HACKS revision 1.34
11.34Scl# $NetBSD: HACKS,v 1.34 2004/05/13 15:44:04 cl Exp $ 21.1Slukem# 31.1Slukem# This file is intended to document workarounds for currently unsolved 41.1Slukem# (mostly) compiler bugs. 51.1Slukem# 61.1Slukem# Format: 71.1Slukem# hack title 81.1Slukem# cdate creation date 91.1Slukem# mdate mod date 101.1Slukem# who responsible developer 111.1Slukem# port ... 121.1Slukem# affected ports, space separated, if not "all" 131.1Slukem# file affected file : revision : line from : line to 141.1Slukem# affected files and revision and line numbers describing hack 151.1Slukem# multiple lines if necessary. 161.1Slukem# pr NNNN ... 171.1Slukem# problem reports this hack works around, if known. Space 181.1Slukem# separated. 191.1Slukem# regress src/regress/directory/where/test/found 201.1Slukem# regression test directories, if available. 211.1Slukem# descr 221.1Slukem# insert short informal description (multi-line). (Longer ones 231.1Slukem# should be in the PR database. More formal descriptions might 241.1Slukem# be in the regress tree. See above). 251.1Slukem# kcah 261.1Slukem# closing bracket. 271.1Slukem# 281.1Slukem# this is a comment. 291.1Slukem 301.1Slukemhack netstat ieee1394 address printing. 311.1Slukemmdate 14 Nov 2000 321.1Slukemwho matt 331.1Slukemfile lib/libc/net/getnameinfo.c : 1.32 : 497 : 503 341.1Slukemdescr 351.1Slukem Because the current implementation of IP over IEEE1394, the 361.1Slukem fw device address contains more than just the IEEE1394 EUI-64. 371.1Slukem So when printing out IEEE1394 addresses, ignore the extra stuff. 381.1Slukemkcah 391.1Slukem 401.1Slukemhack xterm vs. libterm 411.1Slukemmdate 01 Aug 2000 421.1Slukemwho jdc 431.1Slukemfile xsrc/xc/programs/xterm/main.c : 1.2 : 3609 : 3614 441.1Slukempr 10383 451.1Slukemdescr 461.1Slukem In order to extend the termcap string over 1023 bytes, a ZZ entry was 471.1Slukem introduced to point to a memory location containing the full entry. 481.1Slukem Without this hack, xterm will export a termcap containing the ZZ 491.1Slukem entry, which will then be ignored by libterm. As xterm modifies the 501.1Slukem exported termcap, this would cause those modifications to be ignored. 511.1Slukemkcah 521.1Slukem 531.1Slukemhack wi-at-big-endian-bus 541.1Slukemcdate 15 Mar 2002 551.1Slukemwho martin 561.1Slukemfile dev/ic/wireg.h 571.1Slukemdescr Add an option to access the underlying bus in big endian byte order 581.1Slukem to work around deficiencies in bus_space_{read,write}_* macros. 591.1Slukem Those don't allow the implementation of a proper pcmcia bus space 601.1Slukem tag. 611.1Slukemkcah 621.1Slukem 631.3Sjdchack specific knowledge of colours in curses code 641.3Sjdccdate Sun Apr 6 11:05:24 BST 2003 651.3Sjdcwho jdc 661.3Sjdcfile lib/libcurses/color.c : r1.24 671.3Sjdcdescr 681.3Sjdc Swap red/blue and yellow/cyan colours for COLOR_OTHER. 691.3Sjdc Fix is to enhance libtermcap to understand terminfo-style % sequences. 