Commit | Line | Data |
---|---|---|
07c788ce | 1 | -*- text -*- |
0e39a8bb | 2 | |
07c788ce | 3 | PORTING: |
542e1629 | 4 | |
5bed5321 KR |
5 | Sorry, no description of the interfaces is written up yet. Look at existing |
6 | back ends and work from there. | |
07c788ce | 7 | |
77dea544 ILT |
8 | New hosts: The configure script, which is generated by autoconf, |
9 | should handle all host specific configuration. | |
10 | ||
11 | New target formats: Look at the BFD_ASSEMBLER code. The a.out or ELF | |
12 | code might be a fair example. There are no "good" examples yet, | |
13 | unfortunately, nor any good documentation of the changes. | |
07c788ce | 14 | |
5bed5321 KR |
15 | New target processors: Check first to see if the BFD_ASSEMBLER interface is |
16 | supported by the file format code you need to use. | |
07c788ce | 17 | |
07c788ce KR |
18 | DOCUMENTATION: |
19 | ||
1535ff15 | 20 | The internals of gas need documenting. |
07c788ce | 21 | |
607c8625 KR |
22 | The documentation should also contain a "Trouble" section similar to gcc's |
23 | manual: real bugs, common problems, incompatibilities, etc. | |
07c788ce KR |
24 | |
25 | Anyone want to offer to maintain a man page? | |
26 | ||
27 | BFD CONVERSION: | |
28 | ||
5bed5321 KR |
29 | The "#ifdef BFD_ASSEMBLER" code is on its way in; the "#ifndef BFD_ASSEMBLER" |
30 | code is on its way out. The new code uses BFD data structures, and calls BFD | |
31 | for anything that needs to be written to the output file. The old code did all | |
32 | the writing itself, or in a couple of cases, used BFD as a slightly higher | |
33 | level than stdio (i.e., bfd_seek, bfd_write -- these are not the preferred | |
34 | interface). | |
07c788ce KR |
35 | |
36 | Because of this, some of this code is messy. Lots of ifdef's, and the | |
5bed5321 KR |
37 | non-BFD_ASSEMBLER version often has multiple conditional tests inside it for |
38 | various processors or formats. As the various targets get converted over, | |
39 | these will gradually go away. | |
07c788ce | 40 | |
07c788ce KR |
41 | TO DO: |
42 | ||
43 | Remove DONTDEF code, commented-out code. | |
44 | ||
5bed5321 KR |
45 | Eliminate, as much as possible, anything not in config that is conditionalized |
46 | on a CPU, format, or environment. | |
07c788ce | 47 | |
09257909 KR |
48 | Merge COFF support into one version, supporting all the pseudo-ops used in |
49 | either versions now, but using BFD for high-level operations. (See second | |
50 | following item.) Currently there are two versions (plus the new BFD code), | |
51 | which support different features, and are used on different targets. | |
52 | ||
53 | Convert remaining a.out/b.out targets to using the BFD_ASSEMBLER code by | |
54 | default. | |
55 | ||
5bed5321 KR |
56 | Finish conversion to using BFD for all object file writing. (This is the |
57 | BFD_ASSEMBLER code, not BFD or BFD_HEADERS.) VMS might be the tough one here, | |
09257909 KR |
58 | since there's no BFD support for it at all yet. Eliminate the old code. Some |
59 | of this can be done target by target, so doing a target where the CPU or | |
60 | format already supports BFD_ASSEMBLER mode may be easiest. | |
07c788ce | 61 | |
de001789 KR |
62 | Fix lots of uses of empty strings to use null pointers. Will improve |
63 | efficiency, and should make code clearer too. | |
64 | ||
5bed5321 KR |
65 | Clean up comments; lots of 'em are one previous maintainer griping about |
66 | another previous maintainer, unrelated to the code. (And with no names, | |
67 | they're not so fun to read. :-) | |
07c788ce | 68 | |
0c3cab7a KR |
69 | For sparc: "call 0" becomes "jmpl %g0,%l7", and similarly for absolute |
70 | addresses in -4096...4095. (Solaris assembler does this. No | |
71 | relocation required, no absolute symbol needed.) For addresses | |
72 | outside the range, for COFF, keep generating an absolute symbol to use | |
73 | for relocs. | |
74 | ||
09257909 KR |
75 | Get Steve to document H8/500 stuff (and others). |
76 | ||
77 | Improve test suite. Incorporate more reported net bugs, and non-confidential | |
78 | Cygnus customer bugs, and anything else. | |
79 | ||
80 | Add support for i386/i486 16-bit mode, so operating system initialization code | |
81 | doesn't require a separate assembler nor lots of `.byte' directives. | |
de001789 | 82 | |
09257909 KR |
83 | See if it's more maintainable (and not too much of a performance loss) to use |
84 | a yacc grammar for parsing input. The lexer will have to be flexible, and the | |
85 | grammar will have to contain any construct used on any platform, but it may be | |
86 | easier to maintain, instead of having code in most of the back ends. | |
07c788ce | 87 | |
09257909 | 88 | PIC support. |
5bed5321 | 89 | |
74a88e8b KR |
90 | Torbjorn Granlund <tege@cygnus.com> writes, regarding alpha .align: |
91 | ||
92 | Please make sure the .align directive works as in digital's assembler. | |
93 | They fill the space with a sequence of "bis $31,$31,$31;ldq_u $31,0($30)" | |
94 | since these two instructions can dual-issue. Since .align is ued a lot by | |
95 | gcc, it is an important optimization. | |
96 | ||
07c788ce KR |
97 | (From old "NOTES" file to-do list, not really reviewed:) |
98 | ||
99 | fix relocation types for i860, perhaps by adding a ref pointer to fixS? | |
c6cfc5b5 | 100 | |
0e39a8bb | 101 | remove the ifdef's from fx_callj tests? |
0e39a8bb | 102 | |
07c788ce | 103 | space tighten sparc alignment? |
542e1629 | 104 | |
07c788ce | 105 | md_ => tc_ |
0e39a8bb | 106 | |
07c788ce | 107 | share b.out with a.out. |