HACKS revision 1.71
1# $NetBSD: HACKS,v 1.71 2006/05/20 23:41:16 dan 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 uninitialised 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/refclock_shm.c : 1.4 138file dist/ntp/sntp/timing.c : 1.3 139file dist/pppd/pppstats/pppstats.c : 1.3 140file dist/smbfs/lib/smb/rap.c : 1.6 141file dist/tcpdump/print-zephyr.c : 1.5 142file distrib/utils/sysinst/aout2elf.c : 1.12 143file gnu/libexec/uucp/uucico/uucico.c : 1.6 144file lib/libc/citrus/citrus_csmapper.c : 1.6 145file lib/libc/citrus/citrus_pivot_factory.c : 1.5 146file lib/libc/inet/inet_cidr_ntop.c : 1.3 147file lib/libc/inet/inet_ntop.c : 1.3 148file lib/libc/stdio/vfwprintf.c : 1.8 149file libexec/ld.elf_so/arch/m68k/mdreloc.c : 1.20 150file libexec/ld.elf_so/arch/sparc/mdreloc.c : 1.39 151file libexec/ld.elf_so/arch/sparc64/mdreloc.c : 1.39 152file libexec/ld.elf_so/arch/x86_64/mdreloc.c : 1.27 153file sbin/fsck_ext2fs/dir.c : 1.19 154file sbin/routed/rtquery/rtquery.c : 1.18 155file sys/crypto/cast128/cast128.c : 1.9 156file sys/ddb/db_command.c : 1.86 157file sys/dev/cardbus/cardbus_map.c : 1.21 158file sys/dev/fss.c : 1.25 159file sys/dev/ic/igsfb.c : 1.39 160file sys/dev/ic/mb86950.c : 1.5 161file sys/dev/ic/midway.c : 1.71 162file sys/dev/kttcp.c : 1.18 163file sys/dev/rasops/rasops_bitops.h : 1.9 164file sys/dev/pci/cmpci.c : 1.31 165file sys/dev/pci/machfb.c : 1.45 166file sys/dev/usb/ohci.c : 1.174 167file sys/dev/usb/uhci.c : 1.196 168file sys/dev/rasops/rasops_bitops.h : 1.9 169file sys/dist/ipf/netinet/ip_nat.c : 1.10 170file sys/dist/ipf/netinet/ip_rpcb_pxy.c : 1.8 171file sys/dist/pf/net/pf.c : 1.22 172file sys/fs/udf/udf_vnops.c : 1.4 173file sys/kern/kern_sig.c : 1.219 174file sys/kern/tty.c : 1.181 175file sys/net/bpf.c : 1.116 176file sys/net/zlib.c : 1.26 177file sys/netccitt/if_x25subr.c : 1.37 178file sys/netinet/in.c : 1.107 179file sys/netsmb/smb_smb.c : 1.27 180file sys/netsmb/smb_trantcp.c : 1.22 181file sys/nfs/nfs_serv.c : 1.108 182file sys/nfs/nfs_socket.c : 1.129 183file sys/nfs/nfs_syscalls.c : 1.91 184file sys/ufs/lfs/lfs_vfsops.c : 1.207 [also (char *)] 185file usr.bin/ftp/ftp.c : 1.140 186file usr.bin/find/function.c : 1.54 187file usr.bin/mail/tty.c : 1.20 188file usr.bin/nl/nl.c : 1.7 189file usr.bin/systat/keyboard.c : 1.23 190file usr.bin/usbhidctl/usbhid.c : 1.29 191file usr.bin/vi/cl/cl_read.c : 1.5 192file usr.bin/vi/ex/ex_cscope.c : 1.12 193file usr.bin/vi/ex/ex_tag.c : 1.19 194file usr.bin/vi/vi/v_txt.c : 1.15 195file usr.sbin/altq/altqstat/qdisc_rio.c : 1.4 196file usr.sbin/cron/do_command.c : 1.19 197file usr.sbin/timed/timed/slave.c : 1.15 198descr 199 GCC 4.1 gets many uninitialised variable warnings wrong. We should 200 really audit all the old hacks like this when older compilers are 201 removed from the tree, as many are probably no longer required. 202 The problem is that it does not recognize initialization via function 203 call pointer. I.e. 204 int p; 205 foo(&p); 206 does not mark p as initialized. 