HACKS revision 1.42
1# $NetBSD: HACKS,v 1.42 2005/02/22 18:47:09 christos 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 99hack gcc-strict-aliasing 100cdate 6 August 2004 101who junyoung 102pr 26516 103file src/usr.sbin/racoon/racoon/Makefile : 1.21 104descr 105 GCC issues a warning "dereferencing type-punned pointer will break 106 strict-aliasing rules" when compiling 107 crypto/dist/kame/racoon/crypto/openssl.c with -O[23s]. Note that this 108 is due to use of type casting in a non-standard way in the code and 109 not a compiler bug. Once a new release of KAME with a proper fix 110 applied is imported this should be backed out. 111kcah 112 113 114port vax 115 116 hack gcc 2.95/vax cannot handle __builtin_ffs() 117 cdate Fri Aug 23 21:31:15 CEST 2002 118 who ragge 119 file sys/lib/libkern/libkern.h : 1.42 120 descr 121 __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 122 added this feature to GCC 3 so it can be removed when switching. 123 kcah 124 125 hack gcc-vax-libbz2 126 mdate 27 Jun 2002 127 who thorpej 128 file lib/libbz2/Makefile 129 descr 130 libbz2 is mis-compiled with optimization with GCC 2.95.3 131 on VAX. -O0 works around this problem. 132 kcah 133 134 hack gawk 135 mdate 8 Nov 2003 136 who ragge 137 file gnu/usr.bin/gawk/Makefile 138 descr 139 gawk is mis-compiled with optimization with GCC 2.95.3 140 on VAX. -O0 works around this problem. 141 kcah 142 143 hack vax Toolchain bug 144 cdate 28 Mar 2003 145 who he 146 file sys/arch/vax/vax/intvec.S : 1.5 147 descr 148 Workaround for PR toolchain/20924. The assembler apparently 149 tries to range-check byte offsets when it doesn't have 150 sufficient information to make that decision. 151 Workaround uses `brw' instead of `brb' instruction. 152 kcah 153 154 hack vax gcc 2.95.3 -Wuninitialized workarounds 155 cdate 6 Nov 2003 156 who he 157 file sys/dev/bi/if_ni.c : 1.22 158 file sys/arch/vax/vax/ctu.c : 1.20 159 descr 160 Workaround for gcc 2.95.3 failing to detect that certain 161 variables will be initialized; so that this code compiles 162 with -Wuninitialized. Should be reviewed and possibly 163 reverted when gcc 3.3.2 is ready for vax. 164 kcah 165 166 hack vax gcc 2.95.3 structure initialization 167 cdat 13 May 2004 168 who cl 169 file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 170 descr 171 Workaround for gcc 2.95.3 failing to initialize structures 172 and/or unions inside structures using nested designators. 173 Should be reverted when gcc >=3.3.3 is ready for vax. 174 kcah 175 176 hack vax gcc 2.95.3 needs -I. to build nslexer.c 177 cdat 6 Jun 2004 178 who mhitch 179 file lib/libc/arch/vax/Makefile.inc : 1.6 180 descr 181 It appears necessary to add -I. to CPPFLAGS when building 182 nslesxer.c. This may be caused by gcc 2.95.3 being used. 183 Should be reverted when a newer gcc is ready for vax. 184 kcah 185 186port sparc64 187 188 hack binutil-2.11-sparc64-pltrela 189 mdate 14 Aug 2001 190 who eeh 191 file libexec/ld.elf_so/arch/sparc64/mdreloc.c 192 descr 193 The first four PLT entries are reserved. There is some 194 disagreement whether they should have associated relocation 195 entries. Both the SPARC 32-bit and 64-bit ELF specifications 196 say that they should have relocation entries, but the 32-bit 197 SPARC binutils do not generate them, and now the 64-bit SPARC 198 binutils have stopped generating them too. 199 200 To provide binary compatibility, we will check the first entry, 201 if it is reserved it should not be of the type JMP_SLOT. If it 202 is JMP_SLOT, then the 4 reserved entries were not generated and 203 our index is 4 entries too far, so we frob the rela pointer. 204 kcah 205 206 207port arm 208 209 hack gcc-unsigned-compare 210 cdate 09 Mar 2002 211 mdate 18 Mar 2002 212 who bjh21 213 file dist/bind/lib/nameser/ns_parse.c : 1.3 214 file dist/dhcp/minires/ns_parse.c : 1.3 215 file dist/dhcp/omapip/result.c : 1.2 216 file dist/dhcp/server/failover.c : 1.3 217 file gnu/dist/gawk/eval.c : 1.4 218 file gnu/dist/toolchain/bfd/bfd.c : 1.2 219 file gnu/dist/toolchain/bfd/format.c : 1.2 220 file gnu/dist/toolchain/gdb/target.c : 1.2 221 file sys/kern/vfs_subr.c : 1.172 222 descr When checking that a potentially-unsigned enum is >= 0, assign 223 it to an int first. This is necessary to avoid "comparison is 224 always true" warnings with -fshort-enums. Casting to an int 225 really should be enough, but turns out not to be. 226 kcah 227 228 hack gcc 3.3.2/arm sys/kern/uipc_socket.c 229 cdate Tue Oct 28 18:02:16 GMT 2003 230 who skrll 231 pr 23044 232 file sys/arch/arm/conf/Makefile.arm : 1.19 233 descr 234 Performing a build.sh tools on a CATS with NFS mounted 235 obj and src directories results in "panic: receive 1" 236 The hack is to compile sys/kern/uipc_socket.c with 237 -fno-strict-aliasing 238 kcah 239 240 241port pc532 242 243 hack egcs-pc532-ip6_mroute 244 cdate 09 Jul 2002 245 who simonb 246 file sys/arch/pc532/conf/Makefile.pc532 : 1.70 247 file sys/arch/pc532/conf/files.pc532 : 1.47 248 descr 249 egcs 1.1.2 gets an "internal error--insn does not satisfy its 250 constraints" error compiling ip6_mroute.c with -O2 or greater. 