HACKS revision 1.61
1# $NetBSD: HACKS,v 1.61 2006/05/12 03:26:54 mrg 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 113hack gcc4 wrong uninitialised variable 114mdate 10 May 2006 115who mrg 116file bin/ksh/eval.c : 1.6 117file bin/sh/histedit.c : 1.39 118file bin/sh/parser.c : 1.60 119file bin/systrace/systrace.c : 1.33 120file crypto/dist/heimdal/kdc/524.c : 1.10 121file crypto/dist/ssh/sftp.c : 1.20 122file crypto/dist/ssh/ssh-keysign.c : 1.11 123file dist/ipf/lib/hostname.c : 1.2 124file dist/ipf/tools/ipmon.c : 1.8 125file dist/ntp/ntpd/refclock_shm.c : 1.4 126file dist/ntp/sntp/timing.c : 1.3 127file dist/pppd/pppstats/pppstats.c : 1.3 128file dist/smbfs/lib/smb/rap.c : 1.6 129file dist/tcpdump/print-zephyr.c : 1.5 130file gnu/libexec/uucp/uucico/uucico.c : 1.6 131file lib/libc/citrus/citrus_csmapper.c : 1.6 132file lib/libc/citrus/citrus_pivot_factory.c : 1.5 133file lib/libc/inet/inet_cidr_ntop.c : 1.3 134file lib/libc/inet/inet_ntop.c : 1.3 135file lib/libc/stdio/vfwprintf.c : 1.8 136file libexec/ld.elf_so/arch/sparc64/mdreloc.c : 1.39 137file sbin/fsck_ext2fs/dir.c : 1.19 138file sbin/routed/rtquery/rtquery.c : 1.18 139file sys/crypto/cast128/cast128.c : 1.9 140file sys/ddb/db_command.c : 1.86 141file sys/dev/cardbus/cardbus_map.c : 1.xxx 142file sys/dev/fss.c : 1.25 143file sys/dev/ic/mb86950.c : 1.xxx 144file sys/dev/ic/midway.c : 1.71 145file sys/dev/usb/ohci.c : 1.xxx 146file sys/dev/usb/uhci.c : 1.xxx 147file sys/dev/pci/cmpci.c : 1.xxx 148file sys/dev/pci/machfb.c : 1.45 149file sys/dist/ipf/netinet/ip_nat.c : 1.10 150file sys/dist/ipf/netinet/ip_rpcb_pxy.c : 1.8 151file sys/dist/pf/net/pf.c : 1.22 152file sys/fs/udf/udf_vnops.c : 1.4 153file sys/kern/kern_sig.c : 1.219 154file sys/kern/tty.c : 1.181 155file sys/net/bpf.c : 1.116 156file sys/net/zlib.c : 1.26 157file sys/netccitt/if_x25subr.c : 1.xxx 158file sys/netinet/in.c : 1.107 159file sys/netsmb/smb_smb.c : 1.27 160file sys/netsmb/smb_trantcp.c : 1.22 161file sys/nfs/nfs_serv.c : 1.108 162file sys/nfs/nfs_socket.c : 1.129 163file sys/nfs/nfs_syscalls.c : 1.91 164file sys/ufs/lfs/lfs_vfsops.c : 1.207 [also (char *)] 165file usr.bin/ftp/ftp.c : 1.140 166file usr.bin/find/function.c : 1.54 167file usr.bin/mail/tty.c : 1.20 168file usr.bin/nl/nl.c : 1.7 169file usr.bin/systat/keyboard.c : 1.23 170file usr.bin/usbhidctl/usbhid.c : 1.29 171file usr.bin/vi/cl/cl_read.c : 1.5 172file usr.bin/vi/ex/ex_cscope.c : 1.12 173file usr.bin/vi/ex/ex_tag.c : 1.19 174file usr.bin/vi/vi/v_txt.c : 1.15 175file usr.sbin/altq/altqstat/qdisc_rio.c : 1.4 176file usr.sbin/cron/do_command.c : 1.19 177file usr.sbin/timed/timed/slave.c : 1.15 178descr 179 GCC 4.1 gets many uninitialised variable warnings wrong. We should 180 really audit all the old hacks like this when older compilers are 181 removed from the tree, as many are probably no longer required. 