701.3Sjdc See also: 711.5Ssalo http://mail-index.NetBSD.org/tech-userlevel/2003/04/06/0000.html 721.4Sscwkcah 731.4Sscw 741.7Shehack Compensation for differing types of LINUX_USRSTACK and USRSTACK 751.7Shecdate 21 Aug 2003 761.7Shewho he 771.7Shefile sys/miscfs/procfs/procfs_linux.c : 1.14 781.7Shedescr 791.7She Not all ports have LINUX_USRSTACK and/or USRSTACK defined as 801.7She literals/constants, but refer to variables of a type which is 811.7She not "unsigned long", causing compilation of procfs_linux.c to 821.7She fail with "makes integer from pointer without a cast". This 831.7She is observed on e.g. the sun3 port. Ideally the "types" for 841.7She symbols should be consistent across all ports. 851.25Smrgkcah 861.25Smrg 871.25Smrghack GCC2 isn't C99 881.25Smrgcdate 5 December 2003 891.25Smrgwho mrg 901.25Smrgfile gnu/dist/binutils/ld/ldlex.c: r1.2 911.25Smrgdescr 921.25Smrg ldlex.c is generated from ldlex.l and we get macro redefinition 931.25Smrg errors from it if __STDC_VERSION__ is not C99 compliant. This 941.25Smrg hack forces NetBSD 1.6 and later systems to also include <stdint.h> 951.25Smrg to get the relevant definitions. This hack should be backed out 961.25Smrg when all platforms have switched away from GCC2. 971.11Slukemkcah 981.11Slukem 991.13Ssimonb 1001.21Smycroftport vax 1011.21Smycroft 1021.21Smycroft hack gcc 2.95/vax cannot handle __builtin_ffs() 1031.21Smycroft cdate Fri Aug 23 21:31:15 CEST 2002 1041.21Smycroft who ragge 1051.21Smycroft file sys/lib/libkern/libkern.h : 1.42 1061.21Smycroft descr 1071.21Smycroft __builtin_ffs() is ifdef'd away if __vax__. Matt Thomas has 1081.21Smycroft added this feature to GCC 3 so it can be removed when switching. 1091.21Smycroft kcah 1101.21Smycroft 1111.21Smycroft hack gcc-vax-libbz2 1121.21Smycroft mdate 27 Jun 2002 1131.21Smycroft who thorpej 1141.21Smycroft file lib/libbz2/Makefile 1151.21Smycroft descr 1161.21Smycroft libbz2 is mis-compiled with optimization with GCC 2.95.3 1171.24Sragge on VAX. -O0 works around this problem. 1181.24Sragge kcah 1191.24Sragge 1201.24Sragge hack gawk 1211.24Sragge mdate 8 Nov 2003 1221.24Sragge who ragge 1231.24Sragge file gnu/usr.bin/gawk/Makefile 1241.24Sragge descr 1251.24Sragge gawk is mis-compiled with optimization with GCC 2.95.3 1261.21Smycroft on VAX. -O0 works around this problem. 1271.21Smycroft kcah 1281.21Smycroft 1291.21Smycroft hack vax Toolchain bug 1301.21Smycroft cdate 28 Mar 2003 1311.21Smycroft who he 1321.21Smycroft file sys/arch/vax/vax/intvec.S : 1.5 1331.21Smycroft descr 1341.21Smycroft Workaround for PR toolchain/20924. The assembler apparently 1351.21Smycroft tries to range-check byte offsets when it doesn't have 1361.21Smycroft sufficient information to make that decision. 1371.21Smycroft Workaround uses `brw' instead of `brb' instruction. 1381.23She kcah 1391.23She 1401.23She hack vax gcc 2.95.3 -Wuninitialized workarounds 1411.23She cdate 6 Nov 2003 1421.23She who he 1431.23She file sys/dev/bi/if_ni.c : 1.22 1441.23She file sys/arch/vax/vax/ctu.c : 1.20 1451.23She descr 1461.23She Workaround for gcc 2.95.3 failing to detect that certain 1471.23She variables will be initialized; so that this code compiles 1481.23She with -Wuninitialized. Should be reviewed and possibly 1491.23She reverted when gcc 3.3.2 is ready for vax. 1501.21Smycroft kcah 1511.21Smycroft 1521.34Scl hack vax gcc 2.95.3 structure initialization 1531.34Scl cdat 13 May 2004 1541.34Scl who cl 1551.34Scl file sys/miscfs/kernfs/kernfs_vnops.c : 1.104 1561.34Scl descr 1571.34Scl Workaround for gcc 2.95.3 failing to initialize structures 1581.34Scl and/or unions inside structures using nested designators. 