HACKS revision 1.36
1# $NetBSD: HACKS,v 1.36 2004/06/06 17:11:33 he Exp $ 2# 3# This file is intended to document workarounds for currently unsolved 4# (mostly) compiler bugs. 5# 6# Format: 7# hack title 8# cdate creation date 9# mdate mod date 10# who responsible developer 11# port ... 12# affected ports, space separated, if not "all" 13# file affected file : revision : line from : line to 14# affected files and revision and line numbers describing hack 15# multiple lines if necessary. 16# pr NNNN ... 17# problem reports this hack works around, if known. Space 18# separated. 19# regress src/regress/directory/where/test/found 20# regression test directories, if available. 21# descr 22# insert short informal description (multi-line). (Longer ones 23# should be in the PR database. More formal descriptions might 24# be in the regress tree. See above). 25# kcah 26# closing bracket. 27# 28# this is a comment. 29 30hack netstat ieee1394 address printing. 31mdate 14 Nov 2000 32who matt 33file lib/libc/net/getnameinfo.c : 1.32 : 497 : 503 34descr 35 Because the current implementation of IP over IEEE1394, the 36 fw device address contains more than just the IEEE1394 EUI-64. 37 So when printing out IEEE1394 addresses, ignore the extra stuff. 38kcah 39 40hack xterm vs. libterm 41mdate 01 Aug 2000 42who jdc 43file xsrc/xc/programs/xterm/main.c : 1.2 : 3609 : 3614 44pr 10383 45descr 46 In order to extend the termcap string over 1023 bytes, a ZZ entry was 47 introduced to point to a memory location containing the full entry. 48 Without this hack, xterm will export a termcap containing the ZZ 49 entry, which will then be ignored by libterm. As xterm modifies the 50 exported termcap, this would cause those modifications to be ignored. 51kcah 52 53hack wi-at-big-endian-bus 54cdate 15 Mar 2002 55who martin 56file dev/ic/wireg.h 57descr Add an option to access the underlying bus in big endian byte order 58 to work around deficiencies in bus_space_{read,write}_* macros. 59 Those don't allow the implementation of a proper pcmcia bus space 60 tag. 61kcah 62 63hack specific knowledge of colours in curses code 64cdate Sun Apr 6 11:05:24 BST 2003 65who jdc 66file lib/libcurses/color.c : r1.24 67descr 68 Swap red/blue and yellow/cyan colours for COLOR_OTHER. 69 Fix is to enhance libtermcap to understand terminfo-style % sequences. 70 See also: 71 http://mail-index.NetBSD.org/tech-userlevel/2003/04/06/0000.html 72kcah 73 74hack Compensation for differing types of LINUX_USRSTACK and USRSTACK 75cdate 21 Aug 2003 76who he 77file sys/miscfs/procfs/procfs_linux.c : 1.14 78descr 79 Not all ports have LINUX_USRSTACK and/or USRSTACK defined as 80 literals/constants, but refer to variables of a type which is 81 not "unsigned long", causing compilation of procfs_linux.c to 82 fail with "makes integer from pointer without a cast". This 83 is observed on e.g. the sun3 port. Ideally the "types" for 84 symbols should be consistent across all ports. 85kcah 86 87hack GCC2 isn't C99 88cdate 5 December 2003 89who mrg 90file gnu/dist/binutils/ld/ldlex.c: r1.2 91descr 92 ldlex.c is generated from ldlex.l and we get macro redefinition 93 errors from it if __STDC_VERSION__ is not C99 compliant. This 94 hack forces NetBSD 1.6 and later systems to also include <stdint.h> 95 to get the relevant definitions. This hack should be backed out 96 when all platforms have switched away from GCC2. 97kcah 98 99 100port vax 101 102 hack gcc 2.95/vax cannot handle __builtin_ffs() 103 cdate Fri Aug 23 21:31:15 CEST 2002 104 who ragge 105 file sys/lib/libkern/libkern.h : 1.42 106 descr 107 __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 108 added this feature to GCC 3 so it can be removed when switching. 109 kcah 110 111 hack gcc-vax-libbz2 112 mdate 27 Jun 2002 113 who thorpej 114 file lib/libbz2/Makefile 115 descr 116 libbz2 is mis-compiled with optimization with GCC 2.95.3 117 on VAX. -O0 works around this problem. 118 kcah 119 120 hack gawk 121 mdate 8 Nov 2003 122 who ragge 123 file gnu/usr.bin/gawk/Makefile 124 descr 125 gawk is mis-compiled with optimization with GCC 2.95.3 126 on VAX. -O0 works around this problem. 