182kcah 183 184hack gcc4 pointer sign and strict aliasing problems 185mdate 10 May 2006 186who mrg 187file bin/ed/Makefile : 1.33 188file bin/systrace/Makefile : 1.12 189file games/bcd/Makefile : 1.5 190file games/dab/Makefile : 1.5 191file games/larn/Makefile : 1.17 192file games/pom/Makefile : 1.5 193file lib/libasn1/Makefile : 1.26 194file lib/libcrypt/Makefile : 1.17 195file lib/libgssapi/Makefile : 1.16 196file lib/libhdb/Makefile : 1.20 197file lib/libkadm5clnt/Makefile : 1.21 198file lib/libkadm5srv/Makefile : 1.25 199file lib/libkrb5/Makefile : 1.35 200file lib/libssh/Makefile : 1.6 201file lib/libtelnet/Makefile : 1.26 202file libexec/getty/Makefile : 1.14 203file libexec/kadmind/Makefile : 1.19 204file libexec/kpasswdd/Makefile : 1.14 205file sbin/atactl/Makefile : 1.3 206file sbin/cgdconfig/Makefile : 1.7 207file sbin/clri/Makefile : 1.13 208file sbin/dkctl/Makefile : 1.4 209file sbin/dump/Makefile : 1.33 210file sbin/fdisk/Makefile : 1.35 211file sbin/fsck_ext2fs/Makefile : 1.11 212file sbin/fsck_ffs/Makefile : 1.29 213file sbin/fsdb/Makefile : 1.18 214file sbin/mount_smbfs/Makefile : 1.4 215file sbin/newfs/Makefile : 1.30 216file sbin/newfs_sysvbfs/Makefile : 1.2 217file sbin/restore/Makefile : 1.23 218file sbin/veriexecctl/Makefile : 1.11 219file usr.bin/awk/Makefile : 1.9 220file usr.bin/ctags/Makefile : 1.8 221file usr.bin/gzip/Makefile : 1.10 222file usr.bin/ssh/sftp/Makefile : 1.10 223file usr.bin/ssh/ssh/Makefile : 1.25 224file usr.bin/vi/build/Makefile : 1.26 225file usr.bin/telnet/Makefile : 1.40 226file usr.bin/tn3270/tn3270/Makefile : 1.36 227file usr.bin/tr/Makefile : 1.4 228file usr.sbin/amd/amd/Makefile : 1.27 229file usr.sbin/amd/amq/Makefile : 1.14 230file usr.sbin/amd/libamu/Makefile : 1.20 231file usr.sbin/amd/pawd/Makefile : 1.5 232file usr.sbin/bind/Makefile.inc : 1.22 233file usr.sbin/bind/libdns/Makefile : 1.3 234file usr.sbin/bind/named/Makefile : 1.17 235file usr.sbin/bootp/bootptest/Makefile : 1.2 236file usr.sbin/chrtbl/Makefile : 1.6 237file usr.sbin/cron/Makefile : 1.12 238file usr.sbin/dhcp/Makefile.inc : 1.20 239file usr.sbin/dumplfs/Makefile : 1.11 240file usr.sbin/hprop/Makefile : 1.13 241file usr.sbin/installboot/Makefile : 1.35 242file usr.sbin/ipf/ipftest/Makefile : 1.32 243file usr.sbin/isdn/isdnd/Makefile : 1.6 244file usr.sbin/isdn/isdnmonitor/Makefile : 1.3 245file usr.sbin/isdn/isdntel/Makefile : 1.2 246file usr.sbin/isdn/isdntrace/Makefile : 126 247file usr.sbin/mopd/common/Makefile : 1.10 248file usr.sbin/mopd/mopd/Makefile : 1.9 249file usr.sbin/mopd/mopprobe/Makefile : 1.7 250file usr.sbin/makefs/Makefile : 1.17 251file usr.sbin/mscdlabel/Makefile : 1.5 252file usr.sbin/pkg_install/add/Makefile : 1.7 253file usr.sbin/pkg_install/create/Makefile : 1.5 254file usr.sbin/pkg_install/lib/Makefile : 1.28 255file usr.sbin/ntp/ntpd/Makefile : 1.10/1.11 256file usr.sbin/ntp/ntptime/Makefile : 1.4/1.5 257file usr.sbin/pppd/Makefile.inc : 1.3 258file usr.sbin/pppd/pppd/Makefile : 1.38 259file usr.sbin/rarpd/Makefile : 1.10 260file usr.sbin/rbootd/Makefile : 1.10 261file usr.sbin/rpc.pcnfsd/Makefile : 1.17 262file usr.sbin/rtadvd/Makefile : 1.9 263file usr.sbin/tcpdump/Makefile : 1.42 264file usr.sbin/wiconfig/Makefile : 1.3 265file usr.sbin/ypserv/ypserv/Makefile : 1.16 266descr 267 GCC 4.1 warns on pointer sign comparision/assignments and lots of 268 code does not conform. For now we use -Wno-pointer-sign and 269 -fno-strict-aliasing. 