From bugs at busybox.net Thu Feb 1 03:11:49 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Thu, 1 Feb 2007 03:11:49 -0800 Subject: [buildroot 0001190]: PATCH: nfs-tools package - fakeroot install Message-ID: The following issue has been SUBMITTED. ====================================================================== http://busybox.net/bugs/view.php?id=1190 ====================================================================== Reported By: janlana Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1190 Category: Other Reproducibility: always Severity: feature Priority: normal Status: assigned ====================================================================== Date Submitted: 02-01-2007 03:11 PST Last Modified: 02-01-2007 03:11 PST ====================================================================== Summary: PATCH: nfs-tools package - fakeroot install Description: nfs-tools.mk used fakeroot for install because the nfs-tools Makefile wants chmod one directory to nobody. To use fakeroot is a bit dirty but mainly it causes spurious rebuild each time. The attached patch fix it - it set $statduser=`whoami` so the 'make install' does not fail. Because buildroot does not need the dirs (nfs-tools.mk removes them), it does not make any run problems. ====================================================================== Issue History Date Modified Username Field Change ====================================================================== 02-01-07 03:11 janlana New Issue 02-01-07 03:11 janlana Status new => assigned 02-01-07 03:11 janlana Assigned To => uClibc 02-01-07 03:11 janlana File Added: package_nfs.patch ====================================================================== From bugs at busybox.net Thu Feb 1 03:15:49 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Thu, 1 Feb 2007 03:15:49 -0800 Subject: [buildroot 0001191]: fix gcc, grub and lzma pckages spurious rebuild Message-ID: The following issue has been SUBMITTED. ====================================================================== http://busybox.net/bugs/view.php?id=1191 ====================================================================== Reported By: janlana Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1191 Category: Architecture Specific Reproducibility: always Severity: minor Priority: normal Status: assigned ====================================================================== Date Submitted: 02-01-2007 03:15 PST Last Modified: 02-01-2007 03:15 PST ====================================================================== Summary: fix gcc, grub and lzma pckages spurious rebuild Description: see attached patches ====================================================================== Issue History Date Modified Username Field Change ====================================================================== 02-01-07 03:15 janlana New Issue 02-01-07 03:15 janlana Status new => assigned 02-01-07 03:15 janlana Assigned To => uClibc 02-01-07 03:15 janlana File Added: package_lzma.patch ====================================================================== From bugs at busybox.net Thu Feb 1 13:53:06 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Thu, 1 Feb 2007 13:53:06 -0800 Subject: [buildroot 0000967]: SEGV when running openssh Message-ID: A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=967 ====================================================================== Reported By: lerichi Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 967 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: assigned ====================================================================== Date Submitted: 07-23-2006 14:59 PDT Last Modified: 02-01-2007 13:53 PST ====================================================================== Summary: SEGV when running openssh Description: I downloaded and did a standard build of the latest buildroot. Busybox works just fine, but openssh and all its helpers just SEGV. Apart from that, the other non-busybox proggies also die. So there must be a problem with some library. Just make a standard build (gcc-3.4.2), chroot into build_ARCH/root and call sshd. It should die :-( Any ideas? I have a suspicion that xstrdup() might be the cause. Regards- Richard ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-24-07 13:04 ---------------------------------------------------------------------- Try to strace this and perpahs provide a simple testcase for us to reproduce. ---------------------------------------------------------------------- gheeres - 02-01-07 13:53 ---------------------------------------------------------------------- Attached strace output. Build / configuration: uClibc-snapshost (01/26/2007) gcc-v4.1.1 openssh-v3.9p1 Same results attempting to use openssh-v4.5p1 If strace needs to be in different format, please provide instructions for generating. Issue History Date Modified Username Field Change ====================================================================== 07-23-06 14:59 lerichi New Issue 07-23-06 14:59 lerichi Status new => assigned 07-23-06 14:59 lerichi Assigned To => uClibc 07-27-06 04:58 thrakhath Issue Monitored: thrakhath 01-24-07 13:04 bernhardf Note Added: 0002041 02-01-07 13:50 gheeres File Added: uClibc-snapshot-openssh-v3.9p1.strace 02-01-07 13:53 gheeres Note Added: 0002081 ====================================================================== From bugs at busybox.net Thu Feb 1 14:10:59 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Thu, 1 Feb 2007 14:10:59 -0800 Subject: [buildroot 0000967]: SEGV when running openssh Message-ID: A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=967 ====================================================================== Reported By: lerichi Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 967 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: assigned ====================================================================== Date Submitted: 07-23-2006 14:59 PDT Last Modified: 02-01-2007 14:10 PST ====================================================================== Summary: SEGV when running openssh Description: I downloaded and did a standard build of the latest buildroot. Busybox works just fine, but openssh and all its helpers just SEGV. Apart from that, the other non-busybox proggies also die. So there must be a problem with some library. Just make a standard build (gcc-3.4.2), chroot into build_ARCH/root and call sshd. It should die :-( Any ideas? I have a suspicion that xstrdup() might be the cause. Regards- Richard ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-24-07 13:04 ---------------------------------------------------------------------- Try to strace this and perpahs provide a simple testcase for us to reproduce. ---------------------------------------------------------------------- gheeres - 02-01-07 13:53 ---------------------------------------------------------------------- Attached strace output. Build / configuration: uClibc-snapshost (01/26/2007) gcc-v4.1.1 openssh-v3.9p1 Same results attempting to use openssh-v4.5p1 If strace needs to be in different format, please provide instructions for generating. ---------------------------------------------------------------------- gheeres - 02-01-07 14:10 ---------------------------------------------------------------------- Older buildroot using: uClibc-0.9.28 (non-snapshot) gcc-v3.4.6 openssh-v3.9p1 or openssh-v4.5p1 Works correctly. Issue History Date Modified Username Field Change ====================================================================== 07-23-06 14:59 lerichi New Issue 07-23-06 14:59 lerichi Status new => assigned 07-23-06 14:59 lerichi Assigned To => uClibc 07-27-06 04:58 thrakhath Issue Monitored: thrakhath 01-24-07 13:04 bernhardf Note Added: 0002041 02-01-07 13:50 gheeres File Added: uClibc-snapshot-openssh-v3.9p1.strace 02-01-07 13:53 gheeres Note Added: 0002081 02-01-07 14:10 gheeres Note Added: 0002082 ====================================================================== From bugs at busybox.net Thu Feb 1 15:18:48 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Thu, 1 Feb 2007 15:18:48 -0800 Subject: [buildroot 0001156]: coldfire (no mmu) support Message-ID: A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=1156 ====================================================================== Reported By: kendallc Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1156 Category: Architecture Specific Reproducibility: N/A Severity: feature Priority: normal Status: assigned ====================================================================== Date Submitted: 01-11-2007 16:11 PST Last Modified: 02-01-2007 15:18 PST ====================================================================== Summary: coldfire (no mmu) support Description: Here is a patch to add the ability for Buildroot to target the Coldfire processor, which lacks an MMU and FPU. ====================================================================== ---------------------------------------------------------------------- kendallc - 01-11-07 16:18 ---------------------------------------------------------------------- '-m52000' should be '-m5200', sorry. Also, you might be able to use '-m5206e' or '-m5307' to better support you coldfire version. I use '-m5206e' for the MCF5272. ---------------------------------------------------------------------- kendallc - 01-22-07 10:23 ---------------------------------------------------------------------- Buildroot changes quickly, and I frequently need to update this patch to get it to apply. If anyone pursues integrating any of the changes in this patch, post a note here and I'll post an updated patch. ---------------------------------------------------------------------- vapier - 01-27-07 21:28 ---------------------------------------------------------------------- most of your patches are directed to the wrong places fix individual projects, dont dump into buildroot for example: split your uClibc fixes out and e-mail the uClibc list split your elf2flt fixes out and e-mail the uClinux list split out your mtd fixes and e-mail the mtd list etc... ---------------------------------------------------------------------- kendallc - 01-28-07 13:16 ---------------------------------------------------------------------- i'll split out the buildroot makefile patches from the upstream project patches, and i'll do what I can to get the upstream patches to their repective projects; however, most of them might not get back upstream for the following reasons. elf2flt: no patch, only patched the buildroot makefile mtd-utils: does not appear to be maintained for linux 2.4 portmap: maintainership is unclear, but we'll see boa: it's a patch i took from uclinux, they should probably get it to boa busybox: i'll try uclibc: patched against release 0.9.28 because i needed to start with a stable base to get everything else working the buildroot patches are not really tied to the upstream patches except in the case of uclibc. i noticed and took advantage of a rule in the buildroot makefile that handles patches for released uclibc versions. i'll look into the latest uclibc code, but any changes would obviously not go into release 0.9.28. so, i'll break out the buildroot patch. let me know if you have any suggestions on what to do about patching released uclibc versions. ---------------------------------------------------------------------- kendallc - 02-01-07 15:18 ---------------------------------------------------------------------- i just attached a much leaner patch for the buildroot makefiles only. some of the changes are to propagate LDFLAGS so that you can link BFLT binaries with elf2flt. there is a fix to the elf2flt makefile to for cleaning and rebuilding. there are changes to zlib to build it statically. these changes will be needed for most mmu-less builds. the rest of the changes are for the coldfire/nommu support itself. they add a menu option for the coldfire, some associated vars like BR2_ARCH_HAS_NO_MMU, and some changes to the gcc and uclibc makefiles. they also pass CFLAGS for mmu-less builds to packages like busybox if necessary. Issue History Date Modified Username Field Change ====================================================================== 01-11-07 16:11 kendallc New Issue 01-11-07 16:11 kendallc Status new => assigned 01-11-07 16:11 kendallc Assigned To => uClibc 01-11-07 16:11 kendallc File Added: buildroot.patch 01-11-07 16:12 kendallc File Added: buildroot_coldfire.config 01-11-07 16:12 kendallc File Added: buildroot_powerpc.config 01-11-07 16:13 kendallc File Added: busybox.config 01-11-07 16:18 kendallc Note Added: 0001984 01-22-07 10:23 kendallc Note Added: 0002029 01-27-07 21:28 vapier Note Added: 0002063 01-28-07 13:16 kendallc Note Added: 0002064 02-01-07 14:52 kendallc File Added: buildroot-only.patch 02-01-07 15:18 kendallc Note Added: 0002084 ====================================================================== From andersen at uclibc.org Thu Feb 1 16:23:55 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 16:23:55 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet Message-ID: <20070202002355.8DC78485B2@busybox.net> Author: andersen Date: 2007-02-01 16:23:55 -0800 (Thu, 01 Feb 2007) New Revision: 17701 Log: fix obvious bug in ipv4/ipv6 resolving. When not using AF_INET, gethostbyname2_r tries to resolve an ipv6 address from /etc/hosts using get_hosts_byname_r, but with AF_INET instead of the supplied address family. This returns ipv4 addresses marked as ipv6 ones. Fix from nbd. Modified: trunk/uClibc/libc/inet/resolv.c Changeset: Modified: trunk/uClibc/libc/inet/resolv.c =================================================================== --- trunk/uClibc/libc/inet/resolv.c 2007-02-01 23:11:04 UTC (rev 17700) +++ trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:23:55 UTC (rev 17701) @@ -2130,7 +2130,7 @@ int old_errno = errno; /* Save the old errno and reset errno */ __set_errno(0); /* to check for missing /etc/hosts. */ - if ((i=__get_hosts_byname_r(name, AF_INET, result_buf, + if ((i=__get_hosts_byname_r(name, family, result_buf, buf, buflen, result, h_errnop))==0) return i; switch (*h_errnop) { From andersen at uclibc.org Thu Feb 1 16:25:22 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 16:25:22 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/inet Message-ID: <20070202002522.EBD2E485EE@busybox.net> Author: andersen Date: 2007-02-01 16:25:22 -0800 (Thu, 01 Feb 2007) New Revision: 17702 Log: fix obvious bug in ipv4/ipv6 resolving. When not using AF_INET, gethostbyname2_r tries to resolve an ipv6 address from /etc/hosts using get_hosts_byname_r, but with AF_INET instead of the supplied address family. This returns ipv4 addresses marked as ipv6 ones. Fix from nbd. Modified: branches/uClibc_0_9_28/libc/inet/resolv.c Changeset: Modified: branches/uClibc_0_9_28/libc/inet/resolv.