1591.34Scl Should be reverted when gcc >=3.3.3 is ready for vax. 1601.34Scl kcah 1611.21Smycroft 1621.21Smycroftport sparc64 1631.21Smycroft 1641.21Smycroft hack binutil-2.11-sparc64-pltrela 1651.21Smycroft mdate 14 Aug 2001 1661.21Smycroft who eeh 1671.21Smycroft file libexec/ld.elf_so/arch/sparc64/mdreloc.c 1681.21Smycroft descr 1691.21Smycroft The first for PLT entries are reserved. There is some 1701.21Smycroft disagreement whether they should have associated relocation 1711.21Smycroft entries. Both the SPARC 32-bit and 64-bit ELF specifications 1721.21Smycroft say that they should have relocation entries, but the 32-bit 1731.21Smycroft SPARC binutils do not generate them, and now the 64-bit SPARC 1741.21Smycroft binutils have stopped generating them too. 1751.21Smycroft 1761.21Smycroft To provide binary compatibility, we will check the first entry, 1771.21Smycroft if it is reserved it should not be of the type JMP_SLOT. If it 1781.21Smycroft is JMP_SLOT, then the 4 reserved entries were not generated and 1791.21Smycroft our index is 4 entries too far, so we frob the rela pointer. 1801.21Smycroft kcah 1811.21Smycroft 1821.21Smycroft 1831.21Smycroftport arm 1841.21Smycroft 1851.21Smycroft hack gcc-unsigned-compare 1861.21Smycroft cdate 09 Mar 2002 1871.21Smycroft mdate 18 Mar 2002 1881.21Smycroft who bjh21 1891.21Smycroft file dist/bind/lib/nameser/ns_parse.c : 1.3 1901.21Smycroft file dist/dhcp/minires/ns_parse.c : 1.3 1911.21Smycroft file dist/dhcp/omapip/result.c : 1.2 1921.21Smycroft file dist/dhcp/server/failover.c : 1.3 1931.21Smycroft file gnu/dist/gawk/eval.c : 1.4 1941.21Smycroft file gnu/dist/toolchain/bfd/bfd.c : 1.2 1951.21Smycroft file gnu/dist/toolchain/bfd/format.c : 1.2 1961.21Smycroft file gnu/dist/toolchain/gdb/target.c : 1.2 1971.21Smycroft file sys/kern/vfs_subr.c : 1.172 1981.21Smycroft descr When checking that a potentially-unsigned enum is >= 0, assign 1991.21Smycroft it to an int first. This is necessary to avoid "comparison is 2001.21Smycroft always true" warnings with -fshort-enums. Casting to an int 2011.21Smycroft really should be enough, but turns out not to be. 2021.21Smycroft kcah 2031.21Smycroft 2041.21Smycroft hack gcc 3.3.2/arm sys/kern/uipc_socket.c 2051.21Smycroft cdate Tue Oct 28 18:02:16 GMT 2003 2061.21Smycroft who skrll 2071.21Smycroft file sys/conf/arm/Makefile.arm : 1.19 2081.21Smycroft descr 2091.21Smycroft Performing a build.sh tools on a CATS with NFS mounted 2101.21Smycroft obj and src directories results in "panic: receive 1" 2111.21Smycroft The hack is to compile sys/kern/uipc_socket.c with 2121.21Smycroft -fno-strict-aliasing 2131.21Smycroft kcah 2141.21Smycroft 2151.21Smycroft 2161.13Ssimonbport pc532 2171.16Scl 2181.21Smycroft hack egcs-pc532-ip6_mroute 2191.21Smycroft cdate 09 Jul 2002 2201.21Smycroft who simonb 2211.21Smycroft file sys/arch/pc532/conf/Makefile.pc532 : 1.70 2221.21Smycroft file sys/arch/pc532/conf/files.pc532 : 1.47 2231.21Smycroft descr 2241.21Smycroft egcs 1.1.2 gets an "internal error--insn does not satisfy its 2251.21Smycroft constraints" error compiling ip6_mroute.c with -O2 or greater. 