HACKS revision 1.80
1# $NetBSD: HACKS,v 1.80 2006/06/25 14:24:45 tsutsui 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 gcc 3.3.3 wrong uninitialised variable 114mdate 12 May 2006 115who he 116file usr.bin/msgc/msg_sys.def : 1.33-1.34 117file distrib/utils/sysinst/aout2elf.c : 1.11 118descr 119 Gcc 3.3.3 gets some -Wuninitialized warnings wrong. This is 120 just one of many in the tree, marked with 121 /* XXX -Wuninitialized [arch] */ 122 in the source. 123kcah 124 125hack gcc4 wrong uninitialized variable 126mdate 10 May 2006 127who mrg 128file bin/ksh/eval.c : 1.6 129file bin/sh/histedit.c : 1.39 130file bin/sh/parser.c : 1.60 131file bin/systrace/systrace.c : 1.33 132file crypto/dist/heimdal/kdc/524.c : 1.10 133file crypto/dist/ssh/sftp.c : 1.20 134file crypto/dist/ssh/ssh-keysign.c : 1.11 135file dist/ipf/lib/hostname.c : 1.2 136file dist/ipf/tools/ipmon.c : 1.8 137file dist/ntp/ntpd/ntp_request.c : 1.4 138file dist/ntp/ntpd/refclock_shm.c : 1.4 139file dist/ntp/sntp/timing.c : 1.3 140file dist/pppd/pppstats/pppstats.c : 1.3 141file dist/smbfs/lib/smb/rap.c : 1.6 142file dist/tcpdump/print-zephyr.c : 1.5 143file distrib/utils/sysinst/aout2elf.c : 1.12 144file gnu/libexec/uucp/uucico/uucico.c : 1.6 145file lib/libc/citrus/citrus_csmapper.c : 1.6 146file lib/libc/citrus/citrus_pivot_factory.c : 1.5 147file lib/libc/inet/inet_cidr_ntop.c : 1.3 148file lib/libc/inet/inet_ntop.c : 1.3 149file lib/libc/stdio/vfwprintf.c : 1.8 150file libexec/ld.elf_so/arch/m68k/mdreloc.c : 1.20 151file libexec/ld.elf_so/arch/powerpc/ppc_reloc.c : 1.40 152file libexec/ld.elf_so/arch/sh3/mdreloc.c : 1.22 153file libexec/ld.elf_so/arch/sparc/mdreloc.c : 1.39 154file libexec/ld.elf_so/arch/sparc64/mdreloc.c : 1.39 155file libexec/ld.elf_so/arch/vax/mdreloc.c : 1.21 156file libexec/ld.elf_so/arch/x86_64/mdreloc.c : 1.27 157file sbin/fsck_ext2fs/dir.c : 1.19 158file sbin/routed/rtquery/rtquery.c : 1.18 159file sys/arch/amd64/amd64/pmap.c : 1.26 160file sys/arch/i386/pci/piixpcib.c : 1.4 161file sys/arch/m68k/m68k/pmap_motorola.c : 1.4 162file sys/crypto/cast128/cast128.c : 1.9 163file sys/ddb/db_command.c : 1.86 164file sys/dev/cardbus/cardbus_map.c : 1.21 165file sys/dev/fss.c : 1.25 166file sys/dev/ic/igsfb.c : 1.39 167file sys/dev/ic/mb86950.c : 1.5 168file sys/dev/ic/midway.c : 1.71 169file sys/dev/kttcp.c : 1.18 170file sys/dev/rasops/rasops_bitops.h : 1.9 171file sys/dev/pci/cmpci.c : 1.31 172file sys/dev/pci/machfb.c : 1.45 173file sys/dev/usb/ohci.c : 1.174 174file sys/dev/usb/uhci.c : 1.196 175file sys/dev/rasops/rasops_bitops.h : 1.9 176file sys/dist/ipf/netinet/ip_nat.c : 1.10 177file sys/dist/ipf/netinet/ip_rpcb_pxy.c : 1.8 178file sys/dist/pf/net/pf.c : 1.22 179file sys/fs/udf/udf_vnops.c : 1.4 180file sys/kern/kern_sig.c : 1.219 181file sys/kern/tty.c : 1.181 182file sys/net/bpf.c : 1.116 183file sys/net/zlib.c : 1.26 184file sys/netccitt/if_x25subr.c : 1.37 185file sys/netinet/in.c : 1.107 186file sys/netsmb/smb_smb.c : 1.27 187file sys/netsmb/smb_trantcp.c : 1.22 188file sys/nfs/nfs_serv.c : 1.108 189file sys/nfs/nfs_socket.c : 1.129 190file sys/nfs/nfs_syscalls.c : 1.91 191file sys/ufs/lfs/lfs_vfsops.c : 1.207 [also (char *)] 192file usr.bin/ftp/ftp.c : 1.140 193file usr.bin/find/function.c : 1.54 194file usr.bin/mail/tty.c : 1.20 195file usr.bin/nl/nl.c : 1.7 196file usr.bin/systat/keyboard.c : 1.23 197file usr.bin/usbhidctl/usbhid.c : 1.29 198file usr.bin/vi/cl/cl_read.c : 1.5 199file usr.bin/vi/ex/ex_cscope.c : 1.12 200file usr.bin/vi/ex/ex_tag.c : 1.19 201file usr.bin/vi/vi/v_txt.c : 1.15 202file usr.sbin/altq/altqstat/qdisc_rio.c : 1.4 203file usr.sbin/cron/do_command.c : 1.19 204file usr.sbin/timed/timed/slave.c : 1.15 205descr 206 GCC 4.1 gets many uninitialised variable warnings wrong. We should 207 really audit all the old hacks like this when older compilers are 208 removed from the tree, as many are probably no longer required. 209 The problem is that it does not recognize initialization via function 210 call pointer. I.e. 211 int p; 212 foo(&p); 213 does not mark p as initialized. 214kcah 215 216hack gcc4 pointer sign and strict aliasing problems 217mdate 10 May 2006 218who mrg 219file bin/ed/Makefile : 1.33 220file bin/systrace/Makefile : 1.12 221file distrib/utils/sysinst/Makefile.inc : 1.44 222file games/bcd/Makefile : 1.5 223file games/dab/Makefile : 1.5 224file games/larn/Makefile : 1.17 225file games/pom/Makefile : 1.5 226file lib/libasn1/Makefile : 1.26 227file lib/libcrypt/Makefile : 1.17 228file lib/libgssapi/Makefile : 1.16 229file lib/libhdb/Makefile : 1.20 230file lib/libkadm5clnt/Makefile : 1.21 231file lib/libkadm5srv/Makefile : 1.25 232file lib/libkrb5/Makefile : 1.35 233file lib/libssh/Makefile : 1.6 234file lib/libtelnet/Makefile : 1.26 235file libexec/getty/Makefile : 1.14 236file libexec/kadmind/Makefile : 1.19 237file libexec/kpasswdd/Makefile : 1.14 238file sbin/atactl/Makefile : 1.3 239file sbin/cgdconfig/Makefile : 1.7 240file sbin/clri/Makefile : 1.13 241file sbin/dkctl/Makefile : 1.4 242file sbin/dump/Makefile : 1.33 243file sbin/fdisk/Makefile : 1.35 244file sbin/fsck_ext2fs/Makefile : 1.11 245file sbin/fsck_ffs/Makefile : 1.29 246file sbin/fsdb/Makefile : 1.18 247file sbin/mount_smbfs/Makefile : 1.4 248file sbin/newfs/Makefile : 1.30 249file sbin/newfs_sysvbfs/Makefile : 1.2 250file sbin/restore/Makefile : 1.23 251file sbin/veriexecctl/Makefile : 1.11 252file sys/lib/libsa/Makefile : 1.59 253file sys/arch/evbarm/adi_brh/brh_machdep.c : 1.24 254file usr.bin/awk/Makefile : 1.9 255file usr.bin/crontab/Makefile : 1.24 256file usr.bin/ctags/Makefile : 1.8 257file usr.bin/gzip/Makefile : 1.10 258file usr.bin/ssh/sftp/Makefile : 1.10 259file usr.bin/ssh/ssh/Makefile : 1.25 260file usr.bin/vi/build/Makefile : 1.26 261file usr.bin/telnet/Makefile : 1.40 262file usr.bin/tn3270/tn3270/Makefile : 1.36 263file usr.bin/tr/Makefile : 1.4 264file usr.sbin/amd/amd/Makefile : 1.27 265file usr.sbin/amd/amq/Makefile : 1.14 266file usr.sbin/amd/libamu/Makefile : 1.20 267file usr.sbin/amd/pawd/Makefile : 1.5 268file usr.sbin/bind/Makefile.inc : 1.22 269file usr.sbin/bind/libdns/Makefile : 1.3 270file usr.sbin/bind/named/Makefile : 1.17 271file usr.sbin/bootp/bootptest/Makefile : 1.2 272file usr.sbin/chrtbl/Makefile : 1.6 273file usr.sbin/cron/Makefile : 1.12 274file usr.sbin/dhcp/Makefile.inc : 1.20 275file usr.sbin/dumplfs/Makefile : 1.11 276file usr.sbin/hprop/Makefile : 1.13 277file usr.sbin/installboot/Makefile : 1.35 278file usr.sbin/ipf/ipftest/Makefile : 1.32 279file usr.sbin/isdn/isdnd/Makefile : 1.6 280file usr.sbin/isdn/isdnmonitor/Makefile : 1.3 281file usr.sbin/isdn/isdntel/Makefile : 1.2 282file usr.sbin/isdn/isdntrace/Makefile : 126 283file usr.sbin/mopd/common/Makefile : 1.10 284file usr.sbin/mopd/mopd/Makefile : 1.9 285file usr.sbin/mopd/mopprobe/Makefile : 1.7 286file usr.sbin/makefs/Makefile : 1.17 287file usr.sbin/mscdlabel/Makefile : 1.5 288file usr.sbin/pkg_install/add/Makefile : 1.7 289file usr.sbin/pkg_install/create/Makefile : 1.5 290file usr.sbin/pkg_install/lib/Makefile : 1.28 291file usr.sbin/ntp/ntpd/Makefile : 1.10/1.11 292file usr.sbin/ntp/ntptime/Makefile : 1.4/1.5 293file usr.sbin/pppd/Makefile.inc : 1.3 294file usr.sbin/pppd/pppd/Makefile : 1.38 295file usr.sbin/rarpd/Makefile : 1.10 296file usr.sbin/rbootd/Makefile : 1.10 297file usr.sbin/rpc.pcnfsd/Makefile : 1.17 298file usr.sbin/rtadvd/Makefile : 1.9 299file usr.sbin/tcpdump/Makefile : 1.42 300file usr.sbin/wiconfig/Makefile : 1.3 301descr 302 GCC 4.1 warns on pointer sign comparision/assignments and lots of 303 code does not conform. For now we use -Wno-pointer-sign and 304 -fno-strict-aliasing. 305kcah 306 307port vax 308 309 hack gcc 2.95/vax doesn't like abort being used as function pointer 310 cdate Tue Dec 13 05:54:50 GMT 2005 311 who jmc 312 file lib/libc/rpc/svc_vc.c 1.15 313 descr 314 Provide a local definition for abort which doesn't include 315 the noreturn attribute which trips up gcc 2.95 on vax. 316 317 hack gcc 2.95/vax doesn't have stdbool.h and gettext needs it 318 cdate Mon May 9 12:16:22 CDT 2005 319 who jmc 320 file gnu/usr.bin/gettext/Makefile.inc.prog 1.5 321 gnu/usr.bin/gettext/libgrep/Makefile 1.3 322 gnu/usr.bin/gettext/libnlspr/Makefile 1.5 323 gnu/usr.bin/gettext/libnlsut/Makefile 1.3 324 descr 325 The latest gettext assumes a C99 environment or at least 326 configure checks to work around this. As we don't use configure 327 provide a stdbool.h for vax builds locally in directories from 328 the template provided by the gettext code. 329 330 hack gcc 2.95/vax can't handle __used__ 331 cdate Fri May 6 14:30:53 CDT 2005 332 who jmc 333 file src/lib/libc/gmon/mcount.c 1.17 334 src/usr.bin/config/main.c 1.11 335 descr 336 Recent changes to mark this as __used__ to prevent optimization 337 away on sparc64 don't work on 2.95 vax. Ifdef'd back in the 338 old __unused__ if on vax. 339 340 hack gcc 2.95/vax cannot handle __builtin_ffs() 341 cdate Fri Aug 23 21:31:15 CEST 2002 342 who ragge 343 file sys/lib/libkern/libkern.h : 1.42 344 descr 345 __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 346 added this feature to GCC 3 so it can be removed when switching. 