devel

devel@uclibc-ng.org

4 participants
1145 discussions

[PATCH v2] m68k: fix SUPPORT_LD_DEBUG null pointer access

by Peter Seiderer

Fix null pointer access in case SUPPORT_LD_DEBUG configuration option is enabled (hit in _dl_do_reloc, but fix same old_val/reloc_addr usage pattern in _dl_do_lazy_reloc too). Reported-by: Jean-Michel Hautbois <jeanmichel.hautbois(a)yoseli.org> Link: https://mailman.openadk.org/mailman3/hyperkitty/list/devel@uclibc-ng.org/th… Signed-off-by: Peter Seiderer <ps.report(a)gmx.net> --- Changes v1 -> v2: - fix same old_val/reloc_addr usage pattern in _dl_do_lazy_reloc too - reword commit message --- ldso/ldso/m68k/elfinterp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/ldso/ldso/m68k/elfinterp.c b/ldso/ldso/m68k/elfinterp.c index 25ea23067..e7fa117da 100644 --- a/ldso/ldso/m68k/elfinterp.c +++ b/ldso/ldso/m68k/elfinterp.c @@ -193,7 +193,7 @@ _dl_do_reloc(struct elf_resolve *tpnt, struct r_scope_elem *scope, #endif #if defined (__SUPPORT_LD_DEBUG__) - old_val = *reloc_addr; + old_val = reloc_addr ? *reloc_addr : 0; #endif switch (reloc_type) { @@ -292,7 +292,7 @@ _dl_do_lazy_reloc(struct elf_resolve *tpnt, struct r_scope_elem *scope, reloc_type = ELF_R_TYPE(rpnt->r_info); #if defined (__SUPPORT_LD_DEBUG__) - old_val = *reloc_addr; + old_val = reloc_addr ? *reloc_addr : 0; #endif switch (reloc_type) { -- 2.48.1

1 day, 10 hours

[PATCH v1] m68k: fix SUPPORT_LD_DEBUG null pointer access

by Peter Seiderer

Fix null pointer access in case SUPPORT_LD_DEBUG configuration option is enabled. Reported-by: Jean-Michel Hautbois <jeanmichel.hautbois(a)yoseli.org> Link: https://mailman.openadk.org/mailman3/hyperkitty/list/devel@uclibc-ng.org/th… Signed-off-by: Peter Seiderer <ps.report(a)gmx.net> --- ldso/ldso/m68k/elfinterp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/ldso/ldso/m68k/elfinterp.c b/ldso/ldso/m68k/elfinterp.c index 25ea23067..1536bd5b2 100644 --- a/ldso/ldso/m68k/elfinterp.c +++ b/ldso/ldso/m68k/elfinterp.c @@ -193,7 +193,7 @@ _dl_do_reloc(struct elf_resolve *tpnt, struct r_scope_elem *scope, #endif #if defined (__SUPPORT_LD_DEBUG__) - old_val = *reloc_addr; + old_val = reloc_addr ? *reloc_addr : 0; #endif switch (reloc_type) { -- 2.48.1

