broken. It tried to align to a 32 byte boundary (only 8 byte boundary is required) by first adding 15 bytes then masking. It could have aligned down. The minimum frame size was twice the size needed by allocating 8 bytes per saved value instead of the needed 4. This bug offset the bug in the stack alignment code. | ||||
* | Our gcc defines __sparcv9__ not __sparc_v9__ | 2001-09-10 | 1 | -2/+2 | |
| | |||||
* | preparation for switch-based threads | 2000-09-25 | 1 | -0/+39 | |