347 348 349 hack gcc-vax-libbz2 350 mdate 27 Jun 2002 351 who thorpej 352 file lib/libbz2/Makefile 353 descr 354 libbz2 is mis-compiled with optimization with GCC 2.95.3 355 on VAX. -O0 works around this problem. 356 kcah 357 358 hack gawk 359 mdate 8 Nov 2003 360 who ragge 361 file gnu/usr.bin/gawk/Makefile 362 descr 363 gawk is mis-compiled with optimization with GCC 2.95.3 364 on VAX. -O0 works around this problem. 365 kcah 366 367 hack vax Toolchain bug 368 cdate 28 Mar 2003 369 who he 370 file sys/arch/vax/vax/intvec.S : 1.5 371 descr 372 Workaround for PR toolchain/20924. The assembler apparently 373 tries to range-check byte offsets when it doesn't have 374 sufficient information to make that decision. 375 Workaround uses `brw' instead of `brb' instruction. 376 kcah 377 378 hack vax gcc 2.95.3 -Wuninitialized workarounds 379 cdate 6 Nov 2003 380 who he 381 file sys/dev/bi/if_ni.c : 1.22 382 file sys/arch/vax/vax/ctu.c : 1.20 383 descr 384 Workaround for gcc 2.95.3 failing to detect that certain 385 variables will be initialized; so that this code compiles 386 with -Wuninitialized. Should be reviewed and possibly 387 reverted when gcc 3.3.2 is ready for vax. 388 kcah 389 390 hack vax gcc 2.95.3 structure initialization 391 cdat 13 May 2004 392 who cl 393 file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 394 descr 395 Workaround for gcc 2.95.3 failing to initialize structures 396 and/or unions inside structures using nested designators. 397 Should be reverted when gcc >=3.3.3 is ready for vax. 398 kcah 399 400 hack vax gcc 2.95.3 needs -I. to build nslexer.c 401 cdat 6 Jun 2004 402 who mhitch 403 file lib/libc/arch/vax/Makefile.inc : 1.6 404 descr 405 It appears necessary to add -I. to CPPFLAGS when building 406 nslesxer.c. This may be caused by gcc 2.95.3 being used. 407 Should be reverted when a newer gcc is ready for vax. 408 kcah 409 410 hack vax gcc 2.95.3 doesn't understand "-std=c99" 411 cdat 5 April 2005 412 who tron 413 file src/share/mk/bsd.sys.mk 414 descr 415 We cannot use "-std=c99" for compiler warning level 4 and above 416 because gcc 2.95.3 doesn't support that option. 417 kcah 418 419 hack declare boolean_t in two IPF user-mode programs 420 cdate Tue Mar 7 19:19:20 CET 2006 421 who he 422 file dist/ipf/ipsend/iptests.c : 1.8 423 dist/ipf/ipsend/sock.c : 1.7 424 descr 425 The IPF user-mode programs ipsend and iptest first 426 include <sys/types.h> without _KERNEL defined, and 427 later include <sys/file.h> with _KERNEL defined. 428 This causes a build failure when building for vax, 429 since <sys/device.h> ends up being included without 430 bollean_t being defined by <sys/types.h>. 431 Build failure and further details documented in 432 PR#32907. 433 kcah 434 435port sparc64 436 437 hack binutil-2.11-sparc64-pltrela 438 mdate 14 Aug 2001 439 who eeh 440 file libexec/ld.elf_so/arch/sparc64/mdreloc.c 441 descr 442 The first four PLT entries are reserved. There is some 443 disagreement whether they should have associated relocation 444 entries. Both the SPARC 32-bit and 64-bit ELF specifications 445 say that they should have relocation entries, but the 32-bit 446 SPARC binutils do not generate them, and now the 64-bit SPARC 447 binutils have stopped generating them too. 448 449 To provide binary compatibility, we will check the first entry, 450 if it is reserved it should not be of the type JMP_SLOT. If it 451 is JMP_SLOT, then the 4 reserved entries were not generated and 452 our index is 4 entries too far, so we frob the rela pointer. 