270kcah 271 272 273port vax 274 275 hack gcc 2.95/vax doesn't like abort being used as function pointer 276 cdate Tue Dec 13 05:54:50 GMT 2005 277 who jmc 278 file lib/libc/rpc/svc_vc.c 1.15 279 descr 280 Provide a local definition for abort which doesn't include 281 the noreturn attribute which trips up gcc 2.95 on vax. 282 283 hack gcc 2.95/vax doesn't have stdbool.h and gettext needs it 284 cdate Mon May 9 12:16:22 CDT 2005 285 who jmc 286 file gnu/usr.bin/gettext/Makefile.inc.prog 1.5 287 gnu/usr.bin/gettext/libgrep/Makefile 1.3 288 gnu/usr.bin/gettext/libnlspr/Makefile 1.5 289 gnu/usr.bin/gettext/libnlsut/Makefile 1.3 290 descr 291 The latest gettext assumes a C99 environment or at least 292 configure checks to work around this. As we don't use configure 293 provide a stdbool.h for vax builds locally in directories from 294 the template provided by the gettext code. 295 296 hack gcc 2.95/vax can't handle __used__ in mcount.c 297 cdate Fri May 6 14:30:53 CDT 2005 298 who jmc 299 file src/lib/libc/gmon/mcount.c 1.17 300 descr 301 Recent changes to mark this as __used__ to prevent optimization 302 away on sparc64 don't work on 2.95 vax. Ifdef'd back in the 303 old __unused__ if on vax. 304 305 hack gcc 2.95/vax cannot handle __builtin_ffs() 306 cdate Fri Aug 23 21:31:15 CEST 2002 307 who ragge 308 file sys/lib/libkern/libkern.h : 1.42 309 descr 310 __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 311 added this feature to GCC 3 so it can be removed when switching. 312 313 314 hack gcc-vax-libbz2 315 mdate 27 Jun 2002 316 who thorpej 317 file lib/libbz2/Makefile 318 descr 319 libbz2 is mis-compiled with optimization with GCC 2.95.3 320 on VAX. -O0 works around this problem. 321 kcah 322 323 hack gawk 324 mdate 8 Nov 2003 325 who ragge 326 file gnu/usr.bin/gawk/Makefile 327 descr 328 gawk is mis-compiled with optimization with GCC 2.95.3 329 on VAX. -O0 works around this problem. 330 kcah 331 332 hack vax Toolchain bug 333 cdate 28 Mar 2003 334 who he 335 file sys/arch/vax/vax/intvec.S : 1.5 336 descr 337 Workaround for PR toolchain/20924. The assembler apparently 338 tries to range-check byte offsets when it doesn't have 339 sufficient information to make that decision. 340 Workaround uses `brw' instead of `brb' instruction. 341 kcah 342 343 hack vax gcc 2.95.3 -Wuninitialized workarounds 344 cdate 6 Nov 2003 345 who he 346 file sys/dev/bi/if_ni.c : 1.22 347 file sys/arch/vax/vax/ctu.c : 1.20 348 descr 349 Workaround for gcc 2.95.3 failing to detect that certain 350 variables will be initialized; so that this code compiles 351 with -Wuninitialized. Should be reviewed and possibly 352 reverted when gcc 3.3.2 is ready for vax. 353 kcah 354 355 hack vax gcc 2.95.3 structure initialization 356 cdat 13 May 2004 357 who cl 358 file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 359 descr 360 Workaround for gcc 2.95.3 failing to initialize structures 361 and/or unions inside structures using nested designators. 362 Should be reverted when gcc >=3.3.3 is ready for vax. 363 kcah 364 365 hack vax gcc 2.95.3 needs -I. to build nslexer.c 366 cdat 6 Jun 2004 367 who mhitch 368 file lib/libc/arch/vax/Makefile.inc : 1.6 369 descr 370 It appears necessary to add -I. to CPPFLAGS when building 371 nslesxer.c. This may be caused by gcc 2.95.3 being used. 372 Should be reverted when a newer gcc is ready for vax. 373 kcah 374 375 hack vax gcc 2.95.3 doesn't understand "-std=c99" 376 cdat 5 April 2005 377 who tron 378 file src/share/mk/bsd.sys.mk 379 descr 380 We cannot use "-std=c99" for compiler warning level 4 and above 381 because gcc 2.95.3 doesn't support that option. 