1 day, 10 hours

Unable to use LD_DEBUG on a coldfire

by Jean-Michel Hautbois

Hello! I have a custom buildroot image, for a M54418 coldfire. It works fine with uclibc-ng (1.0.50). I wanted to enable the LD_DEBUG support, with SUPPORT_LD_DEBUG but the init fails. I am using init=/bin/bash. When I do not enable the configuration, I have: [ 5.800748] Run /bin/bash as init process [ 5.800807] with arguments: [ 5.800865] /bin/bash [ 5.800918] with environment: [ 5.800964] HOME=/ [ 5.801011] TERM=linux bash: cannot set terminal process group (-1): Inappropriate ioctl for device bash: no job control in this shell bash-5.2# I added SUPPORT_LD_DEBUG_EARLY and here is the output: [ 5.820897] Run /bin/bash as init process [ 5.820962] with arguments: [ 5.835089] /bin/bash [ 5.835177] with environment: [ 5.835231] HOME=/ [ 5.835282] TERM=linux ELF header=0x60000000 First Dynamic section entry=0x60009be0 Scanning DYNAMIC section Done scanning DYNAMIC section About to do library loader relocations Done relocating ldso; we can now use globals and make function calls! _dl_get_ready_to_run:473: Cool, ldso survived making function calls _dl_malloc:243: mmapping more memory _dl_ldsopath_init:157: Lib Loader: (0x60000000) /lib/ld-uClibc.so.0: using path: /lib _dl_get_ready_to_run:1079: Loading: (0x60010000) /usr/lib/libreadline.so.8 _dl_get_ready_to_run:1079: Loading: (0x60052000) /usr/lib/libhistory.so.8 _dl_get_ready_to_run:1079: Loading: (0x6005e000) /usr/lib/libncurses.so.6 _dl_load_elf_shared_library:843: Found TLS header for /lib/libc.so.0 _dl_load_elf_shared_library:870: Relocated TLS initial image from 0x84bb8 to 0x60128bb8 (size = 0x4) _dl_get_ready_to_run:1079: Loading: (0x600a4000) /lib/libc.so.0 _dl_get_ready_to_run:1079: Loading: (0x6005e000) /usr/lib/libncurses.so.6 _dl_get_ready_to_run:1079: Loading: (0x600a4000) /lib/libc.so.0 _dl_get_ready_to_run:1079: Loading: (0x6005e000) /usr/lib/libncurses.so.6 _dl_get_ready_to_run:1079: Loading: (0x600a4000) /lib/libc.so.0 _dl_get_ready_to_run:1079: Loading: (0x600a4000) /lib/libc.so.0 _dl_get_ready_to_run:1079: Loading: (0x60000000) /lib/ld-uClibc.so.0 _dl_get_ready_to_run:1220: Calling init_tls()! _dl_get_ready_to_run:1320: Beginning relocation fixups [ 7.156188] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b [ 7.156278] CPU: 0 UID: 0 PID: 1 Comm: bash Not tainted 6.14.0-rc1-00083-g859f1346c094 #2264 [ 7.156388] Stack from 41849e20: [ 7.156388] 41849e20 415715bd 415715bd 41450b00 00000001 0000000b 41849e44 4100c4be [ 7.156388] 415715bd 41849e64 410034ec 41450bf8 418780ff 0000000b 418780c8 41874000 [ 7.156388] 00000000 41849ea0 410168a0 41545f7b 0000000b 0000000b 41849f80 00000001 [ 7.156388] 418780c8 400004d8 4184e000 41878000 4184e000 00000000 00000000 00000000 [ 7.156388] 41849eb8 41016f88 0000000b 0000000b 41849f80 41849f6c 41849ef0 4101febe [ 7.156388] 0000000b 41849f6c 00000000 8000e4a4 8000e4a4 6000c028 00000001 41849fcc [ 7.157964] Call Trace: [<41450b00>] rtlock_slowlock_locked+0x79c/0x894 [ 7.158153] [<4100c4be>] dump_stack+0x10/0x16 [ 7.158305] [<410034ec>] panic+0xda/0x28e [ 7.158415] [<41450bf8>] rt_spin_lock+0x0/0x7c [ 7.158542] [<410168a0>] do_exit+0x18a/0x71e [ 7.158708] [<41016f88>] do_group_exit+0x26/0x78 [ 7.158844] [<4101febe>] get_signal+0x500/0x602 [ 7.158963] [<4100f150>] test_ti_thread_flag+0x0/0x1a [ 7.159116] [<4100f810>] do_notify_resume+0x3a/0x4da [ 7.159237] [<4101f1ea>] force_sig_fault+0x18/0x1c [ 7.159350] [<4101072e>] send_fault_sig+0x56/0xba [ 7.159471] [<410109a0>] do_page_fault+0x20e/0x2b6 [ 7.159596] [<4101126a>] Lsignal_return+0x14/0x26 [ 7.159726] [ 7.293555] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]--- Any idea would be appreciated :-). Thanks ! JM