453 kcah 454 455 456port arm 457 458 hack gcc-unsigned-compare 459 cdate 09 Mar 2002 460 mdate 18 Mar 2002 461 who bjh21 462 file dist/bind/lib/nameser/ns_parse.c : 1.3 463 file dist/dhcp/minires/ns_parse.c : 1.3 464 file dist/dhcp/omapip/result.c : 1.2 465 file dist/dhcp/server/failover.c : 1.3 466 file gnu/dist/gawk/eval.c : 1.4 467 file gnu/dist/toolchain/bfd/bfd.c : 1.2 468 file gnu/dist/toolchain/bfd/format.c : 1.2 469 file gnu/dist/toolchain/gdb/target.c : 1.2 470 file sys/kern/vfs_subr.c : 1.172 471 descr When checking that a potentially-unsigned enum is >= 0, assign 472 it to an int first. This is necessary to avoid "comparison is 473 always true" warnings with -fshort-enums. Casting to an int 474 really should be enough, but turns out not to be. 475 kcah 476 477 hack gcc 3.3.2/arm sys/kern/uipc_socket.c 478 cdate Tue Oct 28 18:02:16 GMT 2003 479 who skrll 480 pr 23044 481 file sys/arch/arm/conf/Makefile.arm : 1.19 482 descr 483 Performing a build.sh tools on a CATS with NFS mounted 484 obj and src directories results in "panic: receive 1" 485 The hack is to compile sys/kern/uipc_socket.c with 486 -fno-strict-aliasing 487 kcah 488 489 490port pc532 491 492 hack egcs-pc532-ip6_mroute 493 cdate 09 Jul 2002 494 who simonb 495 file sys/arch/pc532/conf/Makefile.pc532 : 1.70 496 file sys/arch/pc532/conf/files.pc532 : 1.47 497 descr 498 egcs 1.1.2 gets an "internal error--insn does not satisfy its 499 constraints" error compiling ip6_mroute.c with -O2 or greater. 500 -O1 works around this problem. 501 kcah 502 503 hack gcc-pc532-duffs_device 504 cdate Mon Oct 27 07:23:05 UTC 2003 505 who simonb 506 file sys/arch/pc532/conf/Makefile.pc532 : 1.71 507 file sys/arch/pc532/conf/files.pc532 : 1.52 508 pr GCC PR optimization/5230 509 descr 510 gcc incorrectly gives an "unreachable code at beginning of 511 switch statement" for a Duff's device construct in 512 arch/pc532/dev/ncr.c. There is no way to disable just this 513 warning, so -Wno-error is turned on for this file. 514 kcah 515 516 hack gcc332-cppmacro-ice 517 cdate Sun Dec 7 12:48:33 UTC 2003 518 who simonb 519 file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 520 descr 521 gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 522 The hack is to use -O1 for this file. Fixed in gcc 523 -current mainline, unable to work out where the fix is 524 right now. 525 kcah 526 527 528port sh3 529 530 hack gcc4-sh3-bz2 531 mdate 20 May 2006 532 who mrg 533 file lib/libbz2/Makefile : 1.10 534 descr 535 The in-tree GCC 4.1-based compiler generated too-far 536 pc-relative addresses. Hack is to build with -O0. 