207kcah 208 209hack gcc4 pointer sign and strict aliasing problems 210mdate 10 May 2006 211who mrg 212file bin/ed/Makefile : 1.33 213file bin/systrace/Makefile : 1.12 214file distrib/utils/sysinst/Makefile.inc : 1.44 215file games/bcd/Makefile : 1.5 216file games/dab/Makefile : 1.5 217file games/larn/Makefile : 1.17 218file games/pom/Makefile : 1.5 219file lib/libasn1/Makefile : 1.26 220file lib/libcrypt/Makefile : 1.17 221file lib/libgssapi/Makefile : 1.16 222file lib/libhdb/Makefile : 1.20 223file lib/libkadm5clnt/Makefile : 1.21 224file lib/libkadm5srv/Makefile : 1.25 225file lib/libkrb5/Makefile : 1.35 226file lib/libssh/Makefile : 1.6 227file lib/libtelnet/Makefile : 1.26 228file libexec/getty/Makefile : 1.14 229file libexec/kadmind/Makefile : 1.19 230file libexec/kpasswdd/Makefile : 1.14 231file sbin/atactl/Makefile : 1.3 232file sbin/cgdconfig/Makefile : 1.7 233file sbin/clri/Makefile : 1.13 234file sbin/dkctl/Makefile : 1.4 235file sbin/dump/Makefile : 1.33 236file sbin/fdisk/Makefile : 1.35 237file sbin/fsck_ext2fs/Makefile : 1.11 238file sbin/fsck_ffs/Makefile : 1.29 239file sbin/fsdb/Makefile : 1.18 240file sbin/mount_smbfs/Makefile : 1.4 241file sbin/newfs/Makefile : 1.30 242file sbin/newfs_sysvbfs/Makefile : 1.2 243file sbin/restore/Makefile : 1.23 244file sbin/veriexecctl/Makefile : 1.11 245file sys/lib/libsa/Makefile : 1.59 246file sys/arch/evbarm/adi_brh/brh_machdep.c : 1.24 247file usr.bin/awk/Makefile : 1.9 248file usr.bin/crontab/Makefile : 1.24 249file usr.bin/ctags/Makefile : 1.8 250file usr.bin/gzip/Makefile : 1.10 251file usr.bin/ssh/sftp/Makefile : 1.10 252file usr.bin/ssh/ssh/Makefile : 1.25 253file usr.bin/vi/build/Makefile : 1.26 254file usr.bin/telnet/Makefile : 1.40 255file usr.bin/tn3270/tn3270/Makefile : 1.36 256file usr.bin/tr/Makefile : 1.4 257file usr.sbin/amd/amd/Makefile : 1.27 258file usr.sbin/amd/amq/Makefile : 1.14 259file usr.sbin/amd/libamu/Makefile : 1.20 260file usr.sbin/amd/pawd/Makefile : 1.5 261file usr.sbin/bind/Makefile.inc : 1.22 262file usr.sbin/bind/libdns/Makefile : 1.3 263file usr.sbin/bind/named/Makefile : 1.17 264file usr.sbin/bootp/bootptest/Makefile : 1.2 265file usr.sbin/chrtbl/Makefile : 1.6 266file usr.sbin/cron/Makefile : 1.12 267file usr.sbin/dhcp/Makefile.inc : 1.20 268file usr.sbin/dumplfs/Makefile : 1.11 269file usr.sbin/hprop/Makefile : 1.13 270file usr.sbin/installboot/Makefile : 1.35 271file usr.sbin/ipf/ipftest/Makefile : 1.32 272file usr.sbin/isdn/isdnd/Makefile : 1.6 273file usr.sbin/isdn/isdnmonitor/Makefile : 1.3 274file usr.sbin/isdn/isdntel/Makefile : 1.2 275file usr.sbin/isdn/isdntrace/Makefile : 126 276file usr.sbin/mopd/common/Makefile : 1.10 277file usr.sbin/mopd/mopd/Makefile : 1.9 278file usr.sbin/mopd/mopprobe/Makefile : 1.7 279file usr.sbin/makefs/Makefile : 1.17 280file usr.sbin/mscdlabel/Makefile : 1.5 281file usr.sbin/pkg_install/add/Makefile : 1.7 282file usr.sbin/pkg_install/create/Makefile : 1.5 283file usr.sbin/pkg_install/lib/Makefile : 1.28 284file usr.sbin/ntp/ntpd/Makefile : 1.10/1.11 285file usr.sbin/ntp/ntptime/Makefile : 1.4/1.5 286file usr.sbin/pppd/Makefile.inc : 1.3 287file usr.sbin/pppd/pppd/Makefile : 1.38 288file usr.sbin/rarpd/Makefile : 1.10 289file usr.sbin/rbootd/Makefile : 1.10 290file usr.sbin/rpc.pcnfsd/Makefile : 1.17 291file usr.sbin/rtadvd/Makefile : 1.9 292file usr.sbin/tcpdump/Makefile : 1.42 293file usr.sbin/wiconfig/Makefile : 1.3 294descr 295 GCC 4.1 warns on pointer sign comparision/assignments and lots of 296 code does not conform. For now we use -Wno-pointer-sign and 297 -fno-strict-aliasing. 