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 00:23:55 UTC (rev 17701) +++ branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 00:25:22 UTC (rev 17702) @@ -2064,7 +2064,7 @@ int old_errno = errno; /* Save the old errno and reset errno */ __set_errno(0); /* to check for missing /etc/hosts. */ - if ((i=__get_hosts_byname_r(name, AF_INET, result_buf, + if ((i=__get_hosts_byname_r(name, family, result_buf, buf, buflen, result, h_errnop))==0) return i; switch (*h_errnop) { From andersen at uclibc.org Thu Feb 1 16:27:53 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 16:27:53 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/inet Message-ID: <20070202002753.D6E5D4861D@busybox.net> Author: andersen Date: 2007-02-01 16:27:53 -0800 (Thu, 01 Feb 2007) New Revision: 17703 Log: Ronald Maeder writes: Hi Erik, Thanks for all your great work. I found a set of bugs in resolv.c . Basically, there is code that looks like: BIGLOCK; __nameserversXX=__nameservers; __nameserverXX=__nameserver; BIGUNLOCK; i = __dns_lookup(dname, type, __nameserversXX, __nameserverXX, &packet, &a); which is a problem because the declarations are int __nameservers; char * __nameserver[MAX_SERVERS]; int __searchdomains; char * __searchdomain[MAX_SEARCH]; so you can see that __nameserver is a pointer. Copying the pointer to __nameserverXX doesn't protect the global variable space. I have attached a patch and the new file. I hope you will incorporate these bug fixes. I spent quite a bit of time tracking them down. Many thanks, Ron Modified: branches/uClibc_0_9_28/libc/inet/resolv.c Changeset: Modified: branches/uClibc_0_9_28/libc/inet/resolv.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 00:25:22 UTC (rev 17702) +++ branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 00:27:53 UTC (rev 17703) @@ -695,7 +695,11 @@ ++local_id; local_id &= 0xffff; h.id = local_id; + BIGLOCK; + /* this is really __nameserver[] which is a global that + needs a lock!! */ dns = nsip[local_ns]; + BIGUNLOCK; h.qdcount = 1; h.rd = 1; @@ -730,7 +734,11 @@ retries+1, NAMESERVER_PORT, dns); #ifdef __UCLIBC_HAS_IPV6__ + BIGLOCK; + /* 'dns' is really __nameserver[] which is a global that + needs a lock!! */ v6 = inet_pton(AF_INET6, dns, &sa6.sin6_addr) > 0; + BIGUNLOCK; fd = socket(v6 ? AF_INET6 : AF_INET, SOCK_DGRAM, IPPROTO_UDP); #else fd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); @@ -751,7 +759,11 @@ #endif sa.sin_family = AF_INET; sa.sin_port = htons(NAMESERVER_PORT); + BIGLOCK; + /* 'dns' is really __nameserver[] which is a global that + needs a lock!! */ sa.sin_addr.s_addr = inet_addr(dns); + BIGUNLOCK; rc = connect(fd, (struct sockaddr *) &sa, sizeof(sa)); #ifdef __UCLIBC_HAS_IPV6__ } From andersen at uclibc.org Thu Feb 1 16:30:30 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 16:30:30 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet Message-ID: <20070202003030.B4EB048616@busybox.net> Author: andersen Date: 2007-02-01 16:30:30 -0800 (Thu, 01 Feb 2007) New Revision: 17704 Log: Ronald Maeder writes: Hi Erik, Thanks for all your great work. I found a set of bugs in resolv.c . Basically, there is code that looks like: BIGLOCK; __nameserversXX=__nameservers; __nameserverXX=__nameserver; BIGUNLOCK; i = __dns_lookup(dname, type, __nameserversXX, __nameserverXX, &packet, &a); which is a problem because the declarations are int __nameservers; char * __nameserver[MAX_SERVERS]; int __searchdomains; char * __searchdomain[MAX_SEARCH]; so you can see that __nameserver is a pointer. Copying the pointer to __nameserverXX doesn't protect the global variable space. I have attached a patch and the new file. I hope you will incorporate these bug fixes. I spent quite a bit of time tracking them down. Many thanks, Ron Modified: trunk/uClibc/libc/inet/resolv.c Changeset: Modified: trunk/uClibc/libc/inet/resolv.c =================================================================== --- trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:27:53 UTC (rev 17703) +++ trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:30:30 UTC (rev 17704) @@ -753,7 +753,11 @@ ++local_id; local_id &= 0xffff; h.id = local_id; + BIGLOCK; + /* this is really __nameserver[] which is a global that + needs a lock!! */ dns = nsip[local_ns]; + BIGUNLOCK; h.qdcount = 1; h.rd = 1; @@ -788,7 +792,11 @@ retries+1, NAMESERVER_PORT, dns); #ifdef __UCLIBC_HAS_IPV6__ + BIGLOCK; + /* 'dns' is really __nameserver[] which is a global that + needs a lock!! */ v6 = inet_pton(AF_INET6, dns, &sa6.sin6_addr) > 0; + BIGUNLOCK; fd = socket(v6 ? AF_INET6 : AF_INET, SOCK_DGRAM, IPPROTO_UDP); #else fd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); @@ -809,7 +817,11 @@ #endif sa.sin_family = AF_INET; sa.sin_port = htons(NAMESERVER_PORT); + BIGLOCK; + /* 'dns' is really __nameserver[] which is a global that + needs a lock!! */ sa.sin_addr.s_addr = inet_addr(dns); + BIGUNLOCK; rc = connect(fd, (struct sockaddr *) &sa, sizeof(sa)); #ifdef __UCLIBC_HAS_IPV6__ } From andersen at uclibc.org Thu Feb 1 16:37:07 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 16:37:07 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet Message-ID: <20070202003707.0313B485F7@busybox.net> Author: andersen Date: 2007-02-01 16:37:06 -0800 (Thu, 01 Feb 2007) New Revision: 17705 Log: fix lock naming Modified: trunk/uClibc/libc/inet/resolv.c Changeset: Modified: trunk/uClibc/libc/inet/resolv.c =================================================================== --- trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:30:30 UTC (rev 17704) +++ trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:37:06 UTC (rev 17705) @@ -753,11 +753,11 @@ ++local_id; local_id &= 0xffff; h.id = local_id; - BIGLOCK; + __UCLIBC_MUTEX_LOCK(mylock); /* this is really __nameserver[] which is a global that needs a lock!! */ dns = nsip[local_ns]; - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(mylock); h.qdcount = 1; h.rd = 1; @@ -792,11 +792,11 @@ retries+1, NAMESERVER_PORT, dns); #ifdef __UCLIBC_HAS_IPV6__ - BIGLOCK; + __UCLIBC_MUTEX_LOCK(mylock); /* 'dns' is really __nameserver[] which is a global that needs a lock!! */ v6 = inet_pton(AF_INET6, dns, &sa6.sin6_addr) > 0; - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(mylock); fd = socket(v6 ? AF_INET6 : AF_INET, SOCK_DGRAM, IPPROTO_UDP); #else fd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); @@ -817,11 +817,11 @@ #endif sa.sin_family = AF_INET; sa.sin_port = htons(NAMESERVER_PORT); - BIGLOCK; + __UCLIBC_MUTEX_LOCK(mylock); /* 'dns' is really __nameserver[] which is a global that needs a lock!! */ sa.sin_addr.s_addr = inet_addr(dns); - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(mylock); rc = connect(fd, (struct sockaddr *) &sa, sizeof(sa)); #ifdef __UCLIBC_HAS_IPV6__ } From andersen at uclibc.org Thu Feb 1 17:01:52 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 17:01:52 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet Message-ID: <20070202010152.340F2485DC@busybox.net> Author: andersen Date: 2007-02-01 17:01:51 -0800 (Thu, 01 Feb 2007) New Revision: 17706 Log: use the correct lock for access to __nameserver[] Modified: trunk/uClibc/libc/inet/resolv.c Changeset: Modified: trunk/uClibc/libc/inet/resolv.c =================================================================== --- trunk/uClibc/libc/inet/resolv.c 2007-02-02 00:37:06 UTC (rev 17705) +++ trunk/uClibc/libc/inet/resolv.c 2007-02-02 01:01:51 UTC (rev 17706) @@ -753,11 +753,11 @@ ++local_id; local_id &= 0xffff; h.id = local_id; - __UCLIBC_MUTEX_LOCK(mylock); + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* this is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ dns = nsip[local_ns]; - __UCLIBC_MUTEX_UNLOCK(mylock); + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); h.qdcount = 1; h.rd = 1; @@ -792,11 +792,11 @@ retries+1, NAMESERVER_PORT, dns); #ifdef __UCLIBC_HAS_IPV6__ - __UCLIBC_MUTEX_LOCK(mylock); + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* 'dns' is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ v6 = inet_pton(AF_INET6, dns, &sa6.sin6_addr) > 0; - __UCLIBC_MUTEX_UNLOCK(mylock); + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); fd = socket(v6 ? AF_INET6 : AF_INET, SOCK_DGRAM, IPPROTO_UDP); #else fd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); @@ -817,11 +817,11 @@ #endif sa.sin_family = AF_INET; sa.sin_port = htons(NAMESERVER_PORT); - __UCLIBC_MUTEX_LOCK(mylock); + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* 'dns' is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ sa.sin_addr.s_addr = inet_addr(dns); - __UCLIBC_MUTEX_UNLOCK(mylock); + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); rc = connect(fd, (struct sockaddr *) &sa, sizeof(sa)); #ifdef __UCLIBC_HAS_IPV6__ } From andersen at uclibc.org Thu Feb 1 17:03:29 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 17:03:29 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/inet Message-ID: <20070202010329.C3A41485DC@busybox.net> Author: andersen Date: 2007-02-01 17:03:29 -0800 (Thu, 01 Feb 2007) New Revision: 17707 Log: use the correct lock to guard access to the __nameserver[] global Modified: branches/uClibc_0_9_28/libc/inet/resolv.c Changeset: Modified: branches/uClibc_0_9_28/libc/inet/resolv.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 01:01:51 UTC (rev 17706) +++ branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 01:03:29 UTC (rev 17707) @@ -695,11 +695,11 @@ ++local_id; local_id &= 0xffff; h.id = local_id; - BIGLOCK; + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* this is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ dns = nsip[local_ns]; - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); h.qdcount = 1; h.rd = 1; @@ -734,11 +734,11 @@ retries+1, NAMESERVER_PORT, dns); #ifdef __UCLIBC_HAS_IPV6__ - BIGLOCK; + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* 'dns' is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ v6 = inet_pton(AF_INET6, dns, &sa6.sin6_addr) > 0; - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); fd = socket(v6 ? AF_INET6 : AF_INET, SOCK_DGRAM, IPPROTO_UDP); #else fd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); @@ -759,11 +759,11 @@ #endif sa.sin_family = AF_INET; sa.sin_port = htons(NAMESERVER_PORT); - BIGLOCK; + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* 'dns' is really __nameserver[] which is a global that - needs a lock!! */ + needs to hold __resolv_lock before access!! */ sa.sin_addr.s_addr = inet_addr(dns); - BIGUNLOCK; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); rc = connect(fd, (struct sockaddr *) &sa, sizeof(sa)); #ifdef __UCLIBC_HAS_IPV6__ } From andersen at uclibc.org Thu Feb 1 17:28:37 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 17:28:37 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/inet Message-ID: <20070202012837.4136048584@busybox.net> Author: andersen Date: 2007-02-01 17:28:36 -0800 (Thu, 01 Feb 2007) New Revision: 17713 Log: Ronald Maeder writes: I have successfully made gethostbyname_r(), res_init(), and gethostid() fully reentrant. In addition, I have added a NULL check to inet_aton(). This is where SEG FAULTs were coming from when gethostbyname_r() was called. Modified: branches/uClibc_0_9_28/libc/inet/addr.c branches/uClibc_0_9_28/libc/inet/hostid.c branches/uClibc_0_9_28/libc/inet/resolv.c Changeset: Modified: branches/uClibc_0_9_28/libc/inet/addr.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/addr.c 2007-02-02 01:19:09 UTC (rev 17712) +++ branches/uClibc_0_9_28/libc/inet/addr.c 2007-02-02 01:28:36 UTC (rev 17713) @@ -52,6 +52,10 @@ int value; int part; + if (cp == NULL) { /* check for NULL arg */ + return 0; + } + addr = 0; for (part = 1; part <= 4; part++) { Modified: branches/uClibc_0_9_28/libc/inet/hostid.