537 kcah 538 539 540port sh5 541 542 hack SuperH SH5 Toolchain Bugs 543 cdate 11 Jul 2002 544 who scw 545 file usr.sbin/ndbootd/ndbootd.c : 1.6 546 file usr.sbin/traceroute/traceroute.c : 1.48 547 descr 548 The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 549 compiler error when assigning a bit-wise inverted value 550 under some circumstances. 551 Work around it by splitting the statement into two. 552 kcah 553 554 hack SuperH SH5 Toolchain Bugs 555 cdate 17 May 2003 556 who scw 557 file lib/libpthread/pthread_lock.c : 1.7 558 descr 559 The SuperH SH5 toolchain generates incorrect PIC code when 560 faced with a symbol which is declared extern, but has local 561 scope due to being defined within an asm() statement (without 562 being declared .globl in that statement). Work around it by 563 adding the .globl. 564 kcah 565 566 hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 567 cdate Sun Dec 7 19:44:05 GMT 2003 568 who scw 569 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 570 descr 571 Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 572 Drop to -O1 for now when building for sh5. 573 kcah 574 575 hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 576 cdate Sun Dec 7 21:10:46 GMT 2003 577 who scw 578 file gnu/lib/libbfd/Makefile : 1.39 579 descr 580 Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 581 and elf64.c at -O2. 582 Drop to -O1 for now when building for sh5. 583 kcah 584 585 586 hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 587 cdate Sun Dec 7 19:44:05 GMT 2003 588 who scw 589 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 590 descr 591 Gcc3 for sh5 issues a warning: 592 .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 593 mutually exclusive equal-tests is always 0. 594 This causes the build to fail because of -Werror. 595 Work-around by defining NOGCCERROR when building for sh5. 596 kcah 597 598 hack gcc 3.3.3 on sh5 - ICE compiling rbtdb.c and rbtdb64.c 599 cdate Thu Nov 11 00:35:33 UTC 2004 600 who he 601 file usr.sbin/bind/libdns/Makefile/1.2 602 descr 603 Gcc3 for sh5 gets an ICE compiling these files at -O2. 604 Drop to -O1 for now when building for sh5. 605 kcah 606 607 hack gcc 3.3.3 nb3 on sh5 - ICE compiling bsd-comp.c 608 cdate Thu May 25 23:05:20 UTC 2005 609 who he 610 file sys/lkm/net/bsdcomp/Makefile : 1.3 611 descr 612 Gcc3 for sh5 gets an ICE compiling this at -O2, 613 "could not split insn". Drop to -O1 for sh5. 614 kcah 615 616 hack gcc 3.3.3 nb3 on sh5 - ICE compiling strtod.c 617 cdate Tue Jan 31 10:51:52 GMT 2006 618 who scw 619 file lib/libc/gdtoa/Makefile.inc : 1.2 620 descr 621 Gcc3 for sh5 gets an ICE compiling this at -O2, 622 "could not split insn". Drop to -O1 for sh5. 623 kcah 624 625port sun2 626 627 hack gcc 3.3.2/mdsetimage 628 cdate Tue Oct 28 18:43:05 EST 2003 629 who fredette 630 file gnu/usr.sbin/mdsetimage/Makefile : 1.15 631 descr 632 mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 633 this hack uses -O0 to avoid the problem. 634 kcah 635 636 hack gcc 3.3.2/installboot/ffs.c 637 cdate Tue Oct 28 18:43:05 EST 2003 638 who fredette 639 file usr.sbin/installboot/Makefile : 1.26 640 descr 641 ffs.c causes cc1 to barf for as-yet unknown reasons. this 642 hack uses -O0 to avoid the problem. 