382 kcah 383 384 hack declare boolean_t in two IPF user-mode programs 385 cdate Tue Mar 7 19:19:20 CET 2006 386 who he 387 file dist/ipf/ipsend/iptests.c : 1.8 388 dist/ipf/ipsend/sock.c : 1.7 389 descr 390 The IPF user-mode programs ipsend and iptest first 391 include <sys/types.h> without _KERNEL defined, and 392 later include <sys/file.h> with _KERNEL defined. 393 This causes a build failure when building for vax, 394 since <sys/device.h> ends up being included without 395 bollean_t being defined by <sys/types.h>. 396 Build failure and further details documented in 397 PR#32907. 398 kcah 399 400port sparc64 401 402 hack binutil-2.11-sparc64-pltrela 403 mdate 14 Aug 2001 404 who eeh 405 file libexec/ld.elf_so/arch/sparc64/mdreloc.c 406 descr 407 The first four PLT entries are reserved. There is some 408 disagreement whether they should have associated relocation 409 entries. Both the SPARC 32-bit and 64-bit ELF specifications 410 say that they should have relocation entries, but the 32-bit 411 SPARC binutils do not generate them, and now the 64-bit SPARC 412 binutils have stopped generating them too. 413 414 To provide binary compatibility, we will check the first entry, 415 if it is reserved it should not be of the type JMP_SLOT. If it 416 is JMP_SLOT, then the 4 reserved entries were not generated and 417 our index is 4 entries too far, so we frob the rela pointer. 418 kcah 419 420 421port arm 422 423 hack gcc-unsigned-compare 424 cdate 09 Mar 2002 425 mdate 18 Mar 2002 426 who bjh21 427 file dist/bind/lib/nameser/ns_parse.c : 1.3 428 file dist/dhcp/minires/ns_parse.c : 1.3 429 file dist/dhcp/omapip/result.c : 1.2 430 file dist/dhcp/server/failover.c : 1.3 431 file gnu/dist/gawk/eval.c : 1.4 432 file gnu/dist/toolchain/bfd/bfd.c : 1.2 433 file gnu/dist/toolchain/bfd/format.c : 1.2 434 file gnu/dist/toolchain/gdb/target.c : 1.2 435 file sys/kern/vfs_subr.c : 1.172 436 descr When checking that a potentially-unsigned enum is >= 0, assign 437 it to an int first. This is necessary to avoid "comparison is 438 always true" warnings with -fshort-enums. Casting to an int 439 really should be enough, but turns out not to be. 440 kcah 441 442 hack gcc 3.3.2/arm sys/kern/uipc_socket.c 443 cdate Tue Oct 28 18:02:16 GMT 2003 444 who skrll 445 pr 23044 446 file sys/arch/arm/conf/Makefile.arm : 1.19 447 descr 448 Performing a build.sh tools on a CATS with NFS mounted 449 obj and src directories results in "panic: receive 1" 450 The hack is to compile sys/kern/uipc_socket.c with 451 -fno-strict-aliasing 452 kcah 453 454 455port pc532 456 457 hack egcs-pc532-ip6_mroute 458 cdate 09 Jul 2002 459 who simonb 460 file sys/arch/pc532/conf/Makefile.pc532 : 1.70 461 file sys/arch/pc532/conf/files.pc532 : 1.47 462 descr 463 egcs 1.1.2 gets an "internal error--insn does not satisfy its 464 constraints" error compiling ip6_mroute.c with -O2 or greater. 465 -O1 works around this problem. 466 kcah 467 468 hack gcc-pc532-duffs_device 469 cdate Mon Oct 27 07:23:05 UTC 2003 470 who simonb 471 file sys/arch/pc532/conf/Makefile.pc532 : 1.71 472 file sys/arch/pc532/conf/files.pc532 : 1.52 473 pr GCC PR optimization/5230 474 descr 475 gcc incorrectly gives an "unreachable code at beginning of 476 switch statement" for a Duff's device construct in 477 arch/pc532/dev/ncr.c. There is no way to disable just this 478 warning, so -Wno-error is turned on for this file. 479 kcah 480 481 hack gcc332-cppmacro-ice 482 cdate Sun Dec 7 12:48:33 UTC 