298kcah 299 300port vax 301 302 hack gcc 2.95/vax doesn't like abort being used as function pointer 303 cdate Tue Dec 13 05:54:50 GMT 2005 304 who jmc 305 file lib/libc/rpc/svc_vc.c 1.15 306 descr 307 Provide a local definition for abort which doesn't include 308 the noreturn attribute which trips up gcc 2.95 on vax. 309 310 hack gcc 2.95/vax doesn't have stdbool.h and gettext needs it 311 cdate Mon May 9 12:16:22 CDT 2005 312 who jmc 313 file gnu/usr.bin/gettext/Makefile.inc.prog 1.5 314 gnu/usr.bin/gettext/libgrep/Makefile 1.3 315 gnu/usr.bin/gettext/libnlspr/Makefile 1.5 316 gnu/usr.bin/gettext/libnlsut/Makefile 1.3 317 descr 318 The latest gettext assumes a C99 environment or at least 319 configure checks to work around this. As we don't use configure 320 provide a stdbool.h for vax builds locally in directories from 321 the template provided by the gettext code. 322 323 hack gcc 2.95/vax can't handle __used__ in mcount.c 324 cdate Fri May 6 14:30:53 CDT 2005 325 who jmc 326 file src/lib/libc/gmon/mcount.c 1.17 327 descr 328 Recent changes to mark this as __used__ to prevent optimization 329 away on sparc64 don't work on 2.95 vax. Ifdef'd back in the 330 old __unused__ if on vax. 331 332 hack gcc 2.95/vax cannot handle __builtin_ffs() 333 cdate Fri Aug 23 21:31:15 CEST 2002 334 who ragge 335 file sys/lib/libkern/libkern.h : 1.42 336 descr 337 __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 338 added this feature to GCC 3 so it can be removed when switching. 339 340 341 hack gcc-vax-libbz2 342 mdate 27 Jun 2002 343 who thorpej 344 file lib/libbz2/Makefile 345 descr 346 libbz2 is mis-compiled with optimization with GCC 2.95.3 347 on VAX. -O0 works around this problem. 348 kcah 349 350 hack gawk 351 mdate 8 Nov 2003 352 who ragge 353 file gnu/usr.bin/gawk/Makefile 354 descr 355 gawk is mis-compiled with optimization with GCC 2.95.3 356 on VAX. -O0 works around this problem. 357 kcah 358 359 hack vax Toolchain bug 360 cdate 28 Mar 2003 361 who he 362 file sys/arch/vax/vax/intvec.S : 1.5 363 descr 364 Workaround for PR toolchain/20924. The assembler apparently 365 tries to range-check byte offsets when it doesn't have 366 sufficient information to make that decision. 367 Workaround uses `brw' instead of `brb' instruction. 368 kcah 369 370 hack vax gcc 2.95.3 -Wuninitialized workarounds 371 cdate 6 Nov 2003 372 who he 373 file sys/dev/bi/if_ni.c : 1.22 374 file sys/arch/vax/vax/ctu.c : 1.20 375 descr 376 Workaround for gcc 2.95.3 failing to detect that certain 377 variables will be initialized; so that this code compiles 378 with -Wuninitialized. Should be reviewed and possibly 379 reverted when gcc 3.3.2 is ready for vax. 380 kcah 381 382 hack vax gcc 2.95.3 structure initialization 383 cdat 13 May 2004 384 who cl 385 file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 386 descr 387 Workaround for gcc 2.95.3 failing to initialize structures 388 and/or unions inside structures using nested designators. 389 Should be reverted when gcc >=3.3.3 is ready for vax. 390 kcah 391 392 hack vax gcc 2.95.3 needs -I. to build nslexer.c 393 cdat 6 Jun 2004 394 who mhitch 395 file lib/libc/arch/vax/Makefile.inc : 1.6 396 descr 397 It appears necessary to add -I. to CPPFLAGS when building 398 nslesxer.c. This may be caused by gcc 2.95.3 being used. 399 Should be reverted when a newer gcc is ready for vax. 400 kcah 401 402 hack vax gcc 2.95.3 doesn't understand "-std=c99" 403 cdat 5 April 2005 404 who tron 405 file src/share/mk/bsd.sys.mk 406 descr 407 We cannot use "-std=c99" for compiler warning level 4 and above 408 because gcc 2.95.3 doesn't support that option. 