| Age | Commit message (Collapse) | Author | Files | Lines | |
|---|---|---|---|---|---|
| 2011-07-02 | cw.1: Update to mention GCC_ROOT | Richard Lowe | 1 | -1/+1 | |
| 2011-07-02 | bootblks: Use GCC_ROOT | Richard Lowe | 1 | -4/+2 | |
| 2011-07-02 | grub: Use GCC_ROOT | Richard Lowe | 1 | -3/+3 | |
| 2011-07-02 | cw: Use GCC_ROOT, not GNU_ROOT | Richard Lowe | 2 | -7/+5 | |
| 2011-07-02 | perl/Kstat: Don't cast pointers directly to int | Richard Lowe | 1 | -1/+1 | |
| 2011-07-02 | quota: Avoid undefined order of operations | Richard Lowe | 1 | -2/+2 | |
| 2011-07-02 | jdtrace: unreachable return from main, to pacify GCC | Richard Lowe | 1 | -1/+2 | |
| 2011-07-02 | hal: explicitly cast sentinel NULL to a pointer type | Richard Lowe | 1 | -1/+1 | |
| GCC's sentinel checks apparently only notice the sentinel NULL if it is of pointer type, and at the point g_strconcat() is used here, NULL is (integer) 0. Cast it to a pointer to silence the warning. | |||||
| 2011-07-02 | ipf: Avoid implicit truncation in IP_HL_A | Richard Lowe | 1 | -1/+1 | |
| 2011-07-02 | [HACK] vi: Cast away unsignedness to match getpass() | Richard Lowe | 1 | -1/+1 | |
| I have no idea why vgetpass() is using unsigned chars in the first place, but it seemed safest for now to preserve that and pacify GCC in the obvious, hacky, fashion. | |||||
| 2011-07-02 | svc: Don't pass NULL as an integer, pass 0 | Richard Lowe | 1 | -7/+7 | |
| 2011-07-02 | svc: Return -1 not NULL from find_manifests, which returns int | Richard Lowe | 1 | -1/+1 | |
| 2011-07-02 | kadmin: Adjust loop to avoid undefined order of operations | Richard Lowe | 1 | -22/+22 | |
| 2011-07-02 | rpcbind: Mark thread function __NORETURN | Richard Lowe | 1 | -36/+36 | |
| 2011-07-02 | nscd: Mark thread function __NORETURN | Richard Lowe | 1 | -1/+1 | |
| 2011-07-02 | ucbhead: Remove CPP "assertion" use, deprecated in GCC | Richard Lowe | 1 | -11/+0 | |
| 2011-07-02 | dnssd: Remove non-existent parameters from javadoc | Richard Lowe | 1 | -13/+7 | |
| 2011-07-02 | libkmsagent: Avoid deprecated string constant conversion | Richard Lowe | 1 | -6/+11 | |
| 2011-07-02 | libnisdb: Avoid deprecated string constant conversion | Richard Lowe | 10 | -91/+81 | |
| 2011-07-02 | libumem: Mark thread function __NORETURN | Richard Lowe | 1 | -3/+1 | |
| 2011-07-02 | libtsol: Squelch pointer to different sized int cast warning | Richard Lowe | 1 | -7/+5 | |
| 2011-07-02 | libsmbns: Explicit truncate cache status while byte swapping | Richard Lowe | 1 | -2/+2 | |
| 2011-07-02 | smbfs: Squelch assignment signedness warning | Richard Lowe | 1 | -9/+10 | |
| 2011-07-02 | libc_db: Squelch pointer to different sized int cast warning | Richard Lowe | 1 | -1/+2 | |
| 2011-07-02 | lvm: Squelch pointer to different-sized int cast warnings | Richard Lowe | 2 | -110/+107 | |
| 2011-07-02 | sun_fc: Add missing #includes | Richard Lowe | 10 | -1/+15 | |
| 2011-07-02 | libparted: Drop workaround for compiler we don't use | Richard Lowe | 1 | -4/+0 | |
| GCC would warn of implicit overflow in the preprocessor conditional used to select the workaround. | |||||
| 2011-07-02 | libnisdb: Use strcmp, not unspecified == of strings | Richard Lowe | 1 | -146/+142 | |
| 2011-07-02 | libbc (uname): Copy nodename in two passes to avoid array index warning | Richard Lowe | 1 | -14/+16 | |
| 2011-07-02 | libbc: fix string quoting | Richard Lowe | 1 | -4/+2 | |
| 2011-07-02 | libbc (usleep): Haul static prototypes out of functions | Richard Lowe | 1 | -10/+7 | |
| Function-local function prototypes may not be static, says GCC. | |||||
| 2011-07-02 | libc: Remove SPARC tls workarounds when using GCC 4 | Richard Lowe | 2 | -4/+6 | |