2003 483 who simonb 484 file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 485 descr 486 gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 487 The hack is to use -O1 for this file. Fixed in gcc 488 -current mainline, unable to work out where the fix is 489 right now. 490 kcah 491 492 493port sh3 494 495 hack gcc-sh3-sed 496 mdate 23 Apr 2002 497 who thorpej 498 file usr.bin/sed/Makefile : 1.9 : 9 : 13 499 descr 500 The in-tree GCC 2.95.3-based compiler ICEs when building 501 with optimization for SuperH. Hack is to build with -O0. 502 kcah 503 504 505port sh5 506 507 hack SuperH SH5 Toolchain Bugs 508 cdate 11 Jul 2002 509 who scw 510 file usr.sbin/ndbootd/ndbootd.c : 1.6 511 file usr.sbin/traceroute/traceroute.c : 1.48 512 descr 513 The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 514 compiler error when assigning a bit-wise inverted value 515 under some circumstances. 516 Work around it by splitting the statement into two. 517 kcah 518 519 hack SuperH SH5 Toolchain Bugs 520 cdate 17 May 2003 521 who scw 522 file lib/libpthread/pthread_lock.c : 1.7 523 descr 524 The SuperH SH5 toolchain generates incorrect PIC code when 525 faced with a symbol which is declared extern, but has local 526 scope due to being defined within an asm() statement (without 527 being declared .globl in that statement). Work around it by 528 adding the .globl. 529 kcah 530 531 hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 532 cdate Sun Dec 7 19:44:05 GMT 2003 533 who scw 534 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 535 descr 536 Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 537 Drop to -O1 for now when building for sh5. 538 kcah 539 540 hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 541 cdate Sun Dec 7 21:10:46 GMT 2003 542 who scw 543 file gnu/lib/libbfd/Makefile : 1.39 544 descr 545 Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 546 and elf64.c at -O2. 547 Drop to -O1 for now when building for sh5. 548 kcah 549 550 551 hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 552 cdate Sun Dec 7 19:44:05 GMT 2003 553 who scw 554 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 555 descr 556 Gcc3 for sh5 issues a warning: 557 .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 558 mutually exclusive equal-tests is always 0. 559 This causes the build to fail because of -Werror. 560 Work-around by defining NOGCCERROR when building for sh5. 561 kcah 562 563 hack gcc 3.3.3 on sh5 - ICE compiling rbtdb.c and rbtdb64.c 564 cdate Thu Nov 11 00:35:33 UTC 2004 565 who he 566 file usr.sbin/bind/libdns/Makefile/1.2 567 descr 568 Gcc3 for sh5 gets an ICE compiling these files at -O2. 569 Drop to -O1 for now when building for sh5. 570 kcah 571 572 hack gcc 3.3.3 nb3 on sh5 - ICE compiling bsd-comp.c 573 cdate Thu May 25 23:05:20 UTC 2005 574 who he 575 file sys/lkm/net/bsdcomp/Makefile : 1.3 576 descr 577 Gcc3 for sh5 gets an ICE compiling this at -O2, 578 "could not split insn". Drop to -O1 for sh5. 579 kcah 580 581 hack gcc 3.3.3 nb3 on sh5 - ICE compiling strtod.c 582 cdate Tue Jan 31 10:51:52 GMT 2006 583 who scw 584 file lib/libc/gdtoa/Makefile.inc : 1.2 585 descr 586 Gcc3 for sh5 gets an ICE compiling this at -O2, 587 "could not split insn". Drop to -O1 for sh5. 