643 kcah 644 645 hack gcc 4.1.1/fsck_ffs/pass1.c 646 cdate Fri Jun 23 10:30:25 PDT 2006 647 who mrg 648 file sbin/fsck_ffs/Makefile : 1.30 649 file sbin/fsdb/Makefile : 1.19 650 file sbin/dump_lfs/Makefile : 1.9 651 file usr.sbin/ndbootd/Makefile : 1.2 652 descr 653 pass1.c causes cc1 to barf for as-yet unknown reasons. this 654 hack uses -O0 to avoid the problem. (these all seem to be the 655 same problem.) 656 kcah 657 658 659port m68000 660 661 hack gcc 3.3.2/gcc/ra-colorize.c 662 cdate Wed Nov 5 11:03:13 EST 2003 663 who mrg 664 file gnu/usr.bin/gcc3/backend/Makefile : 1.11 665 descr 666 ra-colorize.c causes a whole spew of unreferenced .L* local 667 symbols. using -O0 (or even -O2 reportedly) avoids the problem. 668 Note this has been fixed in 3.3.3 at least on m68k, 669 but it hasn't been confirmed on sun2/m68000. 670 (it's too slow even on emulator and unlikely to build 671 toolchains on sun2 with its maximum 4MB RAM) 672 kcah 673 674port sparc64 675 676 hack disable optimzations for uvm_bio.c on 32 bit kernels 677 cdate Sun Mar 21 14:14:04 MET 2004 678 who martin 679 file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 680 file sys/arch/sparc64/conf/files.sparc64 : 1.93 681 descr 682 The sparc compiler miscompiles uvm_bio.c when using 683 any optimization. This results in ubc_release panics. 684 kcah 685 686port macppc 687 688 hack remove inline from rotate_{left,right} 689 cdate Tue Feb 22 13:43:25 EST 2005 690 who christos 691 file src/dist/bind/lib/dns/rbt.c 692 descr 693 The macppc gcc-3.3 compiler miscompiles rbt.c when these 694 functions are inlined. The error is: 695 .* named[368]: rbt.c:1584: INSIST(child != (void *)0) >failed 696 .* named[368]: exiting (due to assertion failure) 697 This is reportedly fixed in gcc-3.4 698 kcah 699 700port hppa 701 702 hack hppa-gcc-3.3.3/grep/dfa.c 703 cdate Sun Aug 8 08:34:32 PDT 2004 704 who chs 705 file gnu/usr.bin/grep/grep/Makefile : 1.4 706 descr 707 gcc 3.3.3 generates bad code for dfa.c with -O2, 708 reduce optimization to -O1 to avoid the bug. 709 kcah 710 711 hack hppa boot-from-disk memory corruption 712 cdate Wed May 18 06:56:07 PDT 2005 713 who chs 714 file src/sys/dev/scsipi/scsipiconf.h : 1.98 715 descr 716 There is some bug with DMA on hppa that corrupts scsipi_xfer 717 structures, but it only shows up when booting from disk. 718 For now, we'll add some padding to scsipi_xfer so that 719 the corrupted memory is not otherwise used. 720 kcah 721 722port mips 723 724 hack mips-shared-linker-load-address 725 cdate Fri Oct 7 08:33:10 UTC 2005 726 who simonb 727 file src/sys/kern/exec_elf32.c : 1.107 728 descr 729 With COMPAT_16 or previous enabled (which enables 730 ELF_INTERP_NON_RELOCATABLE) a recent ld.elf_so will 731 load and run at address 0. The check to fix this in 732 rev 1.107 only checks the first psection of the ELF 733 executable, which may not be loadable. A more correct 734 fix is to check the first loadable psection instead of 735 just the first psection. 736 kcah 737 738