2261.21Smycroft -O1 works around this problem. 2271.21Smycroft kcah 2281.21Smycroft 2291.21Smycroft hack gcc-pc532-duffs_device 2301.21Smycroft cdate Mon Oct 27 07:23:05 UTC 2003 2311.21Smycroft who simonb 2321.21Smycroft file sys/arch/pc532/conf/Makefile.pc532 : 1.71 2331.21Smycroft file sys/arch/pc532/conf/files.pc532 : 1.52 2341.21Smycroft pr GCC PR optimization/5230 2351.21Smycroft descr 2361.21Smycroft gcc incorrectly gives an "unreachable code at beginning of 2371.21Smycroft switch statement" for a Duff's device construct in 2381.21Smycroft arch/pc532/dev/ncr.c. There is no way to disable just this 2391.21Smycroft warning, so -Wno-error is turned on for this file. 2401.21Smycroft kcah 2411.21Smycroft 2421.26Ssimonb hack gcc332-cppmacro-ice 2431.26Ssimonb cdate Sun Dec 7 12:48:33 UTC 2003 2441.26Ssimonb who simonb 2451.26Ssimonb file gnu/usr.bin/gcc3/libcpp/Makefile : 1.2 2461.26Ssimonb descr 2471.26Ssimonb gcc 3.3.2 gets an ICE when compiling cppmacro.c with -O2. 2481.26Ssimonb The hack is to use -O1 for this file. Fixed in gcc 2491.26Ssimonb -current mainline, unable to work out where the fix is 2501.26Ssimonb right now. 2511.26Ssimonb kcah 2521.26Ssimonb 2531.21Smycroft 2541.21Smycroftport sh3 2551.21Smycroft 2561.21Smycroft hack gcc-sh3-sed 2571.21Smycroft mdate 23 Apr 2002 2581.21Smycroft who thorpej 2591.21Smycroft file usr.bin/sed/Makefile : 1.9 : 9 : 13 2601.21Smycroft descr 2611.21Smycroft The in-tree GCC 2.95.3-based compiler ICEs when building 2621.26Ssimonb with optimization for SuperH. Hack is to build with -O0. 2631.21Smycroft kcah 2641.21Smycroft 2651.21Smycroft 2661.21Smycroftport sh5 2671.21Smycroft 2681.21Smycroft hack SuperH SH5 Toolchain Bugs 2691.21Smycroft cdate 11 Jul 2002 2701.21Smycroft who scw 2711.21Smycroft file usr.sbin/ndbootd/ndbootd.c : 1.6 2721.21Smycroft file usr.sbin/traceroute/traceroute.c : 1.48 2731.21Smycroft descr 2741.21Smycroft The SuperH SH5 toolchain (2.97-sh5-010522) gets an internal 2751.21Smycroft compiler error when assigning a bit-wise inverted value 2761.21Smycroft under some circumstances. 2771.21Smycroft Work around it by splitting the statement into two. 2781.21Smycroft kcah 2791.21Smycroft 2801.21Smycroft hack SuperH SH5 Toolchain Bugs 2811.21Smycroft cdate 17 May 2003 2821.21Smycroft who scw 2831.21Smycroft file lib/libpthread/pthread_lock.c : 1.7 2841.21Smycroft descr 2851.21Smycroft The SuperH SH5 toolchain generates incorrect PIC code when 2861.21Smycroft faced with a symbol which is declared extern, but has local 2871.21Smycroft scope due to being defined within an asm() statement (without 2881.21Smycroft being declared .globl in that statement). Work around it by 2891.21Smycroft adding the .globl. 2901.27Sscw kcah 2911.27Sscw 2921.27Sscw hack gcc 3.3.2 on sh5 - ICE compiling reload1.c at -O2 2931.27Sscw cdate Sun Dec 7 19:44:05 GMT 2003 2941.27Sscw who scw 2951.27Sscw file gnu/usr.bin/gcc3/backend/Makefile : 1.13 2961.27Sscw descr 2971.27Sscw Gcc3 for sh5 gets an ICE compiling reload1.c at -O2. 