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/hostid.c 2007-02-02 01:19:09 UTC (rev 17712) +++ branches/uClibc_0_9_28/libc/inet/hostid.c 2007-02-02 01:28:36 UTC (rev 17713) @@ -56,7 +56,19 @@ struct hostent *hp; struct in_addr in; - if ((hp = gethostbyname(host)) == (struct hostent *)NULL) + /* replace gethostbyname() with gethostbyname_r() - ron at zing.net */ + /*if ((hp = gethostbyname(host)) == (struct hostent *)NULL)*/ + { + struct hostent ghbn_h; + char ghbn_buf[sizeof(struct in_addr) + + sizeof(struct in_addr *)*2 + + sizeof(char *)*((2 + 5/*MAX_ALIASES*/ + + 1)/*ALIAS_DIM*/) + + 256/*namebuffer*/ + 32/* margin */]; + int ghbn_errno; + gethostbyname_r(host, &ghbn_h, ghbn_buf, sizeof(ghbn_buf), &hp, &ghbn_errno); + } + if (hp == (struct hostent *)NULL) /* This is not a error if we get here, as all it means is that * this host is not on a network and/or they have not Modified: branches/uClibc_0_9_28/libc/inet/resolv.c =================================================================== --- branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 01:19:09 UTC (rev 17712) +++ branches/uClibc_0_9_28/libc/inet/resolv.c 2007-02-02 01:28:36 UTC (rev 17713) @@ -180,6 +180,9 @@ /* Global stuff (stuff needing to be locked to be thread safe)... */ + +/* We use __resolv_lock to guard access to the + * '__nameservers' and __searchdomains globals */ extern int __nameservers; extern char * __nameserver[MAX_SERVERS]; extern int __searchdomains; @@ -965,7 +968,7 @@ int __searchdomains; char * __searchdomain[MAX_SEARCH]; -__UCLIBC_MUTEX_INIT(__resolv_lock, PTHREAD_MUTEX_INITIALIZER); +__UCLIBC_MUTEX_INIT(__resolv_lock, PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP); /* * we currently read formats not quite the same as that on normal @@ -1096,12 +1099,14 @@ #ifdef L_res_init +/* We use __resolv_lock to guard access to global '_res' */ struct __res_state _res; int res_init(void) { struct __res_state *rp = &(_res); + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* must be a recursive lock! */ __close_nameservers(); __open_nameservers(); rp->retrans = RES_TIMEOUT; @@ -1119,7 +1124,6 @@ /** rp->rhook = NULL; **/ /** rp->_u._ext.nsinit = 0; **/ - __UCLIBC_MUTEX_LOCK(__resolv_lock); if(__searchdomains) { int i; for(i=0; i<__searchdomains; i++) { @@ -1217,8 +1221,14 @@ u_int dots; int trailing_dot, ret, saved_herrno; int got_nodata = 0, got_servfail = 0, tried_as_is = 0; + u_long _res_options; + unsigned _res_ndots; + char **_res_dnsrch; - if ((!name || !answer) || ((_res.options & RES_INIT) == 0 && res_init() == -1)) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!name || !answer) || ((_res_options & RES_INIT) == 0 && res_init() == -1)) { h_errno = NETDB_INTERNAL; return (-1); } @@ -1237,7 +1247,10 @@ * 'as is'. The threshold can be set with the "ndots" option. */ saved_herrno = -1; - if (dots >= _res.ndots) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_ndots = _res.ndots; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (dots >= _res_ndots) { ret = res_querydomain(name, NULL, class, type, answer, anslen); if (ret > 0) return (ret); @@ -1251,11 +1264,15 @@ * - there is at least one dot, there is no trailing dot, * and RES_DNSRCH is set. */ - if ((!dots && (_res.options & RES_DEFNAMES)) || - (dots && !trailing_dot && (_res.options & RES_DNSRCH))) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + _res_dnsrch = _res.dnsrch; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!dots && (_res_options & RES_DEFNAMES)) || + (dots && !trailing_dot && (_res_options & RES_DNSRCH))) { int done = 0; - for (domain = (const char * const *)_res.dnsrch; + for (domain = (const char * const *)_res_dnsrch; *domain && !done; domain++) { @@ -1304,7 +1321,10 @@ * if we got here for some reason other than DNSRCH, * we only wanted one iteration of the loop, so stop. */ - if (!(_res.options & RES_DNSRCH)) + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (!(_res_options & RES_DNSRCH)) done++; } } @@ -1350,14 +1370,21 @@ char nbuf[MAXDNAME]; const char *longname = nbuf; size_t n, d; + u_long _res_options; - if ((!name || !answer) || ((_res.options & RES_INIT) == 0 && res_init() == -1)) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!name || !answer) || ((_res_options & RES_INIT) == 0 && res_init() == -1)) { h_errno = NETDB_INTERNAL; return (-1); } #ifdef DEBUG - if (_res.options & RES_DEBUG) + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (_res_options & RES_DEBUG) printf(";; res_querydomain(%s, %s, %d, %d)\n", name, domain?domain:"", class, type); #endif From andersen at uclibc.org Thu Feb 1 17:29:10 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 17:29:10 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet Message-ID: <20070202012910.410AF48584@busybox.net> Author: andersen Date: 2007-02-01 17:29:10 -0800 (Thu, 01 Feb 2007) New Revision: 17714 Log: Ronald Maeder writes: I have successfully made gethostbyname_r(), res_init(), and gethostid() fully reentrant. In addition, I have added a NULL check to inet_aton(). This is where SEG FAULTs were coming from when gethostbyname_r() was called. Modified: trunk/uClibc/libc/inet/addr.c trunk/uClibc/libc/inet/hostid.c trunk/uClibc/libc/inet/resolv.c Changeset: Modified: trunk/uClibc/libc/inet/addr.c =================================================================== --- trunk/uClibc/libc/inet/addr.c 2007-02-02 01:28:36 UTC (rev 17713) +++ trunk/uClibc/libc/inet/addr.c 2007-02-02 01:29:10 UTC (rev 17714) @@ -55,6 +55,10 @@ int value; int part; + if (cp == NULL) { /* check for NULL arg */ + return 0; + } + addr = 0; for (part = 1; part <= 4; part++) { Modified: trunk/uClibc/libc/inet/hostid.c =================================================================== --- trunk/uClibc/libc/inet/hostid.c 2007-02-02 01:28:36 UTC (rev 17713) +++ trunk/uClibc/libc/inet/hostid.c 2007-02-02 01:29:10 UTC (rev 17714) @@ -73,7 +73,19 @@ struct hostent *hp; struct in_addr in; - if ((hp = gethostbyname(host)) == (struct hostent *)NULL) + /* replace gethostbyname() with gethostbyname_r() - ron at zing.net */ + /*if ((hp = gethostbyname(host)) == (struct hostent *)NULL)*/ + { + struct hostent ghbn_h; + char ghbn_buf[sizeof(struct in_addr) + + sizeof(struct in_addr *)*2 + + sizeof(char *)*((2 + 5/*MAX_ALIASES*/ + + 1)/*ALIAS_DIM*/) + + 256/*namebuffer*/ + 32/* margin */]; + int ghbn_errno; + gethostbyname_r(host, &ghbn_h, ghbn_buf, sizeof(ghbn_buf), &hp, &ghbn_errno); + } + if (hp == (struct hostent *)NULL) /* This is not a error if we get here, as all it means is that * this host is not on a network and/or they have not Modified: trunk/uClibc/libc/inet/resolv.c =================================================================== --- trunk/uClibc/libc/inet/resolv.c 2007-02-02 01:28:36 UTC (rev 17713) +++ trunk/uClibc/libc/inet/resolv.c 2007-02-02 01:29:10 UTC (rev 17714) @@ -1018,12 +1018,14 @@ #ifdef L_opennameservers +/* We use __resolv_lock to guard access to the + * '__nameservers' and __searchdomains globals */ int __nameservers; char * __nameserver[MAX_SERVERS]; int __searchdomains; char * __searchdomain[MAX_SEARCH]; -__UCLIBC_MUTEX_INIT(__resolv_lock, PTHREAD_MUTEX_INITIALIZER); +__UCLIBC_MUTEX_INIT(__resolv_lock, PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP); /* * we currently read formats not quite the same as that on normal @@ -1155,12 +1157,14 @@ #ifdef L_res_init +/* We use __resolv_lock to guard access to global '_res' */ struct __res_state _res; int res_init(void) { struct __res_state *rp = &(_res); + __UCLIBC_MUTEX_LOCK(__resolv_lock); /* must be a recursive lock! */ __close_nameservers(); __open_nameservers(); rp->retrans = RES_TIMEOUT; @@ -1178,7 +1182,6 @@ /** rp->rhook = NULL; **/ /** rp->_u._ext.nsinit = 0; **/ - __UCLIBC_MUTEX_LOCK(__resolv_lock); if(__searchdomains) { int i; for(i=0; i<__searchdomains; i++) { @@ -1278,8 +1281,14 @@ u_int dots; int trailing_dot, ret, saved_herrno; int got_nodata = 0, got_servfail = 0, tried_as_is = 0; + u_long _res_options; + unsigned _res_ndots; + char **_res_dnsrch; - if ((!name || !answer) || ((_res.options & RES_INIT) == 0 && res_init() == -1)) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!name || !answer) || ((_res_options & RES_INIT) == 0 && res_init() == -1)) { h_errno = NETDB_INTERNAL; return (-1); } @@ -1298,7 +1307,10 @@ * 'as is'. The threshold can be set with the "ndots" option. */ saved_herrno = -1; - if (dots >= _res.ndots) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_ndots = _res.ndots; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (dots >= _res_ndots) { ret = res_querydomain(name, NULL, class, type, answer, anslen); if (ret > 0) return (ret); @@ -1312,11 +1324,15 @@ * - there is at least one dot, there is no trailing dot, * and RES_DNSRCH is set. */ - if ((!dots && (_res.options & RES_DEFNAMES)) || - (dots && !trailing_dot && (_res.options & RES_DNSRCH))) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + _res_dnsrch = _res.dnsrch; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!dots && (_res_options & RES_DEFNAMES)) || + (dots && !trailing_dot && (_res_options & RES_DNSRCH))) { int done = 0; - for (domain = (const char * const *)_res.dnsrch; + for (domain = (const char * const *)_res_dnsrch; *domain && !done; domain++) { @@ -1365,7 +1381,10 @@ * if we got here for some reason other than DNSRCH, * we only wanted one iteration of the loop, so stop. */ - if (!(_res.options & RES_DNSRCH)) + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (!(_res_options & RES_DNSRCH)) done++; } } @@ -1411,14 +1430,21 @@ char nbuf[MAXDNAME]; const char *longname = nbuf; size_t n, d; + u_long _res_options; - if ((!name || !answer) || ((_res.options & RES_INIT) == 0 && res_init() == -1)) { + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if ((!name || !answer) || ((_res_options & RES_INIT) == 0 && res_init() == -1)) { h_errno = NETDB_INTERNAL; return (-1); } #ifdef DEBUG - if (_res.options & RES_DEBUG) + __UCLIBC_MUTEX_LOCK(__resolv_lock); + _res_options = _res.options; + __UCLIBC_MUTEX_UNLOCK(__resolv_lock); + if (_res_options & RES_DEBUG) printf(";; res_querydomain(%s, %s, %d, %d)\n", name, domain?domain:"", class, type); #endif From andersen at uclibc.org Thu Feb 1 17:37:44 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 17:37:44 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits Message-ID: <20070202013744.4A5CE485B8@busybox.net> Author: andersen Date: 2007-02-01 17:37:43 -0800 (Thu, 01 Feb 2007) New Revision: 17715 Log: add some ipv6 sockopts used by some folks Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/in.h Changeset: Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/in.h =================================================================== --- branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/in.h 2007-02-02 01:29:10 UTC (rev 17714) +++ branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/in.h 2007-02-02 01:37:43 UTC (rev 17715) @@ -115,6 +115,9 @@ #define IPV6_MTU_DISCOVER 23 #define IPV6_MTU 24 #define IPV6_RECVERR 25 +#define IPV6_V6ONLY 26 +#define IPV6_JOIN_ANYCAST 27 +#define IPV6_LEAVE_ANYCAST 28 #define SCM_SRCRT IPV6_RXSRCRT From vapier at uclibc.org Thu Feb 1 17:42:52 2007 From: vapier at uclibc.org (vapier at uclibc.org) Date: Thu, 1 Feb 2007 17:42:52 -0800 (PST) Subject: svn commit: trunk/uClibc Message-ID: <20070202014252.8FCA3485B8@busybox.net> Author: vapier Date: 2007-02-01 17:42:51 -0800 (Thu, 01 Feb 2007) New Revision: 17716 Log: replace bits/syscalls.h with a comment ... an empty file makes people think this is a bug rather than by design Modified: trunk/uClibc/Makefile.in Changeset: Modified: trunk/uClibc/Makefile.in =================================================================== --- trunk/uClibc/Makefile.in 2007-02-02 01:37:43 UTC (rev 17715) +++ trunk/uClibc/Makefile.in 2007-02-02 01:42:51 UTC (rev 17716) @@ -116,8 +116,7 @@ rm -f tar_exclude printf '#ifndef _LIBC_INTERNAL_H\n#define _LIBC_INTERNAL_H 1\n#endif\n' > \ $(PREFIX)$(DEVEL_PREFIX)include/libc-internal.h - # Don't export _syscall#() macros anymore; they're not portable - echo > $(PREFIX)$(DEVEL_PREFIX)include/bits/syscalls.h + echo '/* Dont use _syscall#() macros; use the syscall() function */' > $(PREFIX)$(DEVEL_PREFIX)include/bits/syscalls.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/dl-osinfo.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/_lfs_64.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/bits/uClibc_uintmaxtostr.h From andersen at uclibc.org Thu Feb 1 18:15:09 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 18:15:09 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/sysdeps/linux/mips Message-ID: <20070202021509.CA08C485C5@busybox.net> Author: andersen Date: 2007-02-01 18:15:09 -0800 (Thu, 01 Feb 2007) New Revision: 17718 Log: backport fix from svn trunk. Do not set SA_RESTORER when the application did not request it. Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/mips/sigaction.c Changeset: Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/mips/sigaction.c =================================================================== --- branches/uClibc_0_9_28/libc/sysdeps/linux/mips/sigaction.c 2007-02-02 01:57:24 UTC (rev 17717) +++ branches/uClibc_0_9_28/libc/sysdeps/linux/mips/sigaction.c 2007-02-02 02:15:09 UTC (rev 17718) @@ -27,9 +27,20 @@ #define SA_RESTORER 0x04000000 +extern __typeof(sigaction) __libc_sigaction; -#if defined __NR_rt_sigaction +#ifdef __NR_rt_sigaction +#if _MIPS_SIM != _ABIO32 + +# ifdef __NR_rt_sigreturn +static void restore_rt (void) asm ("__restore_rt"); +# endif +# ifdef __NR_sigreturn +static void restore (void) asm ("__restore"); +# endif +#endif + /* If ACT is not NULL, change the action for SIG to *ACT. If OACT is not NULL, put the old action for SIG in *OACT. */ int __libc_sigaction (int sig, const struct sigaction *act, struct sigaction *oact) @@ -48,10 +59,12 @@ kact.k_sa_handler = act->sa_handler; memcpy (&kact.sa_mask, &act->sa_mask, sizeof (kact.sa_mask)); kact.sa_flags = act->sa_flags; - - kact.sa_flags = act->sa_flags | SA_RESTORER; #ifdef HAVE_SA_RESTORER +# if _MIPS_SIM == _ABIO32 kact.sa_restorer = act->sa_restorer; +# else + kact.sa_restorer = &restore_rt; +# endif #endif } @@ -73,8 +86,7 @@ #else -#warning "Yes there is a warning here. Don't worry about it." -static void restore (void) asm ("__restore"); +extern void restore (void) asm ("__restore") attribute_hidden; /* If ACT is not NULL, change the action for SIG to *ACT. If OACT is not NULL, put the old action for SIG in *OACT. */ @@ -93,9 +105,12 @@ if (act) { kact.k_sa_handler = act->sa_handler; kact.sa_mask = act->sa_mask.