409 kcah 410 411 hack declare boolean_t in two IPF user-mode programs 412 cdate Tue Mar 7 19:19:20 CET 2006 413 who he 414 file dist/ipf/ipsend/iptests.c : 1.8 415 dist/ipf/ipsend/sock.c : 1.7 416 descr 417 The IPF user-mode programs ipsend and iptest first 418 include <sys/types.h> without _KERNEL defined, and 419 later include <sys/file.h> with _KERNEL defined. 420 This causes a build failure when building for vax, 421 since <sys/device.h> ends up being included without 422 bollean_t being defined by <sys/types.h>. 423 Build failure and further details documented in 424 PR#32907. 425 kcah 426 427port sparc64 428 429 hack binutil-2.11-sparc64-pltrela 430 mdate 14 Aug 2001 431 who eeh 432 file libexec/ld.elf_so/arch/sparc64/mdreloc.c 433 descr 434 The first four PLT entries are reserved. There is some 435 disagreement whether they should have associated relocation 436 entries. Both the SPARC 32-bit and 64-bit ELF specifications 437 say that they should have relocation entries, but the 32-bit 438 SPARC binutils do not generate them, and now the 64-bit SPARC 439 binutils have stopped generating them too. 440 441 To provide binary compatibility, we will check the first entry, 442 if it is reserved it should not be of the type JMP_SLOT. If it 443 is JMP_SLOT, then the 4 reserved entries were not generated and 444 our index is 4 entries too far, so we frob the rela pointer. 445 kcah 446 447 448port arm 449 450 hack gcc-unsigned-compare 451 cdate 09 Mar 2002 452 mdate 18 Mar 2002 453 who bjh21 454 file dist/bind/lib/nameser/ns_parse.c : 1.3 455 file dist/dhcp/minires/ns_parse.c : 1.3 456 file dist/dhcp/omapip/result.c : 1.2 457 file dist/dhcp/server/failover.c : 1.3 458 file gnu/dist/gawk/eval.c : 1.4 459 file gnu/dist/toolchain/bfd/bfd.c : 1.2 460 file gnu/dist/toolchain/bfd/format.c : 1.2 461 file gnu/dist/toolchain/gdb/target.c : 1.2 462 file sys/kern/vfs_subr.c : 1.172 463 descr When checking that a potentially-unsigned enum is >= 0, assign 464 it to an int first. This is necessary to avoid "comparison is 465 always true" warnings with -fshort-enums. Casting to an int 466 really should be enough, but turns out not to be. 467 kcah 468 469 hack gcc 3.3.2/arm sys/kern/uipc_socket.c 470 cdate Tue Oct 28 18:02:16 GMT 2003 471 who skrll 472 pr 23044 473 file sys/arch/arm/conf/Makefile.arm : 1.19 474 descr 475 Performing a build.sh tools on a CATS with NFS mounted 476 obj and src directories results in "panic: receive 1" 477 The hack is to compile sys/kern/uipc_socket.c with 478 -fno-strict-aliasing 479 kcah 480 481 482port pc532 483 484 hack egcs-pc532-ip6_mroute 485 cdate 09 Jul 2002 486 who simonb 487 file sys/arch/pc532/conf/Makefile.pc532 : 1.70 488 file sys/arch/pc532/conf/files.pc532 : 1.47 489 descr 490 egcs 1.1.2 gets an "internal error--insn does not satisfy its 491 constraints" error compiling ip6_mroute.c with -O2 or greater. 492 -O1 works around this problem. 