251 -O1 works around this problem. 252 kcah 253 254 hack gcc-pc532-duffs_device 255 cdate Mon Oct 27 07:23:05 UTC 2003 256 who simonb 257 file sys/arch/pc532/conf/Makefile.pc532 : 1.71 258 file sys/arch/pc532/conf/files.pc532 : 1.52 259 pr GCC PR optimization/5230 260 descr 261 gcc incorrectly gives an "unreachable code at beginning of 262 switch statement" for a Duff's device construct in 263 arch/pc532/dev/ncr.c. There is no way to disable just this 264 warning, so -Wno-error is turned on for this file. 265 kcah 266 267 hack gcc332-cppmacro-ice 268 cdate Sun Dec 7 12:48:33 UTC 2003 269 who simonb 270 file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 271 descr 272 gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 273 The hack is to use -O1 for this file. Fixed in gcc 274 -current mainline, unable to work out where the fix is 275 right now. 276 kcah 277 278 279port sh3 280 281 hack gcc-sh3-sed 282 mdate 23 Apr 2002 283 who thorpej 284 file usr.bin/sed/Makefile : 1.9 : 9 : 13 285 descr 286 The in-tree GCC 2.95.3-based compiler ICEs when building 287 with optimization for SuperH. Hack is to build with -O0. 288 kcah 289 290 291port sh5 292 293 hack SuperH SH5 Toolchain Bugs 294 cdate 11 Jul 2002 295 who scw 296 file usr.sbin/ndbootd/ndbootd.c : 1.6 297 file usr.sbin/traceroute/traceroute.c : 1.48 298 descr 299 The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 300 compiler error when assigning a bit-wise inverted value 301 under some circumstances. 302 Work around it by splitting the statement into two. 303 kcah 304 305 hack SuperH SH5 Toolchain Bugs 306 cdate 17 May 2003 307 who scw 308 file lib/libpthread/pthread_lock.c : 1.7 309 descr 310 The SuperH SH5 toolchain generates incorrect PIC code when 311 faced with a symbol which is declared extern, but has local 312 scope due to being defined within an asm() statement (without 313 being declared .globl in that statement). Work around it by 314 adding the .globl. 315 kcah 316 317 hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 318 cdate Sun Dec 7 19:44:05 GMT 2003 319 who scw 320 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 321 descr 322 Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 323 Drop to -O1 for now when building for sh5. 324 kcah 325 326 hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 327 cdate Sun Dec 7 21:10:46 GMT 2003 328 who scw 329 file gnu/lib/libbfd/Makefile : 1.39 330 descr 331 Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 332 and elf64.c at -O2. 333 Drop to -O1 for now when building for sh5. 334 kcah 335 336 337 hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 338 cdate Sun Dec 7 19:44:05 GMT 2003 339 who scw 340 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 341 descr 342 Gcc3 for sh5 issues a warning: 343 .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 344 mutually exclusive equal-tests is always 0. 345 This causes the build to fail because of -Werror. 346 Work-around by defining NOGCCERROR when building for sh5. 347 kcah 348 349 hack gcc 3.3.3 on sh5 - ICE compiling rbtdb.c and rbtdb64.c 350 cdate Thu Nov 11 00:35:33 UTC 2004 351 who he 352 file usr.sbin/bind/libdns/Makefile/1.2 353 descr 354 Gcc3 for sh5 gets an ICE compiling these files at -O2. 355 Drop to -O1 for now when building for sh5. 356 kcah 357 358 359port sun2 360 361 hack gcc 3.3.2/mdsetimage 362 cdate Tue Oct 28 18:43:05 EST 2003 363 who fredette 364 file gnu/usr.sbin/mdsetimage/Makefile : 1.15 365 descr 366 mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 367 this hack uses -O0 to avoid the problem. 368 kcah 369 370 hack gcc 3.3.2/installboot/ffs.c 371 cdate Tue Oct 28 18:43:05 EST 2003 372 who fredette 373 file usr.sbin/installboot/Makefile : 1.26 374 descr 375 ffs.c causes cc1 to barf for as-yet unknown reasons. this 376 hack uses -O0 to avoid the problem. 377 kcah 378 379 380port sun2, m68k 381 382 hack gcc 3.3.2/gcc/ra-colorize.c 383 cdate Wed Nov 5 11:03:13 EST 2003 384 who mrg 385 file gnu/usr.bin/gcc/backend/Makefile : 1.11 386 descr 387 ra-colorize.c causes a whole spew of unreferenced .L* local 388 symbols. using -O0 (or even -O2 reportedly) avoids the problem. 389 kcah 390 391port sparc64 392 393 hack disable optimzations for uvm_bio.c on 32 bit kernels 394 cdate Sun Mar 21 14:14:04 MET 2004 395 who martin 396 file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 397 file sys/arch/sparc64/conf/files.sparc64 : 1.93 398 descr 399 The sparc compiler miscompiles uvm_bio.c when using 400 any optimization. This results in ubc_release panics. 401 kcah 402 403port macppc 404 405 hack remove inline from rotate_{left,right} 406 cdate Tue Feb 22 13:43:25 EST 2005 407 who christos 408 file src/dist/bind/lib/dns/rbt.c 409 descr 410 The macppc gcc-3.3 compiler miscompiles rbt.c when these 411 functions are inlined. The error is: 412 .* named[368]: rbt.c:1584: INSIST(child != (void *)0) >failed 413 .* named[368]: exiting (due to assertion failure) 414 This is reportedly fixed in gcc-3.4 415 kcah 416