127 kcah 128 129 hack vax Toolchain bug 130 cdate 28 Mar 2003 131 who he 132 file sys/arch/vax/vax/intvec.S : 1.5 133 descr 134 Workaround for PR toolchain/20924. The assembler apparently 135 tries to range-check byte offsets when it doesn't have 136 sufficient information to make that decision. 137 Workaround uses `brw' instead of `brb' instruction. 138 kcah 139 140 hack vax gcc 2.95.3 -Wuninitialized workarounds 141 cdate 6 Nov 2003 142 who he 143 file sys/dev/bi/if_ni.c : 1.22 144 file sys/arch/vax/vax/ctu.c : 1.20 145 descr 146 Workaround for gcc 2.95.3 failing to detect that certain 147 variables will be initialized; so that this code compiles 148 with -Wuninitialized. Should be reviewed and possibly 149 reverted when gcc 3.3.2 is ready for vax. 150 kcah 151 152 hack vax gcc 2.95.3 structure initialization 153 cdat 13 May 2004 154 who cl 155 file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 156 descr 157 Workaround for gcc 2.95.3 failing to initialize structures 158 and/or unions inside structures using nested designators. 159 Should be reverted when gcc >=3.3.3 is ready for vax. 160 kcah 161 162 hack vax gcc 2.95.3 needs -I. to build nslexer.c 163 cdat 6 Jun 2004 164 who mhitch 165 file lib/libc/arch/vax/Makefile.inc : 1.6 166 descr 167 It appears necessary to add -I. to CPPFLAGS when building 168 nslesxer.c. This may be caused by gcc 2.95.3 being used. 169 Should be reverted when a newer gcc is ready for vax. 170 kcah 171 172port sparc64 173 174 hack binutil-2.11-sparc64-pltrela 175 mdate 14 Aug 2001 176 who eeh 177 file libexec/ld.elf_so/arch/sparc64/mdreloc.c 178 descr 179 The first for PLT entries are reserved. There is some 180 disagreement whether they should have associated relocation 181 entries. Both the SPARC 32-bit and 64-bit ELF specifications 182 say that they should have relocation entries, but the 32-bit 183 SPARC binutils do not generate them, and now the 64-bit SPARC 184 binutils have stopped generating them too. 185 186 To provide binary compatibility, we will check the first entry, 187 if it is reserved it should not be of the type JMP_SLOT. If it 188 is JMP_SLOT, then the 4 reserved entries were not generated and 189 our index is 4 entries too far, so we frob the rela pointer. 190 kcah 191 192 193port arm 194 195 hack gcc-unsigned-compare 196 cdate 09 Mar 2002 197 mdate 18 Mar 2002 198 who bjh21 199 file dist/bind/lib/nameser/ns_parse.c : 1.3 200 file dist/dhcp/minires/ns_parse.c : 1.3 201 file dist/dhcp/omapip/result.c : 1.2 202 file dist/dhcp/server/failover.c : 1.3 203 file gnu/dist/gawk/eval.c : 1.4 204 file gnu/dist/toolchain/bfd/bfd.c : 1.2 205 file gnu/dist/toolchain/bfd/format.c : 1.2 206 file gnu/dist/toolchain/gdb/target.c : 1.2 207 file sys/kern/vfs_subr.c : 1.172 208 descr When checking that a potentially-unsigned enum is >= 0, assign 209 it to an int first. This is necessary to avoid "comparison is 210 always true" warnings with -fshort-enums. Casting to an int 211 really should be enough, but turns out not to be. 212 kcah 213 214 hack gcc 3.3.2/arm sys/kern/uipc_socket.c 215 cdate Tue Oct 28 18:02:16 GMT 2003 216 who skrll 217 file sys/conf/arm/Makefile.arm : 1.19 218 descr 219 Performing a build.sh tools on a CATS with NFS mounted 220 obj and src directories results in "panic: receive 1" 221 The hack is to compile sys/kern/uipc_socket.c with 222 -fno-strict-aliasing 223 kcah 224 225 226port pc532 227 228 hack egcs-pc532-ip6_mroute 229 cdate 09 Jul 2002 230 who simonb 231 file sys/arch/pc532/conf/Makefile.pc532 : 1.70 232 file sys/arch/pc532/conf/files.pc532 : 1.47 233 descr 234 egcs 1.1.2 gets an "internal error--insn does not satisfy its 235 constraints" error compiling ip6_mroute.c with -O2 or greater. 236 -O1 works around this problem. 