493 kcah 494 495 hack gcc-pc532-duffs_device 496 cdate Mon Oct 27 07:23:05 UTC 2003 497 who simonb 498 file sys/arch/pc532/conf/Makefile.pc532 : 1.71 499 file sys/arch/pc532/conf/files.pc532 : 1.52 500 pr GCC PR optimization/5230 501 descr 502 gcc incorrectly gives an "unreachable code at beginning of 503 switch statement" for a Duff's device construct in 504 arch/pc532/dev/ncr.c. There is no way to disable just this 505 warning, so -Wno-error is turned on for this file. 506 kcah 507 508 hack gcc332-cppmacro-ice 509 cdate Sun Dec 7 12:48:33 UTC 2003 510 who simonb 511 file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 512 descr 513 gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 514 The hack is to use -O1 for this file. Fixed in gcc 515 -current mainline, unable to work out where the fix is 516 right now. 517 kcah 518 519 520port sh3 521 522 hack gcc-sh3-sed 523 mdate 23 Apr 2002 524 who thorpej 525 file usr.bin/sed/Makefile : 1.9 : 9 : 13 526 descr 527 The in-tree GCC 2.95.3-based compiler ICEs when building 528 with optimization for SuperH. Hack is to build with -O0. 529 kcah 530 531 532port sh5 533 534 hack SuperH SH5 Toolchain Bugs 535 cdate 11 Jul 2002 536 who scw 537 file usr.sbin/ndbootd/ndbootd.c : 1.6 538 file usr.sbin/traceroute/traceroute.c : 1.48 539 descr 540 The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 541 compiler error when assigning a bit-wise inverted value 542 under some circumstances. 543 Work around it by splitting the statement into two. 544 kcah 545 546 hack SuperH SH5 Toolchain Bugs 547 cdate 17 May 2003 548 who scw 549 file lib/libpthread/pthread_lock.c : 1.7 550 descr 551 The SuperH SH5 toolchain generates incorrect PIC code when 552 faced with a symbol which is declared extern, but has local 553 scope due to being defined within an asm() statement (without 554 being declared .globl in that statement). Work around it by 555 adding the .globl. 556 kcah 557 558 hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 559 cdate Sun Dec 7 19:44:05 GMT 2003 560 who scw 561 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 562 descr 563 Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 564 Drop to -O1 for now when building for sh5. 565 kcah 566 567 hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 568 cdate Sun Dec 7 21:10:46 GMT 2003 569 who scw 570 file gnu/lib/libbfd/Makefile : 1.39 571 descr 572 Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 573 and elf64.c at -O2. 574 Drop to -O1 for now when building for sh5. 575 kcah 576 577 578 hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 579 cdate Sun Dec 7 19:44:05 GMT 2003 580 who scw 581 file gnu/usr.bin/gcc3/backend/Makefile : 1.13 582 descr 583 Gcc3 for sh5 issues a warning: 584 .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 585 mutually exclusive equal-tests is always 0. 586 This causes the build to fail because of -Werror. 587 Work-around by defining NOGCCERROR when building for sh5. 588 kcah 589 590 hack gcc 3.3.3 on sh5 - ICE compiling rbtdb.c and rbtdb64.c 591 cdate Thu Nov 11 00:35:33 UTC 2004 592 who he 593 file usr.sbin/bind/libdns/Makefile/1.2 594 descr 595 Gcc3 for sh5 gets an ICE compiling these files at -O2. 596 Drop to -O1 for now when building for sh5. 597 kcah 598 599 hack gcc 3.3.3 nb3 on sh5 - ICE compiling bsd-comp.c 600 cdate Thu May 25 23:05:20 UTC 2005 601 who he 602 file sys/lkm/net/bsdcomp/Makefile : 1.3 603 descr 604 Gcc3 for sh5 gets an ICE compiling this at -O2, 605 "could not split insn". Drop to -O1 for sh5. 606 kcah 607 608 hack gcc 3.3.3 nb3 on sh5 - ICE compiling strtod.c 609 cdate Tue Jan 31 10:51:52 GMT 2006 610 who scw 611 file lib/libc/gdtoa/Makefile.inc : 1.2 612 descr 613 Gcc3 for sh5 gets an ICE compiling this at -O2, 614 "could not split insn". Drop to -O1 for sh5. 