588 kcah 589 590port sun2 591 592 hack gcc 3.3.2/mdsetimage 593 cdate Tue Oct 28 18:43:05 EST 2003 594 who fredette 595 file gnu/usr.sbin/mdsetimage/Makefile : 1.15 596 descr 597 mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 598 this hack uses -O0 to avoid the problem. 599 kcah 600 601 hack gcc 3.3.2/installboot/ffs.c 602 cdate Tue Oct 28 18:43:05 EST 2003 603 who fredette 604 file usr.sbin/installboot/Makefile : 1.26 605 descr 606 ffs.c causes cc1 to barf for as-yet unknown reasons. this 607 hack uses -O0 to avoid the problem. 608 kcah 609 610 611port m68000 612 613 hack gcc 3.3.2/gcc/ra-colorize.c 614 cdate Wed Nov 5 11:03:13 EST 2003 615 who mrg 616 file gnu/usr.bin/gcc3/backend/Makefile : 1.11 617 descr 618 ra-colorize.c causes a whole spew of unreferenced .L* local 619 symbols. using -O0 (or even -O2 reportedly) avoids the problem. 620 Note this has been fixed in 3.3.3 at least on m68k, 621 but it hasn't been confirmed on sun2/m68000. 622 (it's too slow even on emulator and unlikely to build 623 toolchains on sun2 with its maximum 4MB RAM) 624 kcah 625 626port sparc64 627 628 hack disable optimzations for uvm_bio.c on 32 bit kernels 629 cdate Sun Mar 21 14:14:04 MET 2004 630 who martin 631 file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 632 file sys/arch/sparc64/conf/files.sparc64 : 1.93 633 descr 634 The sparc compiler miscompiles uvm_bio.c when using 635 any optimization. This results in ubc_release panics. 636 kcah 637 638port macppc 639 640 hack remove inline from rotate_{left,right} 641 cdate Tue Feb 22 13:43:25 EST 2005 642 who christos 643 file src/dist/bind/lib/dns/rbt.c 644 descr 645 The macppc gcc-3.3 compiler miscompiles rbt.c when these 646 functions are inlined. The error is: 647 .* named[368]: rbt.c:1584: INSIST(child != (void *)0) >failed 648 .* named[368]: exiting (due to assertion failure) 649 This is reportedly fixed in gcc-3.4 650 kcah 651 652port hppa 653 654 hack hppa-gcc-3.3.3/grep/dfa.c 655 cdate Sun Aug 8 08:34:32 PDT 2004 656 who chs 657 file gnu/usr.bin/grep/grep/Makefile : 1.4 658 descr 659 gcc 3.3.3 generates bad code for dfa.c with -O2, 660 reduce optimization to -O1 to avoid the bug. 661 kcah 662 663 hack hppa-gcc-3.3.3 / make/buf.c 664 cdate Mon May 8 13:02:36 CEST 2006 665 who he 666 file usr.bin/make/Makefile : 1.34 667 descr 668 gcc 3.3.3 nb3 gets an "unrecognized insn" internal 669 compiler error when building this file with -O2, 670 reduce optimization to -O1 to avoid the bug. 671 kcah 672 673 hack hppa boot-from-disk memory corruption 674 cdate Wed May 18 06:56:07 PDT 2005 675 who chs 676 file src/sys/dev/scsipi/scsipiconf.h : 1.98 677 descr 678 There is some bug with DMA on hppa that corrupts scsipi_xfer 679 structures, but it only shows up when booting from disk. 680 For now, we'll add some padding to scsipi_xfer so that 681 the corrupted memory is not otherwise used. 682 kcah 683 684port mips 685 686 hack mips-shared-linker-load-address 687 cdate Fri Oct 7 08:33:10 UTC 2005 688 who simonb 689 file src/sys/kern/exec_elf32.c : 1.107 690 descr 691 With COMPAT_16 or previous enabled (which enables 692 ELF_INTERP_NON_RELOCATABLE) a recent ld.elf_so will 693 load and run at address 0. The check to fix this in 694 rev 1.107 only checks the first psection of the ELF 695 executable, which may not be loadable. A more correct 696 fix is to check the first loadable psection instead of 697 just the first psection. 698 kcah 699 700