__val[0]; - kact.sa_flags = act->sa_flags | SA_RESTORER; #ifdef HAVE_SA_RESTORER +# if _MIPS_SIM == _ABIO32 kact.sa_restorer = act->sa_restorer; +# else + kact.sa_restorer = &restore_rt; +# endif #endif } @@ -121,3 +136,31 @@ #endif weak_alias (__libc_sigaction, sigaction) + +/* NOTE: Please think twice before making any changes to the bits of + code below. GDB needs some intimate knowledge about it to + recognize them as signal trampolines, and make backtraces through + signal handlers work right. Important are both the names + (__restore_rt) and the exact instruction sequence. + If you ever feel the need to make any changes, please notify the + appropriate GDB maintainer. */ + +#define RESTORE(name, syscall) RESTORE2 (name, syscall) +#define RESTORE2(name, syscall) \ +asm \ + ( \ + ".align 4\n" \ + "__" #name ":\n" \ + " li $2, " #syscall "\n" \ + " syscall\n" \ + ); + +/* The return code for realtime-signals. */ +#if _MIPS_SIM != _ABIO32 +# ifdef __NR_rt_sigreturn +RESTORE (restore_rt, __NR_rt_sigreturn) +# endif +# ifdef __NR_sigreturn +RESTORE (restore, __NR_sigreturn) +# endif +#endif From andersen at uclibc.org Thu Feb 1 18:21:00 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 18:21:00 -0800 (PST) Subject: svn commit: trunk/uClibc Message-ID: <20070202022100.6C593485C5@busybox.net> Author: andersen Date: 2007-02-01 18:20:59 -0800 (Thu, 01 Feb 2007) New Revision: 17719 Log: use sane line breaks Modified: trunk/uClibc/Makefile.in Changeset: Modified: trunk/uClibc/Makefile.in =================================================================== --- trunk/uClibc/Makefile.in 2007-02-02 02:15:09 UTC (rev 17718) +++ trunk/uClibc/Makefile.in 2007-02-02 02:20:59 UTC (rev 17719) @@ -116,7 +116,8 @@ rm -f tar_exclude printf '#ifndef _LIBC_INTERNAL_H\n#define _LIBC_INTERNAL_H 1\n#endif\n' > \ $(PREFIX)$(DEVEL_PREFIX)include/libc-internal.h - echo '/* Dont use _syscall#() macros; use the syscall() function */' > $(PREFIX)$(DEVEL_PREFIX)include/bits/syscalls.h + echo '/* Dont use _syscall#() macros; use the syscall() function */' > \ + $(PREFIX)$(DEVEL_PREFIX)include/bits/syscalls.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/dl-osinfo.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/_lfs_64.h $(RM) $(PREFIX)$(DEVEL_PREFIX)include/bits/uClibc_uintmaxtostr.h From andersen at uclibc.org Thu Feb 1 18:24:30 2007 From: andersen at uclibc.org (andersen at uclibc.org) Date: Thu, 1 Feb 2007 18:24:30 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28: include/sys Message-ID: <20070202022430.95943485C5@busybox.net> Author: andersen Date: 2007-02-01 18:24:29 -0800 (Thu, 01 Feb 2007) New Revision: 17720 Log: As with uClibc trunk, note that the _syscall#() macros are for uClibc internal use only. User application code should use syscall() instead. Modified: branches/uClibc_0_9_28/Makefile branches/uClibc_0_9_28/include/sys/syscall.h Changeset: Modified: branches/uClibc_0_9_28/Makefile =================================================================== --- branches/uClibc_0_9_28/Makefile 2007-02-02 02:20:59 UTC (rev 17719) +++ branches/uClibc_0_9_28/Makefile 2007-02-02 02:24:29 UTC (rev 17720) @@ -165,6 +165,8 @@ fi ; \ tar -chf - include --exclude .svn --exclude CVS $$extra_exclude \ | tar -xf - -C $(PREFIX)$(DEVEL_PREFIX) + echo '/* Dont use _syscall#() macros; use the syscall() function */' > \ + $(PREFIX)$(DEVEL_PREFIX)include/bits/syscalls.h ifneq ($(strip $(UCLIBC_HAS_FLOATS)),y) # Remove floating point related headers since float support is disabled. $(RM) $(PREFIX)$(DEVEL_PREFIX)include/complex.h Modified: branches/uClibc_0_9_28/include/sys/syscall.h =================================================================== --- branches/uClibc_0_9_28/include/sys/syscall.h 2007-02-02 02:20:59 UTC (rev 17719) +++ branches/uClibc_0_9_28/include/sys/syscall.h 2007-02-02 02:24:29 UTC (rev 17720) @@ -2,24 +2,36 @@ This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Library General Public License as - published by the Free Software Foundation; either version 2 of the - License, or (at your option) any later version. + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Library General Public License for more details. + Lesser General Public License for more details. - You should have received a copy of the GNU Library General Public - License along with the GNU C Library; see the file COPYING.LIB. If not, - write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, - Boston, MA 02111-1307, USA. */ + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ #ifndef _SYSCALL_H #define _SYSCALL_H 1 -/* This file provides us with the nicely useful _syscall[0-5] macros. */ -#include +/* The _syscall#() macros are for uClibc internal use only. + * User application code should use syscall() instead. + * + * The kernel provided _syscall[0-6] macros from asm/unistd.h are not suitable + * for use in uClibc as they lack PIC support etc, so for uClibc we use our own + * local _syscall# macros to be certain all such variations are handled + * properly. + */ +#include +#include +#if defined _LIBC +# include #endif + +#endif From bugs at busybox.net Fri Feb 2 01:13:40 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Fri, 2 Feb 2007 01:13:40 -0800 Subject: [buildroot 0001188]: samba package build sometimes fails Message-ID: A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=1188 ====================================================================== Reported By: janlana Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1188 Category: Other Reproducibility: sometimes Severity: minor Priority: normal Status: assigned ====================================================================== Date Submitted: 01-31-2007 05:06 PST Last Modified: 02-02-2007 01:13 PST ====================================================================== Summary: samba package build sometimes fails Description: because a error in samba makefile the make -j... sometimes fails because include/proto.h is not generated yet. It happens in http://busybox.net/bugs/view.php?id=118#50% with -j3 on my computer. Suggested fix: change $(MAKE) to $(MAKE1) in samba.mk ====================================================================== ---------------------------------------------------------------------- janlana - 02-02-07 01:13 ---------------------------------------------------------------------- it looks that just $(MAKE1) does not solve the problem. Suggest to close the bug for now. I will submit new issue when I find a fix. Issue History Date Modified Username Field Change ====================================================================== 01-31-07 05:06 janlana New Issue 01-31-07 05:06 janlana Status new => assigned 01-31-07 05:06 janlana Assigned To => uClibc 02-02-07 01:13 janlana Note Added: 0002086 ====================================================================== From pkj at uclibc.org Fri Feb 2 04:32:28 2007 From: pkj at uclibc.org (pkj at uclibc.org) Date: Fri, 2 Feb 2007 04:32:28 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/sysdeps/linux/common/bits Message-ID: <20070202123228.D92FA485E3@busybox.net> Author: pkj Date: 2007-02-02 04:32:27 -0800 (Fri, 02 Feb 2007) New Revision: 17721 Log: Make sure struct stat is identical to struct stat64 if __USE_FILE_OFFSET64 is defined. This was the case for all architectures except those that use packed structures by default (i.e., CRIS). Modified: trunk/uClibc/libc/sysdeps/linux/common/bits/stat.h Changeset: Modified: trunk/uClibc/libc/sysdeps/linux/common/bits/stat.h =================================================================== --- trunk/uClibc/libc/sysdeps/linux/common/bits/stat.h 2007-02-02 02:24:29 UTC (rev 17720) +++ trunk/uClibc/libc/sysdeps/linux/common/bits/stat.h 2007-02-02 12:32:27 UTC (rev 17721) @@ -36,10 +36,11 @@ struct stat { __dev_t st_dev; /* Device. */ +#ifndef __USE_FILE_OFFSET64 unsigned short int __pad1; -#ifndef __USE_FILE_OFFSET64 __ino_t st_ino; /* File serial number. */ #else + unsigned int __pad1; __ino_t __st_ino; /* 32bit file serial number. */ #endif __mode_t st_mode; /* File mode. */ @@ -47,10 +48,11 @@ __uid_t st_uid; /* User ID of the file's owner. */ __gid_t st_gid; /* Group ID of the file's group.*/ __dev_t st_rdev; /* Device number, if device. */ +#ifndef __USE_FILE_OFFSET64 unsigned short int __pad2; -#ifndef __USE_FILE_OFFSET64 __off_t st_size; /* Size of file, in bytes. */ #else + unsigned int __pad2; __off64_t st_size; /* Size of file, in bytes. */ #endif __blksize_t st_blksize; /* Optimal block size for I/O. */ From pkj at uclibc.org Fri Feb 2 04:50:17 2007 From: pkj at uclibc.org (pkj at uclibc.org) Date: Fri, 2 Feb 2007 04:50:17 -0800 (PST) Subject: svn commit: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits Message-ID: <20070202125017.F363C485F0@busybox.net> Author: pkj Date: 2007-02-02 04:50:17 -0800 (Fri, 02 Feb 2007) New Revision: 17723 Log: Backported from the main trunk: * Make sure struct stat is identical to struct stat64 if __USE_FILE_OFFSET64 is defined. This was the case for all architectures except those that use packed structures by default (i.e., CRIS). Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/stat.h Changeset: Modified: branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/stat.h =================================================================== --- branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/stat.h 2007-02-02 12:41:16 UTC (rev 17722) +++ branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/stat.h 2007-02-02 12:50:17 UTC (rev 17723) @@ -38,10 +38,11 @@ struct stat { __dev_t st_dev; /* Device. */ +#ifndef __USE_FILE_OFFSET64 unsigned short int __pad1; -#ifndef __USE_FILE_OFFSET64 __ino_t st_ino; /* File serial number. */ #else + unsigned int __pad1; __ino_t __st_ino; /* 32bit file serial number. */ #endif __mode_t st_mode; /* File mode. */ @@ -49,10 +50,11 @@ __uid_t st_uid; /* User ID of the file's owner. */ __gid_t st_gid; /* Group ID of the file's group.*/ __dev_t st_rdev; /* Device number, if device. */ +#ifndef __USE_FILE_OFFSET64 unsigned short int __pad2; -#ifndef __USE_FILE_OFFSET64 __off_t st_size; /* Size of file, in bytes. */ #else + unsigned int __pad2; __off64_t st_size; /* Size of file, in bytes. */ #endif __blksize_t st_blksize; /* Optimal block size for I/O. */ From bugs at busybox.net Sat Feb 3 17:31:49 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sat, 3 Feb 2007 17:31:49 -0800 Subject: [buildroot 0000967]: SEGV when running openssh Message-ID: <58866aaf83d1f7aa10c2d88e9f197a79@bugs.uclibc.org> A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=967 ====================================================================== Reported By: lerichi Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 967 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: assigned ====================================================================== Date Submitted: 07-23-2006 14:59 PDT Last Modified: 02-03-2007 17:31 PST ====================================================================== Summary: SEGV when running openssh Description: I downloaded and did a standard build of the latest buildroot. Busybox works just fine, but openssh and all its helpers just SEGV. Apart from that, the other non-busybox proggies also die. So there must be a problem with some library. Just make a standard build (gcc-3.4.2), chroot into build_ARCH/root and call sshd. It should die :-( Any ideas? I have a suspicion that xstrdup() might be the cause. Regards- Richard ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-24-07 13:04 ---------------------------------------------------------------------- Try to strace this and perpahs provide a simple testcase for us to reproduce. ---------------------------------------------------------------------- gheeres - 02-01-07 13:53 ---------------------------------------------------------------------- Attached strace output. Build / configuration: uClibc-snapshost (01/26/2007) gcc-v4.1.1 openssh-v3.9p1 Same results attempting to use openssh-v4.5p1 If strace needs to be in different format, please provide instructions for generating. ---------------------------------------------------------------------- gheeres - 02-01-07 14:10 ---------------------------------------------------------------------- Older buildroot using: uClibc-0.9.28 (non-snapshot) gcc-v3.4.6 openssh-v3.9p1 or openssh-v4.5p1 Works correctly. ---------------------------------------------------------------------- gheeres - 02-03-07 17:31 ---------------------------------------------------------------------- Checked out newest uClibc build (http://busybox.net/bugs/view.php?id=#02012007), completely rebuilt entire toolchain and build files. SSH is no longer SEGV. Issue History Date Modified Username Field Change ====================================================================== 07-23-06 14:59 lerichi New Issue 07-23-06 14:59 lerichi Status new => assigned 07-23-06 14:59 lerichi Assigned To => uClibc 07-27-06 04:58 thrakhath Issue Monitored: thrakhath 01-24-07 13:04 bernhardf Note Added: 0002041 02-01-07 13:50 gheeres File Added: uClibc-snapshot-openssh-v3.9p1.strace 02-01-07 13:53 gheeres Note Added: 0002081 02-01-07 14:10 gheeres Note Added: 0002082 02-03-07 17:31 gheeres Note Added: 0002092 ====================================================================== From vapier at uclibc.org Sun Feb 4 01:24:57 2007 From: vapier at uclibc.org (vapier at uclibc.org) Date: Sun, 4 Feb 2007 01:24:57 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/sysdeps/linux/mips/bits Message-ID: <20070204092457.5B56F485CE@busybox.net> Author: vapier Date: 2007-02-04 01:24:56 -0800 (Sun, 04 Feb 2007) New Revision: 17753 Log: sync some stuff from glibc to fix test build failure Modified: trunk/uClibc/libc/sysdeps/linux/mips/bits/stat.h Changeset: Modified: trunk/uClibc/libc/sysdeps/linux/mips/bits/stat.h =================================================================== --- trunk/uClibc/libc/sysdeps/linux/mips/bits/stat.h 2007-02-04 02:41:57 UTC (rev 17752) +++ trunk/uClibc/libc/sysdeps/linux/mips/bits/stat.h 2007-02-04 09:24:56 UTC (rev 17753) @@ -1,4 +1,5 @@ -/* Copyright (C) 1992,95,96,97,98,99,2000,2001 Free Software Foundation, Inc. +/* Copyright (C) 1992, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004 + Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or @@ -35,7 +36,7 @@ #define _MKNOD_VER _MKNOD_VER_LINUX /* The bits defined below. */ -#if _MIPS_SIM == _MIPS_SIM_ABI32 +#if _MIPS_SIM == _ABIO32 /* Structure describing file characteristics. */ struct stat { @@ -65,11 +66,11 @@ * st_ctime but we don't have it under Linux. */ __time_t st_atime; /* Time of last access. */ - long int __reserved0; + unsigned long int st_atimensec; /* Nscecs of