615 kcah 616 617port sun2 618 619 hack gcc 3.3.2/mdsetimage 620 cdate Tue Oct 28 18:43:05 EST 2003 621 who fredette 622 file gnu/usr.sbin/mdsetimage/Makefile : 1.15 623 descr 624 mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 625 this hack uses -O0 to avoid the problem. 626 kcah 627 628 hack gcc 3.3.2/installboot/ffs.c 629 cdate Tue Oct 28 18:43:05 EST 2003 630 who fredette 631 file usr.sbin/installboot/Makefile : 1.26 632 descr 633 ffs.c causes cc1 to barf for as-yet unknown reasons. this 634 hack uses -O0 to avoid the problem. 635 kcah 636 637 638port m68000 639 640 hack gcc 3.3.2/gcc/ra-colorize.c 641 cdate Wed Nov 5 11:03:13 EST 2003 642 who mrg 643 file gnu/usr.bin/gcc3/backend/Makefile : 1.11 644 descr 645 ra-colorize.c causes a whole spew of unreferenced .L* local 646 symbols. using -O0 (or even -O2 reportedly) avoids the problem. 647 Note this has been fixed in 3.3.3 at least on m68k, 648 but it hasn't been confirmed on sun2/m68000. 649 (it's too slow even on emulator and unlikely to build 650 toolchains on sun2 with its maximum 4MB RAM) 651 kcah 652 653port sparc64 654 655 hack disable optimzations for uvm_bio.c on 32 bit kernels 656 cdate Sun Mar 21 14:14:04 MET 2004 657 who martin 658 file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 659 file sys/arch/sparc64/conf/files.sparc64 : 1.93 660 descr 661 The sparc compiler miscompiles uvm_bio.c when using 662 any optimization. This results in ubc_release panics. 663 kcah 664 665port macppc 666 667 hack remove inline from rotate_{left,right} 668 cdate Tue Feb 22 13:43:25 EST 2005 669 who christos 670 file src/dist/bind/lib/dns/rbt.c 671 descr 672 The macppc gcc-3.3 compiler miscompiles rbt.c when these 673 functions are inlined. The error is: 674 .* named[368]: rbt.c:1584: INSIST(child != (void *)0) >failed 675 .* named[368]: exiting (due to assertion failure) 676 This is reportedly fixed in gcc-3.4 677 kcah 678 679port hppa 680 681 hack hppa-gcc-3.3.3/grep/dfa.c 682 cdate Sun Aug 8 08:34:32 PDT 2004 683 who chs 684 file gnu/usr.bin/grep/grep/Makefile : 1.4 685 descr 686 gcc 3.3.3 generates bad code for dfa.c with -O2, 687 reduce optimization to -O1 to avoid the bug. 688 kcah 689 690 hack hppa-gcc-3.3.3 / make/buf.c 691 cdate Mon May 8 13:02:36 CEST 2006 692 who he 693 file usr.bin/make/Makefile : 1.34 694 descr 695 gcc 3.3.3 nb3 gets an "unrecognized insn" internal 696 compiler error when building this file with -O2, 697 reduce optimization to -O1 to avoid the bug. 698 kcah 699 700 hack hppa boot-from-disk memory corruption 701 cdate Wed May 18 06:56:07 PDT 2005 702 who chs 703 file src/sys/dev/scsipi/scsipiconf.h : 1.98 704 descr 705 There is some bug with DMA on hppa that corrupts scsipi_xfer 706 structures, but it only shows up when booting from disk. 707 For now, we'll add some padding to scsipi_xfer so that 708 the corrupted memory is not otherwise used. 709 kcah 710 711port mips 712 713 hack mips-shared-linker-load-address 714 cdate Fri Oct 7 08:33:10 UTC 2005 715 who simonb 716 file src/sys/kern/exec_elf32.c : 1.107 717 descr 718 With COMPAT_16 or previous enabled (which enables 719 ELF_INTERP_NON_RELOCATABLE) a recent ld.elf_so will 720 load and run at address 0. The check to fix this in 721 rev 1.107 only checks the first psection of the ELF 722 executable, which may not be loadable. A more correct 723 fix is to check the first loadable psection instead of 724 just the first psection. 725 kcah 726 727