2981.27Sscw Drop to -O1 for now when building for sh5. 2991.27Sscw kcah 3001.28Sscw 3011.28Sscw hack gcc 3.3.2 on sh5 - ICE compiling elf32.c/elf64.c at -O2 3021.28Sscw cdate Sun Dec 7 21:10:46 GMT 2003 3031.28Sscw who scw 3041.28Sscw file gnu/lib/libbfd/Makefile : 1.39 3051.28Sscw descr 3061.28Sscw Gcc3 for sh5 gets an ICE compiling libbfd's elf32.c 3071.28Sscw and elf64.c at -O2. 3081.28Sscw Drop to -O1 for now when building for sh5. 3091.28Sscw kcah 3101.28Sscw 3111.27Sscw 3121.27Sscw hack gcc 3.3.2 on sh5 - fatal warning compiling function.c 3131.27Sscw cdate Sun Dec 7 19:44:05 GMT 2003 3141.27Sscw who scw 3151.27Sscw file gnu/usr.bin/gcc3/backend/Makefile : 1.13 3161.27Sscw descr 3171.27Sscw Gcc3 for sh5 issues a warning: 3181.27Sscw .../gnu/dist/gcc/gcc/function.c:4742: warning: `and' of 3191.27Sscw mutually exclusive equal-tests is always 0. 3201.27Sscw This causes the build to fail because of -Werror. 3211.27Sscw Work-around by defining NOGCCERROR when building for sh5. 3221.27Sscw kcah 3231.27Sscw 3241.27Sscw hack gcc 3.3.2 on sh5 - ICE compiling dist/bind/bin/named/ns_main.c 3251.27Sscw cdate Sun Dec 7 19:44:05 GMT 2003 3261.27Sscw who scw 3271.27Sscw file usr.sbin/bind/named/Makefile : 1.15 3281.27Sscw descr 3291.27Sscw Gcc3 for sh5 gets an ICE compiling ns_main.c at -O2. 3301.27Sscw Drop to -O1 for now when building for sh5. 3311.21Smycroft kcah 3321.17Smrg 3331.17Smrg 3341.17Smrgport sun2 3351.20Sskrll 3361.21Smycroft hack gcc 3.3.2/mdsetimage 3371.21Smycroft cdate Tue Oct 28 18:43:05 EST 2003 3381.21Smycroft who fredette 3391.21Smycroft file gnu/usr.sbin/mdsetimage/Makefile : 1.15 3401.21Smycroft descr 3411.21Smycroft mdsetimage.c causes cc1 to barf for as-yet unknown reasons. 3421.26Ssimonb this hack uses -O0 to avoid the problem. 3431.21Smycroft kcah 3441.21Smycroft 3451.21Smycroft hack gcc 3.3.2/installboot/ffs.c 3461.21Smycroft cdate Tue Oct 28 18:43:05 EST 2003 3471.21Smycroft who fredette 3481.21Smycroft file usr.sbin/installboot/Makefile : 1.26 3491.21Smycroft descr 3501.21Smycroft ffs.c causes cc1 to barf for as-yet unknown reasons. this 3511.26Ssimonb hack uses -O0 to avoid the problem. 3521.22Smrg kcah 3531.22Smrg 3541.22Smrg 3551.22Smrgport sun2, m68k 3561.22Smrg 3571.22Smrg hack gcc 3.3.2/gcc/ra-colorize.c 3581.22Smrg cdate Wed Nov 5 11:03:13 EST 2003 3591.22Smrg who mrg 3601.22Smrg file gnu/usr.bin/gcc/backend/Makefile : 1.11 3611.22Smrg descr 3621.22Smrg ra-colorize.c causes a whole spew of unreferenced .L* local 3631.22Smrg symbols. using -O0 (or even -O2 reportedly) avoids the problem. 3641.21Smycroft kcah 3651.31Smartin 3661.31Smartinport sparc64 3671.31Smartin 3681.31Smartin hack disable optimzations for uvm_bio.c on 32 bit kernels 3691.31Smartin cdate Sun Mar 21 14:14:04 MET 2004 3701.31Smartin who martin 3711.32Smartin file sys/arch/sparc64/conf/Makefile.sparc64 : 1.53-1.54 3721.32Smartin file sys/arch/sparc64/conf/files.sparc64 : 1.93 3731.31Smartin descr 3741.31Smartin The sparc compiler miscompiles uvm_bio.c when using 3751.31Smartin any optimization. This results in ubc_release panics. 3761.31Smartin kcah 377