237 kcah 238 239 hack gcc-pc532-duffs_device 240 cdate Mon Oct 27 07:23:05 UTC 2003 241 who simonb 242 file sys/arch/pc532/conf/Makefile.pc532 : 1.71 243 file sys/arch/pc532/conf/files.pc532 : 1.52 244 pr GCC PR optimization/5230 245 descr 246 gcc incorrectly gives an "unreachable code at beginning of 247 switch statement" for a Duff's device construct in 248 arch/pc532/dev/ncr.c. There is no way to disable just this 249 warning, so -Wno-error is turned on for this file. 250 kcah 251 252 hack gcc332-cppmacro-ice 253 cdate Sun Dec 7 12:48:33 UTC 2003 254 who simonb 255 file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 256 descr 257 gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 258 The hack is to use -O1 for this file. Fixed in gcc 259 -current mainline, unable to work out where the fix is 260 right now. 261 kcah 262 263 264port sh3 265 266 hack gcc-sh3-sed 267 mdate 23 Apr 2002 268 who thorpej 269 file usr.bin/sed/Makefile : 1.9 : 9 : 13 270 descr 271 The in-tree GCC 2.95.3-based compiler ICEs when building 272 with optimization for SuperH. Hack is to build with -O0. 273 kcah 274 275 276port sh5 277 278 hack SuperH SH5 Toolchain Bugs 279 cdate 11 Jul 2002 280 who scw 281 file usr.sbin/ndbootd/ndbootd.c : 1.6 282 file usr.sbin/traceroute/traceroute.c : 1.48 283 descr 284 The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 285 compiler error when assigning a bit-wise inverted value 286 under some circumstances. 287 Work around it by splitting the statement into two. 288 kcah 289 290 hack SuperH SH5 Toolchain Bugs 291 cdate 17 May 2003 292 who scw 293 file lib/libpthread/pthread_lock.c : 1.7 294 descr 295 The SuperH SH5 toolchain generates incorrect PIC code when 296 faced with a symbol which is declared extern, but has local 297 scope due to being defined within an asm() statement (without 298 being declared .globl in that statement). Work around it by 299 adding the .globl. 300 kcah 301 302 hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 303 cdate Sun Dec 7 19:44:05 GMT 2003 304 who scw 305 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 306 descr 307 Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 308 Drop to -O1 for now when building for sh5. 309 kcah 310 311 hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 312 cdate Sun Dec 7 21:10:46 GMT 2003 313 who scw 314 file gnu/lib/libbfd/Makefile : 1.39 315 descr 316 Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 317 and elf64.c at -O2. 318 Drop to -O1 for now when building for sh5. 319 kcah 320 321 322 hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 323 cdate Sun Dec 7 19:44:05 GMT 2003 324 who scw 325 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 326 descr 327 Gcc3 for sh5 issues a warning: 328 .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 329 mutually exclusive equal-tests is always 0. 330 This causes the build to fail because of -Werror. 331 Work-around by defining NOGCCERROR when building for sh5. 332 kcah 333 334 hack gcc 3.3.2 on sh5 - ICE compiling dist/bind/bin/named/ns_main.c 335 cdate Sun Dec 7 19:44:05 GMT 2003 336 who scw 337 file usr.sbin/bind/named/Makefile : 1.15 338 descr 339 Gcc3 for sh5 gets an ICE compiling ns_main.c at -O2. 340 Drop to -O1 for now when building for sh5. 341 kcah 342 343 344port sun2 345 346 hack gcc 3.3.2/mdsetimage 347 cdate Tue Oct 28 18:43:05 EST 2003 348 who fredette 349 file gnu/usr.sbin/mdsetimage/Makefile : 1.15 350 descr 351 mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 352 this hack uses -O0 to avoid the problem. 353 kcah 354 355 hack gcc 3.3.2/installboot/ffs.c 356 cdate Tue Oct 28 18:43:05 EST 2003 357 who fredette 358 file usr.sbin/installboot/Makefile : 1.26 359 descr 360 ffs.c causes cc1 to barf for as-yet unknown reasons. this 361 hack uses -O0 to avoid the problem. 362 kcah 363 364 365port sun2, m68k 366 367 hack gcc 3.3.2/gcc/ra-colorize.c 368 cdate Wed Nov 5 11:03:13 EST 2003 369 who mrg 370 file gnu/usr.bin/gcc/backend/Makefile : 1.11 371 descr 372 ra-colorize.c causes a whole spew of unreferenced .L* local 373 symbols. using -O0 (or even -O2 reportedly) avoids the problem. 374 kcah 375 376port sparc64 377 378 hack disable optimzations for uvm_bio.c on 32 bit kernels 379 cdate Sun Mar 21 14:14:04 MET 2004 380 who martin 381 file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 382 file sys/arch/sparc64/conf/files.sparc64 : 1.93 383 descr 384 The sparc compiler miscompiles uvm_bio.c when using 385 any optimization. This results in ubc_release panics. 386 kcah 387