1 day, 12 hours

[PATCH] inet/resolv.c: set h_errnop if /etc/hosts cannot be opened

by Marcus Haehnel

From: Jan Klötzke <jan.kloetzke(a)kernkonzept.com> If gethostent_r fails, the h_errnop error code should be set. Signed-off-by: Marcus Haehnel <marcus.haehnel(a)kernkonzept.com> --- libc/inet/resolv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libc/inet/resolv.c b/libc/inet/resolv.c index d7a659a8c..130a9a505 100644 --- a/libc/inet/resolv.c +++ b/libc/inet/resolv.c @@ -2699,7 +2699,7 @@ int gethostent_r(struct hostent *result_buf, char *buf, size_t buflen, hostp = __open_etc_hosts(); if (hostp == NULL) { *result = NULL; - ret = TRY_AGAIN; + *h_errnop = ret = TRY_AGAIN; goto DONE; } } -- 2.47.1

1 week, 2 days

[PATCH] nan.h: define NAN only if not already defined

by Marcus Haehnel

From: Frank Mehnert <frank.mehnert(a)kernkonzept.com> Clang-19 defines NAN in <float.h>. Check if NAN is already defined to prevent a compiler warning. Change-Id: I29ebcb2a7a0c1eade46a73074a1701ea67869128 Signed-off-by: Marcus Haehnel <marcus.haehnel(a)kernkonzept.com> --- libc/sysdeps/linux/common/bits/nan.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/libc/sysdeps/linux/common/bits/nan.h b/libc/sysdeps/linux/common/bits/nan.h index 00cb405f1..46cfb613f 100644 --- a/libc/sysdeps/linux/common/bits/nan.h +++ b/libc/sysdeps/linux/common/bits/nan.h @@ -25,7 +25,9 @@ #if __GNUC_PREREQ(3,3) -# define NAN (__builtin_nanf ("")) +# ifndef NAN +# define NAN (__builtin_nanf ("")) +# endif #elif defined __GNUC__ -- 2.47.1

1 week, 4 days

LTP test fstatat01.c FTBFS on Buildroot bootlin-sh4-uclibc toolchain

by Petr Vorel

Hi all, LTP test fstatat01.c [1] FTBFS on Buildroot bootlin-sh4-uclibc toolchain (uclibc: 1.0.50, kernel headers: 4.19.255). It looks to me that other bootlin uclibc based toolchains are working, therefore it's probably sh4 problem. Looking into include/sys/stat.h in uclibc-ng [2], extern int fstatat (int __fd, ...) definition is guarded only by: # ifdef __REDIRECT_NTH. But in glibc io/sys/stat.h [3] guardes it with # ifdef __USE_TIME64_REDIRECTS added glibc 2.40 a4ed0471d7 ("Always define __USE_TIME_BITS64 when 64 bit time_t is used") which replaced previous __USE_TIME_BITS64 added in in glibc 2.34 47f24c21ee [4] ("y2038: Add support for 64-bit time on legacy ABIs") [5]. Should similar change be added to uclibc-ng? Kind regards, Petr $ cd ~/br-test-pkg/bootlin-sh4-uclibc/build/ltp-testsuite-728759506cbe08612183275b3543007d1c47f7f4/testcases/kernel/syscalls/fstatat $ make V=1 fstatat01 br-test-pkg/bootlin-sh4-uclibc/host/bin/sh4-linux-gcc -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_FORTIFY_SOURCE=2 -I../../../../include -I../../../../include -I../../../../include/old/ -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -g0 -Ibr-test-pkg/bootlin-sh4-uclibc/host/bin/../sh4-buildroot-linux-uclibc/sysroot/usr/include/tirpc -g -O2 -fno-strict-aliasing -pipe -Wall -W -Wold-style-definition -std=gnu99 -L../../../../lib fstatat01.c -Lbr-test-pkg/bootlin-sh4-uclibc/host/bin/../sh4-buildroot-linux-uclibc/sysroot/usr/lib -ltirpc -lfts -lltp -o fstatat01 fstatat01.c:63:5: error: conflicting types for ‘fstatat’; have ‘int(int, const char *, struct stat64 *, int)’ 63 | int fstatat(int dirfd, const char *filename, struct stat64 *statbuf, int flags) | ^~~~~~~ In file included from br-test-pkg/bootlin-sh4-uclibc/host/sh4-buildroot-linux-uclibc/sysroot/usr/include/features.h:409, from br-test-pkg/bootlin-sh4-uclibc/host/sh4-buildroot-linux-uclibc/sysroot/usr/include/sys/types.h:26, from fstatat01.c:27: br-test-pkg/bootlin-sh4-uclibc/host/sh4-buildroot-linux-uclibc/sysroot/usr/include/sys/stat.h:243:12: note: previous declaration of ‘fstatat’ with type ‘int(int, const char * restrict, struct stat * restrict, int)’ 243 | extern int __REDIRECT_NTH (fstatat, (int __fd, const char *__restrict __file, | ^~~~~~~~~~~~~~ [1] https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/sysc… [2] https://cgit.openadk.org/cgi/cgit/uclibc-ng.git/tree/include/sys/stat.h#n247 [3] https://sourceware.org/git/?p=glibc.git;a=blob;f=io/sys/stat.h;h=4bea9e9a77… [4] https://sourceware.org/git/?p=glibc.git;a=commit;h=a4ed0471d71739928a0d0fa3… [5] https://sourceware.org/git/?p=glibc.git;a=commit;h=47f24c21ee38701ae275aa9e…

3 weeks, 1 day

[PATCH] libc/string/x86_64: fix p2align max number of bytes to skip

by Marcus Haehnel

From: Frank Mehnert <frank.mehnert(a)kernkonzept.com> In a .p2align directive, the maximum number of bytes to skip shall be at most the intended alignment minus 1. Certain assemblers warn otherwise: spn.S:97:14: warning: maximum bytes expression exceeds alignment and has no effect .p2align 3,,8 ^ Change-Id: I5ad48f6f3c3036d8d33d3710ed8d5b5a35bf1fea Signed-off-by: Marcus Haehnel <marcus.haehnel(a)kernkonzept.com> --- libc/string/x86_64/strcat.S | 2 +- libc/string/x86_64/strcspn.S | 2 +- libc/string/x86_64/strlen.S | 2 +- libc/string/x86_64/strspn.S | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/libc/string/x86_64/strcat.S b/libc/string/x86_64/strcat.S index 55e09e5f1..209e19062 100644 --- a/libc/string/x86_64/strcat.S +++ b/libc/string/x86_64/strcat.S @@ -106,7 +106,7 @@ ENTRY (BP_SYM (strcat)) /* Align, it is a jump target. */ /* Next 3 insns are 8 bytes total, make sure we decode them in one go */ - .p2align 3,,8 + .p2align 3,,7 3: subq $8,%rax /* correct pointer increment. */ diff --git a/libc/string/x86_64/strcspn.S b/libc/string/x86_64/strcspn.S index 7a06c8867..5ef565db7 100644 --- a/libc/string/x86_64/strcspn.S +++ b/libc/string/x86_64/strcspn.S @@ -94,7 +94,7 @@ L(1): leaq -4(%rdx), %rax /* prepare loop */ /* but it will also align entire function to 16 bytes, */ /* potentially creating largish padding at link time. */ /* We are aligning to 8 bytes instead: */ - .p2align 3,,8 + .p2align 3,,7 L(3): addq $4, %rax /* adjust pointer for full loop round */ diff --git a/libc/string/x86_64/strlen.S b/libc/string/x86_64/strlen.S index 9e84326c2..2fe2f58b2 100644 --- a/libc/string/x86_64/strlen.S +++ b/libc/string/x86_64/strlen.S @@ -102,7 +102,7 @@ ENTRY (strlen) /* Align, it is a jump target. */ /* Next 3 insns are 8 bytes total, make sure we decode them in one go */ - .p2align 3,,8 + .p2align 3,,7 3: subq $8,%rax /* correct pointer increment. */ diff --git a/libc/string/x86_64/strspn.S b/libc/string/x86_64/strspn.S index 366377649..8dc42656b 100644 --- a/libc/string/x86_64/strspn.S +++ b/libc/string/x86_64/strspn.S @@ -89,7 +89,7 @@ L(1): leaq -4(%rdx), %rax /* prepare loop */ /* but it will also align entire function to 16 bytes, */ /* potentially creating largish padding at link time. */ /* We are aligning to 8 bytes instead: */ - .p2align 3,,8 + .p2align 3,,7 L(3): addq $4, %rax /* adjust pointer for full loop round */ -- 2.47.1

3 weeks, 3 days

error: '__NR__dl_fstat' undeclared, uClibc-ng 1.0.50, headers 4.4

by Thomas Petazzoni

Hello, uClibc-ng 1.0.50 + Linux headers 4.4 fails to build with: In file included from ./include/sys/syscall.h:33, from ./ldso/include/dl-syscall.h:12, from ./ldso/include/ldso.h:52, from ldso/ldso/ldso.c:32: ./ldso/include/dl-syscall.h: In function '_dl_fstat': ./include/bits/syscalls-common.h:34:63: error: '__NR__dl_fstat' undeclared (first use in this function); did you mean '_dl_fstat'? # define INLINE_SYSCALL(name, nr, args...) INLINE_SYSCALL_NCS(__NR_##name, nr, args) This is reproducible with Buildroot's arcturus_ucp1020_defconfig, and a log is visible at: https://gitlab.com/buildroot.org/buildroot/-/jobs/8864051076 With the full log at: https://buildroot.org.gitlab.io/-/buildroot/-/jobs/8864051076/artifacts/bui… Any clue? Thomas -- Thomas Petazzoni, co-owner and CEO, Bootlin Embedded Linux and Kernel engineering and training https://bootlin.com

1 month

[PATCH] uclibc: prevent warning about setting an unused variable

by Marcus Haehnel

From: Frank Mehnert <frank.mehnert(a)kernkonzept.com> This variable is used either if __UCLIBC_HAS_FLOATS__ is defined or if __UCLIBC_HAS_GLIBC_DIGIT_GROUPING__ is defined. Instead of adding more complicated #ifdef magic, just mark this variable as possibly unused to satisfy the compiler. Signed-off-by: marcus.haehnel(a)kernkonzept.com --- libc/stdio/_scanf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libc/stdio/_scanf.c b/libc/stdio/_scanf.c index 3f3000d6f..cb72d14ac 100644 --- a/libc/stdio/_scanf.c +++ b/libc/stdio/_scanf.c @@ -1715,7 +1715,7 @@ int attribute_hidden __psfs_do_numeric(psfs_t *psfs, struct scan_cookie *sc) #define MAX_DIGITS 65 /* Allow one leading 0. */ unsigned char buf[MAX_DIGITS+2+ 100]; unsigned char usflag, base; - unsigned char nonzero = 0; + unsigned char nonzero __attribute__((unused)) = 0; unsigned char seendigit = 0; #ifndef __UCLIBC_HAS_FLOATS__ -- 2.47.1

1 month

1.0.51 released

by Waldemar Brodkorb

Hi uClibc-ng hackers, I released 1.0.51. See here for changes: Dmitry Chestnykh (2): mips64n32, time64: Select correct _dl_fstat impl Revert "mips64: time64 for n32 ABI breaks a lot of tests, disable it for now" Frank Mehnert (4): uclibc: prevent warning about setting an unused variable uclibc: prevent warning about declaring an unused function prevent warning about setting an unused variable fix possible overflow in pointer arithmetics strnlen() Mingxiang Lu (1): fix TLS memory leak with dlopen Nadav Tasher (3): features.h: error out when attempting to compile with TIME64 for 32bit targets on Linux versions prior to 5.1.0 Revert "features.h: error out when attempting to compile with TIME64 for 32bit targets on Linux versions prior to 5.1.0" time64_helpers.h: error out when attempting to compile with TIME64 for 32bit targets on Linux versions prior to 5.1.0 Waldemar Brodkorb (4): add process_vm_{read,write}v syscall wrapper from glibc mips64: time64 for n32 ABI breaks a lot of tests, disable it for now clock_adjtime: warn instead of error out bump version for 1.0.51 release best regards Waldemar

1 month

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

devel