| 2011-07-02 | libc: make install target depend on generated headers | Richard Lowe | 1 | -1/+1 | |
| There a handful of generated headers in libc (assym.h), which need to exist for the build. They're removed by clobber but only created via install_h. Have 'install' depend on them as well so working in libc is less tedious. | |||||
| 2011-07-02 | libc: Implement __floatundisf | Richard Lowe | 3 | -0/+49 | |
| 2011-07-02 | libc: Implement __floatundidf | Richard Lowe | 3 | -0/+49 | |
| 2011-07-02 | libc: Mark %g7 as #scratch in _curthread, and implement directly | Richard Lowe | 1 | -5/+7 | |
| We do this because GCC 4.4 will spit out a .register declaration for %g7 which differs from those in __curthread if we do not. | |||||
| 2011-07-02 | libc: Don't let GCC optimize away __rtld | Richard Lowe | 1 | -1/+10 | |
| 2011-07-02 | openssl: Backport relevant sections of GCC 4.2 compilation fixes | Richard Lowe | 3 | -25/+66 | |
| GCC 4.2 and above absolutely forbid casts of function pointers (they in fact emit a fatal trap such that any code doing so fails). These are the fixes from OpenSSL CVS to cease any function pointer casting while retaining type checks via other means. | |||||
| 2011-07-02 | hfsops, ufsops: Use offsetof, not an expanded version | Richard Lowe | 2 | -6/+3 | |
| 2011-07-02 | bootops: Don't cast pointer directly down to int | Richard Lowe | 1 | -3/+1 | |
| 2011-07-02 | dis: Allow decoding of 66 (%cs:) segment override even when amd64 | Richard Lowe | 1 | -1/+1 | |
| This is now generated by GCC 4 as part of an extended multi-byte nop as variable-length executable padding of the form: 66 66 66 2e 0f 1f 84 Where the number of 66 (address size) prefixes varies according to the amount of pad required. | |||||
| 2011-07-02 | dis: Decode multibyte nops generated by GCC | Richard Lowe | 1 | -1/+1 | |
| This decodes 0f 1f (nopl) and 66 0f 1f (nopw), which seem to be the ones most commonly generated by gcc (determined by disassembling /kernel). unused opcodes in [0f18, 0f1f) are technically reserved nops, and used by Intel as "hinting nops", but I haven't found them generated. It doesn't seem to deal with '66 2e 0f 1f' (nopw %cs:) but I'm not certain as to why. This seems to mostly be generated inter-function however, and so this may not be such a big deal. | |||||
| 2011-07-02 | pcisch: Avoid GCC type promotion leading to format string warnings | Richard Lowe | 1 | -2/+3 | |
| Apparently GCC promotes a couple of cmn_err arguments to long long, which then causes it to complain that the %lx format is inappropriate. Explicitly cast to uint64_t. It is unclear why the promotion is occurring. | |||||
| 2011-07-02 | starcat/drmach: Haul static prototypes out of functions | Richard Lowe | 1 | -12/+13 | |
| Function-local function prototypes may not be static, says GCC. | |||||
| 2011-07-02 | starcat/mboxsc: Use offsetof() not a homebrew copy | Richard Lowe | 1 | -4/+3 | |
| 2011-07-02 | starcat/schpc: Remove impossible status value (3rd bit of a 2 bit field) | Richard Lowe | 2 | -10/+0 | |
| 2011-07-02 | starcat/cvc: Use offsetof, don't hand-roll | Richard Lowe | 1 | -2/+2 | |
| 2011-07-02 | opl/drmach: Haul static prototypes out of functions | Richard Lowe | 1 | -10/+10 | |
| Function-local function prototypes may not be static, says GCC. | |||||
| 2011-07-02 | cpu/opl: Haul static prototypes out of functions | Richard Lowe | 1 | -4/+4 | |
| Function-local function prototypes may not be static, says GCC. | |||||
| 2011-07-02 | serengeti/sbdp: Don't cast addresses down to int | Richard Lowe | 1 | -1/+2 | |
 |
index : illumos-joyent | |
| [no description] | Dyson |
| summaryrefslogtreecommitdiff |