last access. */ __time_t st_mtime; /* Time of last modification. */ - long int __reserved1; + unsigned long int st_mtimensec; /* Nsecs of last modification. */ __time_t st_ctime; /* Time of last status change. */ - long int __reserved2; + unsigned long int st_ctimensec; /* Nsecs of last status change. */ __blksize_t st_blksize; /* Optimal block size for I/O. */ #ifndef __USE_FILE_OFFSET64 __blkcnt_t st_blocks; /* Number of 512-byte blocks allocated. */ @@ -110,7 +111,7 @@ #endif /* N32 || N64 */ #ifdef __USE_LARGEFILE64 -#if _MIPS_SIM == _MIPS_SIM_ABI32 +#if _MIPS_SIM == _ABIO32 struct stat64 { __dev_t st_dev; From vapier at uclibc.org Sun Feb 4 01:34:14 2007 From: vapier at uclibc.org (vapier at uclibc.org) Date: Sun, 4 Feb 2007 01:34:14 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/sysdeps/linux/sh Message-ID: <20070204093414.2ABAB4861B@busybox.net> Author: vapier Date: 2007-02-04 01:34:13 -0800 (Sun, 04 Feb 2007) New Revision: 17754 Log: fix compiling and touchup style Modified: trunk/uClibc/libc/sysdeps/linux/sh/pread_write.c Changeset: Modified: trunk/uClibc/libc/sysdeps/linux/sh/pread_write.c =================================================================== --- trunk/uClibc/libc/sysdeps/linux/sh/pread_write.c 2007-02-04 09:24:56 UTC (rev 17753) +++ trunk/uClibc/libc/sysdeps/linux/sh/pread_write.c 2007-02-04 09:34:13 UTC (rev 17754) @@ -1,12 +1,13 @@ -/* vi: set sw=4 ts=4: - * +/* vi: set sw=4 ts=4: */ +/* * Copyright (C) 2000-2006 Erik Andersen * * Licensed under the LGPL v2.1, see the file COPYING.LIB in this tarball. */ -/* Based in part on the files +/* + * Based in part on the files * ./sysdeps/unix/sysv/linux/pwrite.c, - * ./sysdeps/unix/sysv/linux/pread.c, + * ./sysdeps/unix/sysv/linux/pread.c, * sysdeps/posix/pread.c * sysdeps/posix/pwrite.c * from GNU libc 2.2.5, but reworked considerably... @@ -15,6 +16,7 @@ #include #include #include +#include #ifdef __NR_pread64 /* Newer kernels renamed but it's the same. */ # ifdef __NR_pread @@ -25,20 +27,20 @@ #ifdef __NR_pread extern __typeof(pread) __libc_pread; -# define __NR___syscall_pread __NR_pread -static inline _syscall6(ssize_t, __syscall_pread, int, fd, void *, buf, +# define __NR___syscall_pread __NR_pread +static inline _syscall6(ssize_t, __syscall_pread, int, fd, void *, buf, size_t, count, int, dummy, off_t, offset_hi, off_t, offset_lo); ssize_t __libc_pread(int fd, void *buf, size_t count, off_t offset) -{ +{ return(__syscall_pread(fd,buf,count,0,__LONG_LONG_PAIR((off_t)0,offset))); } weak_alias(__libc_pread,pread) -# ifdef __UCLIBC_HAS_LFS__ +# ifdef __UCLIBC_HAS_LFS__ extern __typeof(pread64) __libc_pread64; ssize_t __libc_pread64(int fd, void *buf, size_t count, off64_t offset) -{ +{ uint32_t low = offset & 0xffffffff; uint32_t high = offset >> 32; return(__syscall_pread(fd, buf, count, 0, __LONG_LONG_PAIR (high, low))); @@ -58,20 +60,20 @@ #ifdef __NR_pwrite extern __typeof(pwrite) __libc_pwrite; -# define __NR___syscall_pwrite __NR_pwrite -static inline _syscall6(ssize_t, __syscall_pwrite, int, fd, const void *, buf, +# define __NR___syscall_pwrite __NR_pwrite +static inline _syscall6(ssize_t, __syscall_pwrite, int, fd, const void *, buf, size_t, count, int, dummy, off_t, offset_hi, off_t, offset_lo); ssize_t __libc_pwrite(int fd, const void *buf, size_t count, off_t offset) -{ +{ return(__syscall_pwrite(fd,buf,count,0,__LONG_LONG_PAIR((off_t)0,offset))); } weak_alias(__libc_pwrite,pwrite) -# ifdef __UCLIBC_HAS_LFS__ +# ifdef __UCLIBC_HAS_LFS__ extern __typeof(pwrite64) __libc_pwrite64; ssize_t __libc_pwrite64(int fd, const void *buf, size_t count, off64_t offset) -{ +{ uint32_t low = offset & 0xffffffff; uint32_t high = offset >> 32; return(__syscall_pwrite(fd, buf, count, 0, __LONG_LONG_PAIR (high, low))); From vapier at uclibc.org Sun Feb 4 01:42:39 2007 From: vapier at uclibc.org (vapier at uclibc.org) Date: Sun, 4 Feb 2007 01:42:39 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/inet/rpc Message-ID: <20070204094239.AF7C74861B@busybox.net> Author: vapier Date: 2007-02-04 01:42:38 -0800 (Sun, 04 Feb 2007) New Revision: 17755 Log: the sun rpc code has too many alias violations to worry over now; just disable strict-aliasing until someone gets the balls to clean this stuff up Modified: trunk/uClibc/libc/inet/rpc/Makefile.in Changeset: Modified: trunk/uClibc/libc/inet/rpc/Makefile.in =================================================================== --- trunk/uClibc/libc/inet/rpc/Makefile.in 2007-02-04 09:34:13 UTC (rev 17754) +++ trunk/uClibc/libc/inet/rpc/Makefile.in 2007-02-04 09:42:38 UTC (rev 17755) @@ -5,6 +5,8 @@ # Licensed under the LGPL v2.1, see the file COPYING.LIB in this tarball. # +CFLAGS-rpc := -fno-strict-aliasing + ifneq ($(UCLIBC_HAS_FULL_RPC),y) # For now, only compile the stuff needed to do an NFS mount.... CSRC:= authunix_prot.c auth_none.c auth_unix.c bindresvport.c \ From vapier at uclibc.org Sun Feb 4 01:43:12 2007 From: vapier at uclibc.org (vapier at uclibc.org) Date: Sun, 4 Feb 2007 01:43:12 -0800 (PST) Subject: svn commit: trunk/uClibc/libc/sysdeps/linux/sh Message-ID: <20070204094312.224344861B@busybox.net> Author: vapier Date: 2007-02-04 01:43:11 -0800 (Sun, 04 Feb 2007) New Revision: 17756 Log: fixup includes so we compile again Modified: trunk/uClibc/libc/sysdeps/linux/sh/clone.S trunk/uClibc/libc/sysdeps/linux/sh/vfork.S Changeset: Modified: trunk/uClibc/libc/sysdeps/linux/sh/clone.S =================================================================== --- trunk/uClibc/libc/sysdeps/linux/sh/clone.S 2007-02-04 09:42:38 UTC (rev 17755) +++ trunk/uClibc/libc/sysdeps/linux/sh/clone.S 2007-02-04 09:43:11 UTC (rev 17756) @@ -20,10 +20,9 @@ and invokes a function in the right context after its all over. */ #include -#include -#define _ERRNO_H 1 +#include +#define _ERRNO_H #include -#define _SYSCALL_H #include Modified: trunk/uClibc/libc/sysdeps/linux/sh/vfork.S =================================================================== --- trunk/uClibc/libc/sysdeps/linux/sh/vfork.S 2007-02-04 09:42:38 UTC (rev 17755) +++ trunk/uClibc/libc/sysdeps/linux/sh/vfork.S 2007-02-04 09:43:11 UTC (rev 17756) @@ -21,9 +21,8 @@ */ #include -#define _SYSCALL_H -#include -#define _ERRNO_H 1 +#include +#define _ERRNO_H #include #include From bugs at busybox.net Sun Feb 4 03:24:40 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 03:24:40 -0800 Subject: [buildroot 0000967]: SEGV when running openssh Message-ID: The following issue requires your FEEDBACK. ====================================================================== http://busybox.net/bugs/view.php?id=967 ====================================================================== Reported By: lerichi Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 967 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: feedback ====================================================================== Date Submitted: 07-23-2006 14:59 PDT Last Modified: 02-04-2007 03:24 PST ====================================================================== Summary: SEGV when running openssh Description: I downloaded and did a standard build of the latest buildroot. Busybox works just fine, but openssh and all its helpers just SEGV. Apart from that, the other non-busybox proggies also die. So there must be a problem with some library. Just make a standard build (gcc-3.4.2), chroot into build_ARCH/root and call sshd. It should die :-( Any ideas? I have a suspicion that xstrdup() might be the cause. Regards- Richard ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-24-07 13:04 ---------------------------------------------------------------------- Try to strace this and perpahs provide a simple testcase for us to reproduce. ---------------------------------------------------------------------- gheeres - 02-01-07 13:53 ---------------------------------------------------------------------- Attached strace output. Build / configuration: uClibc-snapshost (01/26/2007) gcc-v4.1.1 openssh-v3.9p1 Same results attempting to use openssh-v4.5p1 If strace needs to be in different format, please provide instructions for generating. ---------------------------------------------------------------------- gheeres - 02-01-07 14:10 ---------------------------------------------------------------------- Older buildroot using: uClibc-0.9.28 (non-snapshot) gcc-v3.4.6 openssh-v3.9p1 or openssh-v4.5p1 Works correctly. ---------------------------------------------------------------------- gheeres - 02-03-07 17:31 ---------------------------------------------------------------------- Checked out newest uClibc build (http://busybox.net/bugs/view.php?id=#02012007), completely rebuilt entire toolchain and build files. SSH is no longer SEGV. ---------------------------------------------------------------------- bernhardf - 02-04-07 03:24 ---------------------------------------------------------------------- I consider this fixed now and shall close this issue within two weeks from now on, unless somebody comes up to let us know that 0.9.28.1 and current svn still has problems. Thanks, all, for your help with testing. Issue History Date Modified Username Field Change ====================================================================== 07-23-06 14:59 lerichi New Issue 07-23-06 14:59 lerichi Status new => assigned 07-23-06 14:59 lerichi Assigned To => uClibc 07-27-06 04:58 thrakhath Issue Monitored: thrakhath 01-24-07 13:04 bernhardf Note Added: 0002041 02-01-07 13:50 gheeres File Added: uClibc-snapshot-openssh-v3.9p1.strace 02-01-07 13:53 gheeres Note Added: 0002081 02-01-07 14:10 gheeres Note Added: 0002082 02-03-07 17:31 gheeres Note Added: 0002092 02-04-07 03:24 bernhardf Note Added: 0002094 02-04-07 03:24 bernhardf Status assigned => feedback ====================================================================== From bugs at busybox.net Sun Feb 4 08:39:15 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:39:15 -0800 Subject: [buildroot 0001191]: fix gcc, grub and lzma pckages spurious rebuild Message-ID: <486303e16fc1a446611d0fe4fa4ad659@busybox.net> The following issue has been CLOSED ====================================================================== http://busybox.net/bugs/view.php?id=1191 ====================================================================== Reported By: janlana Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1191 Category: Architecture Specific Reproducibility: always Severity: minor Priority: normal Status: closed Resolution: open Fixed in Version: ====================================================================== Date Submitted: 02-01-2007 03:15 PST Last Modified: 02-04-2007 08:39 PST ====================================================================== Summary: fix gcc, grub and lzma pckages spurious rebuild Description: see attached patches ====================================================================== ---------------------------------------------------------------------- bernhardf - 02-04-07 08:39 ---------------------------------------------------------------------- Fixed in svn r17763. Thanks! Issue History Date Modified Username Field Change ====================================================================== 02-01-07 03:15 janlana New Issue 02-01-07 03:15 janlana Status new => assigned 02-01-07 03:15 janlana Assigned To => uClibc 02-01-07 03:15 janlana File Added: package_lzma.patch 02-01-07 03:16 janlana File Added: toolchain_gcc.patch 02-01-07 03:16 janlana File Added: target_x86_grub.patch 02-04-07 08:39 bernhardf Status assigned => closed 02-04-07 08:39 bernhardf Note Added: 0002097 ====================================================================== From bugs at busybox.net Sun Feb 4 08:40:33 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:40:33 -0800 Subject: [buildroot 0001185]: 17669 won't compile Busybox Message-ID: The following issue has been CLOSED ====================================================================== http://busybox.net/bugs/view.php?id=1185 ====================================================================== Reported By: jcoppens Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1185 Category: Architecture Specific Reproducibility: always Severity: minor Priority: normal Status: closed Resolution: open Fixed in Version: ====================================================================== Date Submitted: 01-30-2007 12:32 PST Last Modified: 02-04-2007 08:40 PST ====================================================================== Summary: 17669 won't compile Busybox Description: I downloaded SVN 17669, before applying the patch from http://busybox.net/bugs/view.php?id=1182 This is surely unrelated, but now I cannot finish compilation of Busybox. I'm not sure this is before or after the above patch. John ====================================================================== ---------------------------------------------------------------------- bagland - 01-30-07 13:11 ---------------------------------------------------------------------- I had the same error when I updated to 17677, I had to change the configuration file for busybox to package/busybox/busybox-1.2.2.1.config Rgds, Benjamin ---------------------------------------------------------------------- jcoppens - 01-30-07 14:55 ---------------------------------------------------------------------- Great! Thanks. Just to be clear, maybe someone else hits the same problem. The file to edit is .config: # BR2_BUSYBOX_VERSION_1_0_1 is not set # BR2_BUSYBOX_VERSION_1_1_3 is not set BR2_BUSYBOX_VERSION_1_2_2_1=y # BR2_BUSYBOX_VERSION_1_4_0 is not set # BR2_PACKAGE_BUSYBOX_SNAPSHOT is not set BR2_PACKAGE_BUSYBOX_INSTALL_SYMLINKS=y BR2_PACKAGE_BUSYBOX_CONFIG="target/device/x86/i386/busybox.config" And the last line to be changed to BR2_PACKAGE_BUSYBOX_CONFIG="package/busybox/busybox-1.2.2.1.config" Thanks, Benjamin. (I suppose this should've happened automatically?) ---------------------------------------------------------------------- bernhardf - 02-04-07 08:40 ---------------------------------------------------------------------- This was fixed last week. Thanks for the report. Issue History Date Modified Username Field Change ====================================================================== 01-30-07 12:32 jcoppens New Issue 01-30-07 12:32 jcoppens Status new => assigned 01-30-07 12:32 jcoppens Assigned To => uClibc 01-30-07 13:11 bagland Note Added: 0002078 01-30-07 14:55 jcoppens Note Added: 0002079 02-04-07 08:40 bernhardf Status assigned => closed 02-04-07 08:40 bernhardf Note Added: 0002098 ====================================================================== From bugs at busybox.net Sun Feb 4 08:42:46 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:42:46 -0800 Subject: [buildroot 0001186]: 17669 - problem with number of arguments in sprintsigmask Message-ID: <3d5e7d8c53a6c619a4f1fcb04f887694@bugs.busybox.net> The following issue requires your FEEDBACK. ====================================================================== http://busybox.net/bugs/view.php?id=1186 ====================================================================== Reported By: jcoppens Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1186 Category: Architecture Specific Reproducibility: always Severity: minor Priority: normal Status: feedback ====================================================================== Date Submitted: 01-30-2007 17:13 PST Last Modified: 02-04-2007 08:42 PST ====================================================================== Summary: 17669 - problem with number of arguments in sprintsigmask Description: Thanks to all other suggestions... I hope I'm not tireing out everyone with these bugreports... Installed 17669 and applied the mod from Benjamin to compile busybox. Now I get the error below, which is probably due to some incorrect version somewhere. Suggestions? John ====================================================================== ---------------------------------------------------------------------- bernhardf - 02-04-07 08:42 ---------------------------------------------------------------------- Please retry with an unpatched, clean trunk of buildroot. If it still fails, then please provide your egrep "^(#|$)" .config of buildroot and uClibc, so i can try to reproduce this locally. Thanks, Issue History Date Modified Username Field Change ====================================================================== 01-30-07 17:13 jcoppens New Issue 01-30-07 17:13 jcoppens Status new => assigned 01-30-07 17:13 jcoppens Assigned To => uClibc 02-04-07 08:42 bernhardf Note Added: 0002099 02-04-07 08:42 bernhardf Status assigned => feedback ====================================================================== From bugs at busybox.net Sun Feb 4 08:44:37 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:44:37 -0800 Subject: [buildroot 0001182]: MTD: Missing definition of jint_16/jint_32 svn (17567) Message-ID: <09346da0d235da46aadd6333a21a3630@bugs.busybox.net> The following issue has been UPDATED. ====================================================================== http://busybox.net/bugs/view.php?id=1182 ====================================================================== Reported By: jcoppens Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1182 Category: Architecture Specific Reproducibility: always Severity: minor Priority: normal Status: assigned ====================================================================== Date Submitted: 01-28-2007 05:53 PST Last Modified: 02-04-2007 08:44 PST ====================================================================== Summary: MTD: Missing definition of jint_16/jint_32 svn (17567) Description: Definition of jint_16 and _32 is missing in flash_eraseall.c. (See error message below) I tried to add #include "jffs2.h", but that produces duplicate definitions. (second part - below) ====================================================================== ---------------------------------------------------------------------- jcoppens - 01-28-07 20:18 ---------------------------------------------------------------------- Sorry, that was of course jff3.h I tried to add (not jff2.h as commented above). Just a typo - the problem is still there. ---------------------------------------------------------------------- bagland - 01-30-07 04:55 ---------------------------------------------------------------------- Removing an include path fixes the problem for me (against rev 17540): Index: package/mtd/mtd.mk =================================================================== --- package/mtd/mtd.mk (revision 17540) +++ package/mtd/mtd.mk (working copy) @@ -116,7 +116,7 @@ MTD_BUILD_TARGETS := $(addprefix $(MTD_DIR)/util/, $(MTD_TARGETS_y)) $(MTD_BUILD_TARGETS): $(MTD_DIR)/.unpacked - $(MAKE) CC=$(TARGET_CC) CFLAGS="-I$(LINUX_HEADERS_DIR)/include -I$(MTD_DIR)/include $(TARGET_CFLAGS)" LDFLAGS="$(TARGET_LDFLAGS)" LINUXDIR=$(LINUX_DIR) -C $(MTD_DIR)/util + $(MAKE) CC=$(TARGET_CC) CFLAGS="-I$(MTD_DIR)/include $(TARGET_CFLAGS)" LDFLAGS="$(TARGET_LDFLAGS)" LINUXDIR=$(LINUX_DIR) -C $(MTD_DIR)/util MTD_TARGETS := $(addprefix $(TARGET_DIR)/usr/sbin/, $(MTD_TARGETS_y)) Rgds, Benjamin ---------------------------------------------------------------------- jcoppens - 01-30-07 12:34 ---------------------------------------------------------------------- Another issue popped up, and I'm not sure yet this patch solved the problem. (http://busybox.net/bugs/view.php?id=1185) Thanks John (will report if I get past the other problem) Issue History Date Modified Username Field Change ====================================================================== 01-28-07 05:53 jcoppens New Issue 01-28-07 05:53 jcoppens Status new => assigned 01-28-07 05:53 jcoppens Assigned To => uClibc 01-28-07 20:18 jcoppens Note Added: 0002065 01-30-07 04:55 bagland Note Added: 0002071 01-30-07 12:34 jcoppens Note Added: 0002077 02-04-07 08:44 bernhardf Summary Missing definition of jint_16/jint_32 svn (17567) => MTD: Missing definition of jint_16/jint_32 svn (17567) ====================================================================== From bugs at busybox.net Sun Feb 4 08:47:30 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:47:30 -0800 Subject: [buildroot 0001173]: initial make error "Unrecognised emulation mode: armelf_linux_eabi" Message-ID: <36a1be21995d8479d1e504de69e6a8b1@bugs.busybox.net> A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=1173 ====================================================================== Reported By: pferrell Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1173 Category: Architecture Specific Reproducibility: always Severity: block Priority: normal Status: assigned ====================================================================== Date Submitted: 01-24-2007 16:46 PST Last Modified: 02-04-2007 08:47 PST ====================================================================== Summary: initial make error "Unrecognised emulation mode: armelf_linux_eabi" Description: Initial make errors out as follows. ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-24-07 18:09 ---------------------------------------------------------------------- Please provide at least your .config. Also i think that this is a duplicate of a nofpu eabi report, so please check. ---------------------------------------------------------------------- stephaneC - 01-25-07 00:58 ---------------------------------------------------------------------- I have the same problem with gcc-3.4.6 and arm eabi : see bug 1114 ---------------------------------------------------------------------- pferrell - 01-29-07 12:12 ---------------------------------------------------------------------- I looked at bug 1114 and it seems to be the opposite. I make all of stephaneC's fixes, and the problem is the same. The make continues to show "armelf_linux_eabi" as the unregnized emulation, with "armelf_linux" and "armelf" as the accepted ones. -------- egrep -v "^(#|$)" .config BR2_HAVE_DOT_CONFIG=y BR2_arm=y BR2_xscale=y BR2_ARM_TYPE="ARM_XSCALE" BR2_ARM_EABI=y BR2_ARCH="arm" BR2_ENDIAN="LITTLE" BR2_WGET="toolchain/getter_script.sh" BR2_SVN="svn co" BR2_TAR_OPTIONS="" BR2_DL_DIR="$(BASE_DIR)/dl" BR2_SOURCEFORGE_MIRROR="easynews" BR2_STAGING_DIR="$(BUILD_DIR)/staging_dir" BR2_TOPDIR_PREFIX="" BR2_TOPDIR_SUFFIX="" BR2_GNU_BUILD_SUFFIX="pc-linux-gnu" BR2_GNU_TARGET_SUFFIX="linux-uclibc" BR2_JLEVEL=3 BR2_UCLIBC_VERSION_SNAPSHOT=y BR2_USE_UCLIBC_SNAPSHOT="20061006" BR2_PTHREADS_NATIVE=y BR2_BINUTILS_VERSION_2_16_1=y BR2_BINUTILS_VERSION="2.16.1" BR2_GCC_VERSION_4_1_1=y BR2_GCC_VERSION="4.1.1" BR2_EXTRA_GCC_CONFIG_OPTIONS="--without-fp" BR2_CCACHE=y BR2_LARGEFILE=y BR2_SOFT_FLOAT=y BR2_TARGET_OPTIMIZATION="-pipe -Os -march=armv5te -mtune=iwmmxt -Wa,-mcpu=iwmmxt" BR2_PACKAGE_BUSYBOX=y BR2_PACKAGE_BUSYBOX_INSTALL_SYMLINKS=y BR2_PACKAGE_BUSYBOX_CONFIG="target/device/Gumstix/basix-connex/busybox.config" BR2_PACKAGE_BLUEZ=y BR2_PACKAGE_BOA=y BR2_PACKAGE_BONJOUR=y BR2_PACKAGE_DROPBEAR=y BR2_PACKAGE_DROPBEAR_URANDOM=y BR2_PACKAGE_HOTPLUG=y BR2_PACKAGE_LIBSYSFS=y BR2_PACKAGE_LRZSZ=y BR2_PACKAGE_MODULE_INIT_TOOLS=y BR2_PACKAGE_MTD=y BR2_PACKAGE_MTD_ORIG=y BR2_PACKAGE_MTD_ORIG_STRING="mtd-utils-1.0.0.tar.gz" BR2_PACKAGE_MTD_FLASH_LOCK=y BR2_PACKAGE_MTD_FLASH_UNLOCK=y BR2_PACKAGE_MTD_FLASHCP=y BR2_PACKAGE_PCMCIAUTILS=y BR2_PACKAGE_PXAREGS=y BR2_QTE_TMAKE_VERSION="1.13" BR2_PACKAGE_UDEV=y BR2_PACKAGE_CF8385=y BR2_PACKAGE_WIRELESS_TOOLS=y BR2_PACKAGE_ZLIB=y BR2_TARGET_ROOTFS_JFFS2=y BR2_TARGET_ROOTFS_JFFS2_EBSIZE=0x20000 BR2_TARGET_ROOTFS_JFFS2_LE=y BR2_TARGET_ROOTFS_JFFS2_SQUASH=y BR2_TARGET_ROOTFS_JFFS2_SUMTOOL=y BR2_TARGET_ROOTFS_JFFS2_OUTPUT="$(IMAGE).jffs2" BR2_TARGET_ROOTFS_JFFS2_COPYTO="" BR2_TARGET_UBOOT=y BR2_TARGET_UBOOT_FWPRINT=y BR2_TARGET_GUMSTIX_BASIXCONNEX=y GUMSTIX_400MHZ=y BR2_PACKAGE_LINUX=y BR2_PACKAGE_ETC_GUMSTIX_RELEASE=y -------egrep -v "^(#|$)" ./toolchain_build_arm_nofpu/uClibc/.config TARGET_arm=y TARGET_ARCH="arm" FORCE_OPTIONS_FOR_ARCH=y CONFIG_ARM_EABI=y USE_BX=y CONFIG_ARM_XSCALE=y ARCH_ANY_ENDIAN=y ARCH_LITTLE_ENDIAN=y ARCH_WANTS_LITTLE_ENDIAN=y ARCH_HAS_MMU=y ARCH_USE_MMU=y UCLIBC_HAS_FLOATS=y UCLIBC_HAS_SOFT_FLOAT=y DO_C99_MATH=y KERNEL_SOURCE="/root/gumstix-buildroot/build_arm_nofpu/linux-2.6.18gum" HAVE_DOT_CONFIG=y DOPIC=y HAVE_SHARED=y FORCE_SHAREABLE_TEXT_SEGMENTS=y LDSO_LDD_SUPPORT=y LDSO_CACHE_SUPPORT=y LDSO_BASE_FILENAME="ld.so" UCLIBC_CTOR_DTOR=y UCLIBC_HAS_THREADS=y PTHREADS_DEBUG_SUPPORT=y LINUXTHREADS_OLD=y UCLIBC_HAS_LFS=y MALLOC=y MALLOC_GLIBC_COMPAT=y UCLIBC_DYNAMIC_ATEXIT=y UCLIBC_HAS_SHADOW=y UCLIBC_HAS___PROGNAME=y UNIX98PTY_ONLY=y ASSUME_DEVPTS=y UCLIBC_HAS_TM_EXTENSIONS=y UCLIBC_HAS_TZ_CACHING=y UCLIBC_HAS_TZ_FILE=y UCLIBC_HAS_TZ_FILE_READ_MANY=y UCLIBC_TZ_FILE_PATH="/etc/TZ" UCLIBC_USE_NETLINK=y UCLIBC_HAS_STRING_GENERIC_OPT=y UCLIBC_HAS_STRING_ARCH_OPT=y UCLIBC_HAS_CTYPE_TABLES=y UCLIBC_HAS_CTYPE_SIGNED=y UCLIBC_HAS_CTYPE_CHECKED=y UCLIBC_HAS_WCHAR=y UCLIBC_PRINTF_SCANF_POSITIONAL_ARGS=9 UCLIBC_HAS_STDIO_BUFSIZ_4096=y UCLIBC_HAS_STDIO_BUILTIN_BUFFER_NONE=y UCLIBC_HAS_STDIO_GETC_MACRO=y UCLIBC_HAS_STDIO_PUTC_MACRO=y UCLIBC_HAS_STDIO_AUTO_RW_TRANSITION=y UCLIBC_HAS_GLIBC_CUSTOM_STREAMS=y UCLIBC_HAS_ERRNO_MESSAGES=y UCLIBC_HAS_SIGNUM_MESSAGES=y UCLIBC_HAS_GNU_GETOPT=y UCLIBC_HAS_GNU_GETSUBOPT=y UCLIBC_HAS_REGEX=y UCLIBC_HAS_REGEX_OLD=y UCLIBC_HAS_FNMATCH=y UCLIBC_HAS_FNMATCH_OLD=y UCLIBC_HAS_WORDEXP=y UCLIBC_HAS_GLOB=y SHARED_LIB_LOADER_PREFIX="/lib" RUNTIME_PREFIX="/" DEVEL_PREFIX="/usr/" UCLIBC_HAS_SSP=y PROPOLICE_BLOCK_SEGV=y UCLIBC_BUILD_SSP=y UCLIBC_BUILD_RELRO=y UCLIBC_BUILD_NOW=y UCLIBC_BUILD_NOEXECSTACK=y CROSS_COMPILER_PREFIX="/root/gumstix-buildroot/build_arm_nofpu/staging_dir/bin/arm-linux-uclibcgnueabi-" DOSTRIP=y WARNINGS="-Wall" ---------------------------------------------------------------------- bernhardf - 01-29-07 14:16 ---------------------------------------------------------------------- Did any of you read the fine helptext of BR2_GNU_TARGET_SUFFIX ? ---------------------------------------------------------------------- stephaneC - 01-30-07 06:41 ---------------------------------------------------------------------- pferrell : BR2_GNU_TARGET_SUFFIX="linux-uclibc" [snip]/arm-linux-uclibcgnueabi-ld: unrecognised emulation mode: armelf_linux_eabi Supported emulations: armelf_linux armelf me : BR2_GNU_TARGET_SUFFIX="linux-uclibcgnueabi" [snip]/arm-linux-uclibcgnueabi/bin/ld: unrecognised emulation mode: armelf_linux Supported emulations: armelf_linux_eabi armelfb_linux_eabi I'm a little bit confuse about that ! If I understand the solution is : for pferrell : add "gnueabi" or "eabi" to BR2_GNU_TARGET_SUFFIX => BR2_GNU_TARGET_SUFFIX="linux-uclibcgnueabi" for me : remove "gnueabi" from BR2_GNU_TARGET_SUFFIX => BR2_GNU_TARGET_SUFFIX="linux-uclibc" correct ? (now I work with latest gcc 4.2 and it's ok) ---------------------------------------------------------------------- pferrell - 01-30-07 08:42 ---------------------------------------------------------------------- bernhardf, I read the fine help text, changed my BR2_GNU_TARGET_SUFFIX, and re-ran with the same error. I then deleted my entire buildroot directory, re-did it from stratch and still get the same result. Nothing I change seems to have any effect on the "Supported Emulations". The only other notes I can raise are that I am using a VMWare emulator, and have tried fedora core 5 and 6. /root/gumstix-buildroot/build_arm_nofpu/staging_dir/bin/arm-linux-uclibcgnueabi- ld: unrecognised emulation mode: armelf_linux_eabi Supported emulations: armelf_linux armelf collect2: ld returned 1 exit status make[4]: *** [libgcc_s.so] Error 1 make[4]: Leaving directory `/root/gumstix-buildroot/toolchain_build_arm_nofpu/gc c-4.1.1-final/gcc' make[3]: *** [stmp-multilib] Error 2 make[3]: Leaving directory `/root/gumstix-buildroot/toolchain_build_arm_nofpu/gc c-4.1.1-final/gcc' make[2]: *** [all-gcc] Error 2 make[2]: Leaving directory `/root/gumstix-buildroot/toolchain_build_arm_nofpu/gc c-4.1.1-final' make[1]: *** [all] Error 2 make[1]: Leaving directory `/root/gumstix-buildroot/toolchain_build_arm_nofpu/gc c-4.1.1-final' make: *** [/root/gumstix-buildroot/toolchain_build_arm_nofpu/gcc-4.1.1-final/.co mpiled] Error 2 [root at localhost gumstix-buildroot]# egrep -v "^(#|$)" .config BR2_HAVE_DOT_CONFIG=y BR2_arm=y BR2_xscale=y BR2_ARM_TYPE="ARM_XSCALE" BR2_ARM_EABI=y BR2_ARCH="arm" BR2_ENDIAN="LITTLE" BR2_WGET="toolchain/getter_script.sh" BR2_SVN="svn co" BR2_TAR_OPTIONS="" BR2_DL_DIR="$(BASE_DIR)/dl" BR2_SOURCEFORGE_MIRROR="easynews" BR2_STAGING_DIR="$(BUILD_DIR)/staging_dir" BR2_TOPDIR_PREFIX="" BR2_TOPDIR_SUFFIX="" BR2_GNU_BUILD_SUFFIX="pc-linux-gnu" BR2_GNU_TARGET_SUFFIX="linux-uclibcgnueabi" BR2_JLEVEL=3 BR2_UCLIBC_VERSION_SNAPSHOT=y BR2_USE_UCLIBC_SNAPSHOT="20061006" BR2_PTHREADS_NATIVE=y BR2_BINUTILS_VERSION_2_16_1=y BR2_BINUTILS_VERSION="2.16.1" BR2_GCC_VERSION_4_1_1=y BR2_GCC_VERSION="4.1.1" BR2_EXTRA_GCC_CONFIG_OPTIONS="--without-fp" BR2_CCACHE=y BR2_LARGEFILE=y BR2_SOFT_FLOAT=y BR2_TARGET_OPTIMIZATION="-pipe -Os -march=armv5te -mtune=iwmmxt -Wa,-mcpu=iwmmxt" BR2_PACKAGE_BUSYBOX=y BR2_PACKAGE_BUSYBOX_INSTALL_SYMLINKS=y BR2_PACKAGE_BUSYBOX_CONFIG="target/device/Gumstix/basix-connex/busybox.config" BR2_PACKAGE_BLUEZ=y BR2_PACKAGE_BOA=y BR2_PACKAGE_BONJOUR=y BR2_PACKAGE_DROPBEAR=y BR2_PACKAGE_DROPBEAR_URANDOM=y BR2_PACKAGE_HOTPLUG=y BR2_PACKAGE_LIBSYSFS=y BR2_PACKAGE_LRZSZ=y BR2_PACKAGE_MODULE_INIT_TOOLS=y BR2_PACKAGE_MTD=y BR2_PACKAGE_MTD_ORIG=y BR2_PACKAGE_MTD_ORIG_STRING="mtd-utils-1.0.0.tar.gz" BR2_PACKAGE_MTD_FLASH_LOCK=y BR2_PACKAGE_MTD_FLASH_UNLOCK=y BR2_PACKAGE_MTD_FLASHCP=y BR2_PACKAGE_PCMCIAUTILS=y BR2_PACKAGE_PXAREGS=y BR2_QTE_TMAKE_VERSION="1.13" BR2_PACKAGE_UDEV=y BR2_PACKAGE_CF8385=y BR2_PACKAGE_WIRELESS_TOOLS=y BR2_PACKAGE_ZLIB=y BR2_TARGET_ROOTFS_JFFS2=y BR2_TARGET_ROOTFS_JFFS2_EBSIZE=0x20000 BR2_TARGET_ROOTFS_JFFS2_LE=y BR2_TARGET_ROOTFS_JFFS2_SQUASH=y BR2_TARGET_ROOTFS_JFFS2_SUMTOOL=y BR2_TARGET_ROOTFS_JFFS2_OUTPUT="$(IMAGE).jffs2" BR2_TARGET_ROOTFS_JFFS2_COPYTO="" BR2_TARGET_UBOOT=y BR2_TARGET_UBOOT_FWPRINT=y BR2_TARGET_GUMSTIX_BASIXCONNEX=y GUMSTIX_400MHZ=y BR2_PACKAGE_LINUX=y BR2_PACKAGE_ETC_GUMSTIX_RELEASE=y ---------------------------------------------------------------------- bernhardf - 02-04-07 08:47 ---------------------------------------------------------------------- During the last week, i changed quite a bit in that area. arm for an integrator with EABI now builds for me flawlessly. An XScale EABI with a default for big-endian builds too, fwiw. Please svn up and rebuild and let me know if that's better now. Issue History Date Modified Username Field Change ====================================================================== 01-24-07 16:46 pferrell New Issue 01-24-07 16:46 pferrell Status new => assigned 01-24-07 16:46 pferrell Assigned To => uClibc 01-24-07 16:46 pferrell File Added: emulator error txt 01-24-07 18:09 bernhardf Note Added: 0002042 01-25-07 00:58 stephaneC Note Added: 0002044 01-29-07 12:12 pferrell Note Added: 0002067 01-29-07 14:16 bernhardf Note Added: 0002069 01-30-07 06:41 stephaneC Note Added: 0002072 01-30-07 08:27 pferrell Note Added: 0002074 01-30-07 08:29 pferrell Note Edited: 0002074 01-30-07 08:42 pferrell Note Edited: 0002074 02-04-07 08:47 bernhardf Note Added: 0002100 ====================================================================== From bugs at busybox.net Sun Feb 4 08:49:30 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:49:30 -0800 Subject: [buildroot 0001152]: gcc toolchain built for target fails Message-ID: <8435f10a7d5633874b2d89af80e1f059@bugs.busybox.net> The following issue requires your FEEDBACK. ====================================================================== http://busybox.net/bugs/view.php?id=1152 ====================================================================== Reported By: Steven_Carr Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1152 Category: Other Reproducibility: always Severity: minor Priority: normal Status: feedback ====================================================================== Date Submitted: 01-09-2007 22:49 PST Last Modified: 02-04-2007 08:49 PST ====================================================================== Summary: gcc toolchain built for target fails Description: When building a "Generic development system" the compile fails. I believe this to be a typo in the gcc-uclibc-3.x.mk file. Please see the Additional Information for potential patch. ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-22-07 11:36 ---------------------------------------------------------------------- What were the exact steps you did to get to this? unpack snapshot tarball / checkout source, step into the buildroot dir and then.. ---------------------------------------------------------------------- bernhardf - 02-04-07 08:49 ---------------------------------------------------------------------- I will close this issue unless i receive instructions on how to reproduce this within a week. Thanks, Issue History Date Modified Username Field Change ====================================================================== 01-09-07 22:49 Steven_Carr New Issue 01-09-07 22:49 Steven_Carr Status new => assigned 01-09-07 22:49 Steven_Carr Assigned To => uClibc 01-22-07 11:36 bernhardf Note Added: 0002032 02-04-07 08:49 bernhardf Note Added: 0002101 02-04-07 08:49 bernhardf Status assigned => feedback ====================================================================== From bugs at busybox.net Sun Feb 4 08:50:43 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 08:50:43 -0800 Subject: [buildroot 0000912]: Can't make -- bails with error: Unknown floating point type used in --with-float=no Message-ID: <9c11aef0c592f0417db755e2982ae6bc@bugs.busybox.net> A NOTE has been added to this issue. ====================================================================== http://busybox.net/bugs/view.php?id=912 ====================================================================== Reported By: bowerymarc Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 912 Category: Architecture Specific Reproducibility: always Severity: block Priority: normal Status: feedback ====================================================================== Date Submitted: 06-19-2006 16:20 PDT Last Modified: 02-04-2007 08:50 PST ====================================================================== Summary: Can't make -- bails with error: Unknown floating point type used in --with-float=no Description: trying to compile buildroot with GCC 4.2 and ARM920T (config file attached) and getting this error every time: checking for clock_t... yes checking for .preinit_array/.init_array/.fini_array support... yes checking if mkdir takes one argument... no For real value is For real value is Unknown floating point type used in --with-float=no make[1]: *** [configure-gcc] Error 1 make[1]: Leaving directory `/home/marc/buildroot/toolchain_build_arm_nofpu/gcc-4.2-initial' make: *** [/home/marc/buildroot/toolchain_build_arm_nofpu/gcc-4.2-initial/.compiled] Error 2 ====================================================================== ---------------------------------------------------------------------- bernhardf - 01-21-07 06:43 ---------------------------------------------------------------------- Can you retry with current trunk (revision 17434 or later) and let me know if this works now for gcc-4.2, latest binutils and latest kernel-headers? Thanks, ---------------------------------------------------------------------- bernhardf - 02-04-07 08:50 ---------------------------------------------------------------------- Ping. Please update, retry and let me know if that works any better now or not. Issue History Date Modified Username Field Change ====================================================================== 06-19-06 16:20 bowerymarc New Issue 06-19-06 16:20 bowerymarc Status new => assigned 06-19-06 16:20 bowerymarc Assigned To => uClibc 06-19-06 16:20 bowerymarc File Added: current.config 01-21-07 06:43 bernhardf Note Added: 0002017 01-21-07 06:43 bernhardf Status assigned => feedback 02-04-07 08:50 bernhardf Note Added: 0002102 ====================================================================== From bugs at busybox.net Sun Feb 4 09:05:00 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:05:00 -0800 Subject: [buildroot 0000861]: buildroot jail unable to compile busybox --- "missing" output directory Message-ID: <490836a61268c3650c7b5588ba44baba@bugs.busybox.net> The following issue has been CLOSED ====================================================================== http://busybox.net/bugs/view.php?id=861 ====================================================================== Reported By: zugvogel Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 861 Category: Other Reproducibility: always Severity: major Priority: normal Status: closed Resolution: open Fixed in Version: ====================================================================== Date Submitted: 05-01-2006 19:33 PDT Last Modified: 02-04-2007 09:04 PST ====================================================================== Summary: buildroot jail unable to compile busybox --- "missing" output directory Description: Background: I am attempting to use buildroot as a root jail for making components to be linked against uClibc. For the jail, I am using an unwrapped copy of the ext2 target file system as produced by buildroot. Symptom: I am attempting to rebuild busybox inside the jail with my own setting. Unfortunately, "make menuconfig" (or just "make" with a planted .config) fails with the following ominous error: /tmp/busybox $ make Makefile:72: *** output directory "/tmp/busybox" does not exist. Stop. The listed output directory does exist (relative to chroot) as I am currently in it. Disabling the check in the makefile results in a segfault instead--- so I would guess that the "missing" output directory is just a symptom of something more sinister going on. ====================================================================== ---------------------------------------------------------------------- zugvogel - 05-04-06 18:34 ---------------------------------------------------------------------- Tried the same trick on a plain vanilla Debian 3.1r1. Same result (i.e. I get a toolchain that does not compile busybox). Will now try to reproduce the toolchain using the busybox/uClibc snapshots that were used for the "preassembled" i386 toolchain (basically, a sanity check). Will update once I have some results. Frank ---------------------------------------------------------------------- bernhardf - 12-21-06 04:05 ---------------------------------------------------------------------- Any update on this? busybox now has approached 1.3.0 and buildroot recently received quite some changes, so please retest. ---------------------------------------------------------------------- bernhardf - 02-04-07 09:04 ---------------------------------------------------------------------- Assumed to be fixed with current setups. Issue History Date Modified Username Field Change ====================================================================== 05-01-06 19:33 zugvogel New Issue 05-01-06 19:33 zugvogel Status new => assigned 05-01-06 19:33 zugvogel Assigned To => uClibc 05-04-06 18:34 zugvogel Note Added: 0001344 12-21-06 04:05 bernhardf Note Added: 0001903 12-21-06 04:05 bernhardf Status assigned => feedback 02-04-07 09:04 bernhardf Status feedback => closed 02-04-07 09:04 bernhardf Note Added: 0002103 ====================================================================== From bugs at busybox.net Sun Feb 4 09:05:45 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:05:45 -0800 Subject: [buildroot 0001055]: ccache-2.4 causes 'make clean' to fail Message-ID: The following issue requires your FEEDBACK. ====================================================================== http://busybox.net/bugs/view.php?id=1055 ====================================================================== Reported By: agaffney Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1055 Category: Other Reproducibility: always Severity: minor Priority: normal Status: feedback ====================================================================== Date Submitted: 09-27-2006 17:06 PDT Last Modified: 02-04-2007 09:05 PST ====================================================================== Summary: ccache-2.4 causes 'make clean' to fail Description: When running 'make clean' from the root of my buildroot svn checkout, I get: make -j4 -C /root/405gp/buildroot/toolchain_build_powerpc_nofpu/ccache-2.4 uninstall make[1]: Entering directory `/root/405gp/buildroot/toolchain_build_powerpc_nofpu/ccache-2.4' make[1]: *** No rule to make target `uninstall'. Stop. make[1]: Leaving directory `/root/405gp/buildroot/toolchain_build_powerpc_nofpu/ccache-2.4' make: *** [ccache-clean] Error 2 ====================================================================== ---------------------------------------------------------------------- bernhardf - 02-04-07 09:05 ---------------------------------------------------------------------- Should be fixed by andersee yesterday. Please retest. Issue History Date Modified Username Field Change ====================================================================== 09-27-06 17:06 agaffney New Issue 09-27-06 17:06 agaffney Status new => assigned 09-27-06 17:06 agaffney Assigned To => uClibc 10-09-06 12:20 jruere Issue Monitored: jruere 02-04-07 09:05 bernhardf Note Added: 0002104 02-04-07 09:05 bernhardf Status assigned => feedback ====================================================================== From bugs at busybox.net Sun Feb 4 09:06:56 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:06:56 -0800 Subject: [buildroot 0001010]: Build-root make error - missing shadow.h Message-ID: The following issue has been CLOSED ====================================================================== http://busybox.net/bugs/view.php?id=1010 ====================================================================== Reported By: mboutte Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1010 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: closed Resolution: reopened Fixed in Version: ====================================================================== Date Submitted: 08-27-2006 15:57 PDT Last Modified: 02-04-2007 09:06 PST ====================================================================== Summary: Build-root make error - missing shadow.h Description: I am new to build-root so forgive my possible lack of knowledge. I am attempting a new run of build-root for an x86 host to powerpc target. As a first attempt I left most of the defaults as-is, but used fairly new gcc 4.1.1 headers and a linux 2.6 type system. The initial make menuconfig went OK, and the svn checkout seemed to work fine during the make process too, except that after about 15 minutes of work the compilation stops with an error -- "/home/mike/projects/software/snip875/build-tools/uclibc/buildroot/build_powerpc /busybox/include/shadow_.h:23:20: error: shadow.h: No such file or directory" I googled your site and see other mentions of this error, but nothing I try seems to resolve it, including older gcc, linux 2.4, enabling or disableing shadow password support from the make config in the busybox directory. I also tried a make clean during this. I'm lost and have not more ideas now. Note that svn checked out the new vresion 0.9.28. Mike ====================================================================== ---------------------------------------------------------------------- vapier - 08-28-06 11:19 ---------------------------------------------------------------------- your uClibc is configured without shadow support ... that means you need to either enable shadow support in uClibc or disable everything in busybox that requries shadow ---------------------------------------------------------------------- mboutte - 08-28-06 14:38 ---------------------------------------------------------------------- Vapier, Not to belabor this issue, but I had tried modifying the busybox configuration by going to the busybox folder and running "make menuconfig" to no avail. On more testing I saw when I went back to busybox configuration after attempting a new build, that the configuration had reverted to that before my changes. A more careful read of your docs showed copying the new busybox configuration back to the package/busybox/busybox.config file. When I did that the build worked. So, for me the problem was that I assumed that using make menuconfig on busybox actually modified the configuration that build-root would use. Again, this is my first attempt at using this incredible tool, and I was unaware of that requirement. ---------------------------------------------------------------------- bernhardf - 02-04-07 09:06 ---------------------------------------------------------------------- This was a bug in busybox that was fixed in the 1.4.x series. Issue History Date Modified Username Field Change ====================================================================== 08-27-06 15:57 mboutte New Issue 08-27-06 15:57 mboutte Status new => assigned 08-27-06 15:57 mboutte Assigned To => uClibc 08-28-06 11:19 vapier Note Added: 0001596 08-28-06 11:19 vapier Status assigned => closed 08-28-06 11:19 vapier Resolution open => no change required 08-28-06 14:38 mboutte Status closed => feedback 08-28-06 14:38 mboutte Resolution no change required => reopened 08-28-06 14:38 mboutte Note Added: 0001597 02-04-07 09:06 bernhardf Status feedback => closed 02-04-07 09:06 bernhardf Note Added: 0002105 ====================================================================== From bugs at busybox.net Sun Feb 4 09:09:03 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:09:03 -0800 Subject: [buildroot 0001114]: Cannot build ARM eabi toolchain Message-ID: The following issue requires your FEEDBACK. ====================================================================== http://busybox.net/bugs/view.php?id=1114 ====================================================================== Reported By: mikewhit Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1114 Category: Architecture Specific Reproducibility: always Severity: major Priority: normal Status: feedback ====================================================================== Date Submitted: 12-13-2006 01:58 PST Last Modified: 02-04-2007 09:09 PST ====================================================================== Summary: Cannot build ARM eabi toolchain Description: Building of arm-eabi toolchain fails due to missing linker emulation: >> /home/fred/buildroot/build_arm_nofpu/staging_dir/arm-linux-uclibcgnueabi/bin/ld: unrecognised emulation mode: armelf_linux Supported emulations: armelf_linux_eabi collect2: ld returned 1 exit status << ====================================================================== ---------------------------------------------------------------------- mikewhit - 12-13-06 02:43 ---------------------------------------------------------------------- This is actually a Blocking issue, since cannot use Buildroot without EABI support. ---------------------------------------------------------------------- bernhardf - 12-21-06 03:50 ---------------------------------------------------------------------- post your egrep -v "^(#|$)" .config egrep -v "^(#|$)" uClibc.config egrep -v "^(#|$)" busybox.config egrep -v "^(#|$)" kernel.config So i can try to reproduce it locally. ---------------------------------------------------------------------- stephaneC - 01-09-07 05:06 ---------------------------------------------------------------------- I have the same problem with : buildroot rev 17204 ARM920T / EABI / gcc-3.4.6 see below for my .config and uClibc.config 1st error is bug 1131 : file toolchain_build_arm_nofpu/uClibc/libc/misc/glob/glob.c line 364 #if !defined COMPILE_GLOB64 to #if defined __UCLIBC_HAS_LFS__ && !defined COMPILE_GLOB64 2nd error : --- [snip]buildroot/build_arm_nofpu/staging_dir/arm-linux-uclibcgnueabi/bin/ld: unrecognised emulation mode: armelf_linux Supported emulations: armelf_linux_eabi armelfb_linux_eabi collect2: ld returned 1 exit status make[3]: *** [libgcc/./_udivsi3.oS] Erreur 1 make[3]: quittant le r?pertoire ? [snip]buildroot/toolchain_build_arm_nofpu/gcc-3.4.6-final/gcc ? make[2]: *** [libgcc.a] Erreur 2 --- file toolchain_build_arm_nofpu/gcc-3.4.6/gcc/config/arm/linux-elf.h line 44 #define TARGET_LINKER_EMULATION "armelf_linux" to #define TARGET_LINKER_EMULATION "armelf_linux_eabi" after that all is ok I hope this can help you Stephane -------------- egrep -v "^(#|$)" .config ------------------- BR2_HAVE_DOT_CONFIG=y BR2_arm=y BR2_arm920t=y BR2_ARM_TYPE="ARM920T" BR2_ARM_EABI=y BR2_ARCH="arm" BR2_ENDIAN="LITTLE" BR2_WGET="wget --passive-ftp" BR2_SVN="svn co" BR2_ZCAT="zcat" BR2_BZCAT="bzcat" BR2_TAR_OPTIONS="" BR2_DL_DIR="$(BASE_DIR)/dl" BR2_SOURCEFORGE_MIRROR="easynews" BR2_STAGING_DIR="$(BUILD_DIR)/staging_dir" BR2_TOPDIR_PREFIX="" BR2_TOPDIR_SUFFIX="" BR2_GNU_BUILD_SUFFIX="pc-linux-gnu" BR2_GNU_TARGET_SUFFIX="linux-uclibcgnueabi" BR2_JLEVEL=1 BR2_KERNEL_HEADERS_2_6_19_1=y BR2_DEFAULT_KERNEL_HEADERS="2.6.19.1" BR2_UCLIBC_VERSION_SNAPSHOT=y BR2_USE_UCLIBC_SNAPSHOT="snapshot" BR2_PTHREADS_OLD=y BR2_BINUTILS_VERSION_2_17_50_0_8=y BR2_BINUTILS_VERSION="2.17.50.0.8" BR2_EXTRA_BINUTILS_CONFIG_OPTIONS="" BR2_GCC_VERSION_3_4_6=y BR2_GCC_VERSION="3.4.6" BR2_EXTRA_GCC_CONFIG_OPTIONS="" BR2_GCC_SHARED_LIBGCC=y BR2_ENABLE_MULTILIB=y BR2_SOFT_FLOAT=y BR2_TARGET_OPTIMIZATION="-Os -pipe" BR2_CROSS_TOOLCHAIN_TARGET_UTILS=y BR2_PACKAGE_BUSYBOX=y BR2_PACKAGE_BUSYBOX_SNAPSHOT=y BR2_PACKAGE_BUSYBOX_INSTALL_SYMLINKS=y BR2_PACKAGE_BUSYBOX_CONFIG="/home/be/Buildroot/config_busybox" BR2_TARGET_ROOTFS_EXT2=y BR2_TARGET_ROOTFS_EXT2_BLOCKS=0 BR2_TARGET_ROOTFS_EXT2_INODES=0 BR2_TARGET_ROOTFS_EXT2_RESBLKS=0 BR2_TARGET_ROOTFS_EXT2_SQUASH=y BR2_TARGET_ROOTFS_EXT2_OUTPUT="$(IMAGE).ext2" BR2_TARGET_ROOTFS_EXT2_NONE=y BR2_TARGET_ROOTFS_EXT2_COPYTO="" ----------------------------------------------------------------------------------------- -------------- egrep -v "^(#|$)" ./toolchain_build_arm_nofpu/uClibc/.config ---------------- TARGET_arm=y TARGET_ARCH="arm" FORCE_OPTIONS_FOR_ARCH=y CONFIG_ARM_EABI=y USE_BX=y CONFIG_ARM920T=y ARCH_ANY_ENDIAN=y ARCH_LITTLE_ENDIAN=y ARCH_WANTS_LITTLE_ENDIAN=y ARCH_HAS_MMU=y ARCH_USE_MMU=y UCLIBC_HAS_FLOATS=y UCLIBC_HAS_SOFT_FLOAT=y DO_C99_MATH=y KERNEL_HEADERS="/home/be/Buildroot/buildroot_EABI_3.4.6_bug/buildroot/toolchain_build_arm_nofpu/linux/include" HAVE_DOT_CONFIG=y DOPIC=y HAVE_SHARED=y LDSO_LDD_SUPPORT=y LDSO_CACHE_SUPPORT=y LDSO_BASE_FILENAME="ld.so" UCLIBC_CTOR_DTOR=y UCLIBC_HAS_THREADS=y PTHREADS_DEBUG_SUPPORT=y LINUXTHREADS_OLD=y MALLOC_STANDARD=y MALLOC_GLIBC_COMPAT=y UCLIBC_DYNAMIC_ATEXIT=y UCLIBC_SUSV3_LEGACY_MACROS=y UCLIBC_HAS_SHADOW=y UCLIBC_HAS___PROGNAME=y ASSUME_DEVPTS=y UCLIBC_HAS_TM_EXTENSIONS=y UCLIBC_HAS_TZ_CACHING=y UCLIBC_HAS_TZ_FILE=y UCLIBC_HAS_TZ_FILE_READ_MANY=y UCLIBC_TZ_FILE_PATH="/etc/TZ" UCLIBC_HAS_IPV6=y UCLIBC_HAS_RPC=y UCLIBC_HAS_FULL_RPC=y UCLIBC_HAS_REENTRANT_RPC=y UCLIBC_HAS_STRING_GENERIC_OPT=y UCLIBC_HAS_STRING_ARCH_OPT=y UCLIBC_HAS_CTYPE_TABLES=y UCLIBC_HAS_CTYPE_SIGNED=y UCLIBC_HAS_CTYPE_CHECKED=y UCLIBC_HAS_WCHAR=y UCLIBC_HAS_HEXADECIMAL_FLOATS=y UCLIBC_HAS_GLIBC_CUSTOM_PRINTF=y UCLIBC_PRINTF_SCANF_POSITIONAL_ARGS=9 UCLIBC_HAS_SCANF_GLIBC_A_FLAG=y UCLIBC_HAS_STDIO_BUFSIZ_4096=y UCLIBC_HAS_STDIO_BUILTIN_BUFFER_NONE=y UCLIBC_HAS_STDIO_GETC_MACRO=y UCLIBC_HAS_STDIO_PUTC_MACRO=y UCLIBC_HAS_STDIO_AUTO_RW_TRANSITION=y UCLIBC_HAS_FOPEN_EXCLUSIVE_MODE=y UCLIBC_HAS_GLIBC_CUSTOM_STREAMS=y UCLIBC_HAS_PRINTF_M_SPEC=y UCLIBC_HAS_ERRNO_MESSAGES=y UCLIBC_HAS_SIGNUM_MESSAGES=y UCLIBC_HAS_GNU_GETOPT=y UCLIBC_HAS_GNU_GETSUBOPT=y UCLIBC_HAS_REGEX=y UCLIBC_HAS_REGEX_OLD=y UCLIBC_HAS_FNMATCH=y UCLIBC_HAS_FNMATCH_OLD=y UCLIBC_HAS_FTW=y UCLIBC_HAS_GLOB=y UCLIBC_HAS_GNU_GLOB=y SHARED_LIB_LOADER_PREFIX="/lib" RUNTIME_PREFIX="/" DEVEL_PREFIX="/usr/" UCLIBC_BUILD_RELRO=y UCLIBC_BUILD_NOW=y CROSS_COMPILER_PREFIX="/home/be/Buildroot/buildroot_EABI_3.4.6_bug/buildroot/build_arm_nofpu/staging_dir/bin/arm-linux-uclibcgnueabi-" DOSTRIP=y WARNINGS="-Wall" ----------------------------------------------------------------------------------------- ---------------------------------------------------------------------- bernhardf - 02-04-07 09:09 ---------------------------------------------------------------------- I think this is fixed in current trunk. Please retest. Issue History Date Modified Username Field Change ====================================================================== 12-13-06 01:58 mikewhit New Issue 12-13-06 01:58 mikewhit Status new => assigned 12-13-06 01:58 mikewhit Assigned To => uClibc 12-13-06 02:43 mikewhit Note Added: 0001860 12-21-06 03:50 bernhardf Note Added: 0001900 01-09-07 05:06 stephaneC Note Added: 0001974 01-09-07 05:07 stephaneC Issue Monitored: stephaneC 02-04-07 09:09 bernhardf Note Added: 0002106 02-04-07 09:09 bernhardf Status assigned => feedback ====================================================================== From bugs at busybox.net Sun Feb 4 09:09:53 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:09:53 -0800 Subject: [buildroot 0001105]: Missing mkfs.jffs2 option for NAND Flash Message-ID: The following issue has been CLOSED ====================================================================== http://busybox.net/bugs/view.php?id=1105 ====================================================================== Reported By: artemys Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1105 Category: Architecture Specific Reproducibility: always Severity: trivial Priority: normal Status: closed Resolution: open Fixed in Version: ====================================================================== Date Submitted: 11-29-2006 04:12 PST Last Modified: 02-04-2007 09:09 PST ====================================================================== Summary: Missing mkfs.jffs2 option for NAND Flash Description: When generating a JFFS2 rootfs for a NAND flash, an option should be added to mkfs.jffs2 otherwise Linux kernel is printing a lot of warning messages (during rootfs mounting) like: CLEANMARKER node found at 0x00...... has totlen 0xc != normal 0x0 These messages are considerably slowing down system boot. "-n" option should be added to JFFS2_OPTS variable in target/jffs2/jffs2root.mk Makefile. ====================================================================== ---------------------------------------------------------------------- bernhardf - 02-04-07 09:09 ---------------------------------------------------------------------- Fixed in svn a week or two ago. Thanks for the report! Issue History Date Modified Username Field Change ====================================================================== 11-29-06 04:12 artemys New Issue 11-29-06 04:12 artemys Status new => assigned 11-29-06 04:12 artemys Assigned To => uClibc 11-29-06 09:13 artemys File Added: 003-buildroot-cleanmarkers.diff 02-04-07 09:09 bernhardf Status assigned => closed 02-04-07 09:09 bernhardf Note Added: 0002107 ====================================================================== From bugs at busybox.net Sun Feb 4 09:10:24 2007 From: bugs at busybox.net (bugs at busybox.net) Date: Sun, 4 Feb 2007 09:10:24 -0800 Subject: [buildroot 0001052]: MMU: Build fails for MIPS-1 target, vfork.c, `fork' undeclared here Message-ID: <83645a1db97373c4ff02e83bc59becf2@bugs.busybox.net> The following issue has been UPDATED. ====================================================================== http://busybox.net/bugs/view.php?id=1052 ====================================================================== Reported By: dash77 Assigned To: uClibc ====================================================================== Project: buildroot Issue ID: 1052 Category: Architecture Specific Reproducibility: always Severity: block Priority: