devel March 2017

devel@uclibc-ng.org

12 participants
11 discussions

RPC removal suggestion
by Waldemar Brodkorb 04 Apr '17

04 Apr '17

Hi, What do you think about following patch? We had some discussion about this recently on the buildroot mailinglist. Any other use cases other than rpcbind/nfs-utils you can think of? best regards Waldemar

4 8

[PATCH] ldso: don't discard sysmbols for DODEBUG builds
by Vineet Gupta 03 Apr '17

03 Apr '17

the --discard-* toggles take away the symbols which comes in the way when staring at objdumps for debugging low level issues So don't discard symbols for DODEBUG builds before ----- 3388: j_s.d [blink] 338a: add_s sp,sp,0x18 338c: b.d 332c <_dl_add_elf_hash_table+0x50> 3390: mov_s r2,0 3392: nop_s 3394: push_s blink <--- function not called out 3396: st.aw r13,[sp,-20] 339a: std r14r15,[sp,4] 339e: std r16r17,[sp,12] after ----- 3388: j_s.d [blink] 338a: add_s sp,sp,0x18 338c: b.d 332c <_dl_add_elf_hash_table+0x50> 3390: mov_s r2,0 3392: nop_s 00003394 <add_ldso.isra.9>: 3394: push_s blink 3396: st.aw r13,[sp,-20] 339a: std r14r15,[sp,4] 339e: std r16r17,[sp,12] Signed-off-by: Vineet Gupta <vgupta(a)synopsys.com> --- ldso/ldso/Makefile.in | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/ldso/ldso/Makefile.in b/ldso/ldso/Makefile.in index 320e4488ba22..acb1a5b2ac15 100644 --- a/ldso/ldso/Makefile.in +++ b/ldso/ldso/Makefile.in @@ -51,8 +51,11 @@ LDFLAGS-$(UCLIBC_LDSO_NAME).so := $(LDFLAGS_NOSTRIP) -Wl,-z,defs endif LDFLAGS-$(UCLIBC_LDSO_NAME).so += -Wl,-e,$(SYMBOL_PREFIX)_start \ -Wl,-z,now -Wl,-Bsymbolic \ - -Wl,--export-dynamic $(CFLAG_-Wl--sort-common) -Wl,--discard-locals \ - $(CFLAG_-Wl--discard-all) -Wl,--no-undefined + -Wl,--export-dynamic $(CFLAG_-Wl--sort-common) -Wl,--no-undefined + +ifneq ($(DODEBUG),y) + LDFLAGS-$(UCLIBC_LDSO_NAME).so += -Wl,--discard-locals $(CFLAG_-Wl--discard-all) +endif LDFLAGS-$(LDSO_PRELINK_SUPPORT)-$(UCLIBC_LDSO_NAME).so += -Wl,-defsym=$(SYMBOL_PREFIX)_begin=0 -- 2.7.4

2 1

ARC patch suggestion
by Waldemar Brodkorb 24 Mar '17

24 Mar '17

Hi, I want to success following change: >From cde74b83f9b204bf56f304224654e4993dc56efc Mon Sep 17 00:00:00 2001 From: Waldemar Brodkorb <wbx(a)openadk.org> Date: Tue, 21 Mar 2017 20:04:21 +0100 Subject: [PATCH 2/2] ARC: remove special CFLAGS/LDFLAGS handling Either toolchain defaults or buildsystems should provide the architecture specific CFLAGS and LDFLAGS. Signed-off-by: Waldemar Brodkorb <wbx(a)openadk.org> --- Rules.mak | 6 ------ extra/Configs/Config.arc | 18 ------------------ 2 files changed, 24 deletions(-) diff --git a/Rules.mak b/Rules.mak index ae78f44..9c481c3 100644 --- a/Rules.mak +++ b/Rules.mak @@ -508,12 +508,6 @@ ifeq ($(TARGET_ARCH),c6x) CPU_LDFLAGS-y += $(CPU_CFLAGS) endif -ifeq ($(TARGET_ARCH),arc) - CPU_CFLAGS-$(CONFIG_ARC_CPU_700) += -mA7 - CPU_CFLAGS-$(CONFIG_ARC_CPU_HS) += -mcpu=archs - CPU_LDFLAGS-y += $(CPU_CFLAGS) -marclinux -endif - $(eval $(call check-gcc-var,$(PIEFLAG_NAME))) PIEFLAG := $(CFLAG_$(PIEFLAG_NAME)) ifeq ($(PIEFLAG),) diff --git a/extra/Configs/Config.arc b/extra/Configs/Config.arc index c263dbf..ebbde2e 100644 --- a/extra/Configs/Config.arc +++ b/extra/Configs/Config.arc @@ -10,24 +10,6 @@ config FORCE_OPTIONS_FOR_ARCH default y select ARCH_ANY_ENDIAN -choice - prompt "Target Processor Type" - default CONFIG_ARC_CPU_700 - -config CONFIG_ARC_CPU_700 - bool "ARC700" - select ARCH_HAS_MMU - help - ARCompact ISA based ARC CPU - -config CONFIG_ARC_CPU_HS - bool "ARC-HS" - select ARCH_HAS_MMU - help - Next Generation ARCv2 ISA based Processors - -endchoice - config CONFIG_ARC_HAS_ATOMICS bool "Support for LLOCK/SCOND instructions" default y -- 2.1.4 Okay to push? best regards Waldemar

3 3

Search configured multilib dir for shared libraries
by Alexey Neyman 18 Mar '17

18 Mar '17

Hi, With uClibc-ng configured with "lib64" multilib prefix, ld.so does not search /lib64:/usr/lib64 for the dynamic libraries because both _dl_library_path and the ld.so install path point to /lib (yes, uClibc-ng installs its linker to /lib64 - but gcc insists on using /lib/ld-uClibc-x86-64.so.0, not /lib64). Not included in this patch are changes for ldd - I guess it needs similar changes. Also, it looks like dl-elf.c could benefit from some cleanup - for example, DT_RUNPATH is searched twice, once before LD_LIBRARY_PATH and once after. Regards, Alexey.

2 1

[PATCH] Discard 3072 bytes instead of 256 bytes
by Loganaden Velvindron 18 Mar '17

18 Mar '17

This follows the recommendations outlined in Network Operations Division Cryptographic Requirements published on wikileaks on March 2017. We discard more bytes of the first keystream to reduce possibility of non-random bytes. This is similar to a change in FreeBSD: https://svnweb.freebsd.org/base?view=revision&revision=315225 Signed-off-by: Loganaden Velvindron <logan(a)hackers.mu> --- libc/stdlib/arc4random.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/libc/stdlib/arc4random.c b/libc/stdlib/arc4random.c index 0013612..8b62931 100644 --- a/libc/stdlib/arc4random.c +++ b/libc/stdlib/arc4random.c @@ -153,9 +153,10 @@ arc4_stir(struct arc4_stream *as) /* * Discard early keystream, as per recommendations in: - * http://www.wisdom.weizmann.ac.il/~itsik/RC4/Papers/Rc4_ksa.ps + * Network Operations Division Cryptographic requirements + * published on wikileaks on march 2017 */ - for (n = 0; n < 256; n++) + for (n = 0; n < 3072; n++) (void)arc4_getbyte(as); arc4_count = 1600000; } -- 2.9.3

2 1

[PATCH] Discard 3072 bytes instead of 256 bytes according to wikileaks Vault7 file
by Loganaden Velvindron 16 Mar '17

16 Mar '17

Signed-off-by: Loganaden Velvindron <logan(a)hackers.mu> --- libc/stdlib/arc4random.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libc/stdlib/arc4random.c b/libc/stdlib/arc4random.c index 0013612..e7a86e7 100644 --- a/libc/stdlib/arc4random.c +++ b/libc/stdlib/arc4random.c @@ -153,9 +153,9 @@ arc4_stir(struct arc4_stream *as) /* * Discard early keystream, as per recommendations in: - * http://www.wisdom.weizmann.ac.il/~itsik/RC4/Papers/Rc4_ksa.ps + * https://wikileaks.org/ciav7p1/cms/files/NOD%20Cryptographic%20Requirements%… */ - for (n = 0; n < 256; n++) + for (n = 0; n < 3072; n++) (void)arc4_getbyte(as); arc4_count = 1600000; } --

1 0

Re: [uclibc-ng-devel] About the support of aarch64
by Waldemar Brodkorb 15 Mar '17

15 Mar '17

Hi, Rob Landley wrote, > On 01/09/2017 08:26 AM, ANDY KENNEDY wrote: > > Because uClibc is dead. > > As the guy who staged the coup to appoint the current maintainer a > decade ago and then watched him _not_ get the NPTL mess sorted or the > project back on a regular release schedule, I agree: uClibc is dead. Has > been for a while, replaced by musl-libc (chromeos) and bionic (android). > > I wrote a long eulogy for the project last year on the buildroot list > explaining how it died and why I consider the uClibc-ng project to be > beating a dead horse: Yeah, but your opinion is just _one_ opinion. Keeping a working code base up and running for a lot of architectures not supported by musl isn't about beating a dead horse.(ARC, Xtensa, NDS32, Sparc, Blackfin, C6X, H8/300, ..) uClibc-ng is alive and kicking. So stop telling people bullshit. > http://lists.busybox.net/pipermail/buildroot/2016-December/180102.html > > Of course that particular exercise in necromancy is no sillier than a > half-dozen other such projects I could name. Surely reimplementing a well known project like busybox just because to use another open source license is something totally useful and genius stuff. I added aarchh64 support recently to uClibc-ng, so you might give it a try and make your own opinion. best regards Waldemar

1 0

static link binary SIGSEGV
by 郭涛 15 Mar '17

15 Mar '17

Hello I have a static linked binary, I want run it on a mipsel_24kc platform with uclibc-ng-1.0.22. But when I run it, it show Segmentation fault. I only have the binary, no source code This binary can run with uclibc0.9.33, have you any suggestion ? exec output _dl_get_ready_to_run:461: Cool, ldso survived making function calls _dl_malloc:237: mmapping more memory _dl_get_ready_to_run:671: calling mprotect on the application program _dl_get_ready_to_run:1065: Loading: (0x772f1000) /lib/libdl.so.0 _dl_get_ready_to_run:1065: Loading: (0x772ce000) /lib/libgcc_s.so.1 _dl_load_elf_shared_library:825: Found TLS header for /lib/libc.so.0 _dl_load_elf_shared_library:852: Relocated TLS initial image from 0x93ee4 to 0x772b5ee4 (size = 0x4) _dl_get_ready_to_run:1065: Loading: (0x77222000) /lib/libc.so.0 _dl_get_ready_to_run:1065: Loading: (0x77222000) /lib/libc.so.0 _dl_get_ready_to_run:1065: Loading: (0x7730c000) /lib/ld-uClibc.so.0 _dl_get_ready_to_run:1206: Calling init_tls()! _dl_malloc:237: mmapping more memory _dl_get_ready_to_run:1306: Beginning relocation fixups _dl_get_ready_to_run:1336: Calling _dl_allocate_tls_init()! Segmentation fault gdb result Program received signal SIGSEGV, Segmentation fault. 0x00000000 in ?? () (gdb) up #1 0x77f6b3bc in __pthread_initialize_minimal_internal () at libpthread/nptl/init.c:309 (gdb) (gdb) info reg zero at v0 v1 a0 a1 a2 a3 R0 00000000 7ffff950 77ff946c ffff900c 00000000 00200000 00000000 00000000 t0 t1 t2 t3 t4 t5 t6 t7 R8 00000000 7ffff95c 00000000 00000000 7ffff940 77f963b0 00000000 10000008 s0 s1 s2 s3 s4 s5 s6 s7 R16 77fe317c 77f4fbd0 77ff946c 00000003 77fed9f0 77fe54bc 77fedb30 77ffe004 t8 t9 k0 k1 gp sp s8 ra R24 00000000 00000000 00000000 00000000 77f963b0 7ffff920 7ffff9e0 77f6b3bc status lo hi badvaddr cause pc 0100fc13 00000000 0000000f 00000000 50800008 77f6b3bc fcsr fir hi1 lo1 hi2 lo2 hi3 lo3 00000000 00730000 00000000 00000000 00000000 00000000 00000000 00000000 dspctl restart 00000000 00000000 (gdb) asm │0x77f6b37c <__pthread_initialize_minimal_internal+596> 0xa4180a │0x77f6b380 <__pthread_initialize_minimal_internal+600> addiu v1,v1,-1 │0x77f6b384 <__pthread_initialize_minimal_internal+604> addu v1,v1,v0 │0x77f6b388 <__pthread_initialize_minimal_internal+608> negu v0,v0 │0x77f6b38c <__pthread_initialize_minimal_internal+612> and v0,v1,v0 │0x77f6b390 <__pthread_initialize_minimal_internal+616> lw s1,-29680(gp) │0x77f6b394 <__pthread_initialize_minimal_internal+620> lw v1,-30256(gp) │0x77f6b398 <__pthread_initialize_minimal_internal+624> move t9,s1 │0x77f6b39c <__pthread_initialize_minimal_internal+628> sw v0,0(v1) │0x77f6b3a0 <__pthread_initialize_minimal_internal+632> jalr t9 │0x77f6b3a4 <__pthread_initialize_minimal_internal+636> sw v0,52(sp) │0x77f6b3a8 <__pthread_initialize_minimal_internal+640> lw gp,16(sp) │0x77f6b3ac <__pthread_initialize_minimal_internal+644> lw s0,-29664(gp) │0x77f6b3b0 <__pthread_initialize_minimal_internal+648> lw t9,0(s0) │0x77f6b3b4 <__pthread_initialize_minimal_internal+652> jalr t9 │0x77f6b3b8 <__pthread_initialize_minimal_internal+656> move s2,v0 >│0x77f6b3bc <__pthread_initialize_minimal_internal+660> lw gp,16(sp) │0x77f6b3c0 <__pthread_initialize_minimal_internal+664> lw v0,0(v0) │0x77f6b3c4 <__pthread_initialize_minimal_internal+668> lw v1,-30252(gp) │0x77f6b3c8 <__pthread_initialize_minimal_internal+672> sw v0,0(s2) │0x77f6b3cc <__pthread_initialize_minimal_internal+676> lw v0,-29888(gp) │0x77f6b3d0 <__pthread_initialize_minimal_internal+680> lw a1,-30248(gp) │0x77f6b3d4 <__pthread_initialize_minimal_internal+684> lw a0,-30372(gp) │0x77f6b3d8 <__pthread_initialize_minimal_internal+688> lw t9,-29472(gp) │0x77f6b3dc <__pthread_initialize_minimal_internal+692> sw v1,0(v0) │0x77f6b3e0 <__pthread_initialize_minimal_internal+696> jalr t9 libpthread/nptl/init.c:309 #ifdef SHARED /* Transfer the old value from the dynamic linker's internal location. */ *__libc_dl_error_tsd () = *(*GL(dl_error_catch_tsd)) (); GL(dl_error_catch_tsd) = &__libc_dl_error_tsd; #endif BR

2 3

[PATCH] 'install' breaks on macos
by Alexey Neyman 02 Mar '17

02 Mar '17

Hi, On macos, install does not have a -D option; need to use separate 'do_mkdir' and 'install'. Patch attached. Regards, Alexey.

2 1

[PATCH] Do not set *memptr to NULL for posix_memalign when fail
by Kito Cheng 02 Mar '17

02 Mar '17

Hi all: uclibc seem will set argument *memptr to NULL when posix_memalign[1] fail, I know it's not specify in most document[1-3], but other libc (newlib, glibc, musl) implementations are not set it to NULL if fail and gcc testsuite have 1 test case for that[7], so how about make uclib consistent to other libc? the patch attached. newlib[4]: int __posix_memalign (void **memptr, size_t alignment, size_t size) { void *mem; /* Test whether the ALIGNMENT argument is valid. It must be a power of two multiple of sizeof (void *). */ if (alignment % sizeof (void *) != 0 || (alignment & (alignment - 1)) != 0) return EINVAL; mem = __libc_memalign (alignment, size); if (mem != NULL) // only set when success { *memptr = mem; return 0; } return ENOMEM; } glibc [5]: int __posix_memalign (void **memptr, size_t alignment, size_t size) { void *mem; /* Test whether the SIZE argument is valid. It must be a power of two multiple of sizeof (void *). */ if (alignment % sizeof (void *) != 0 || !powerof2 (alignment / sizeof (void *)) != 0 || alignment == 0) return EINVAL; /* Call the hook here, so that caller is posix_memalign's caller and not posix_memalign itself. */ void *(*hook) (size_t, size_t, const void *) = force_reg (__memalign_hook); if (__builtin_expect (hook != NULL, 0)) mem = (*hook)(alignment, size, RETURN_ADDRESS (0)); else mem = __libc_memalign (alignment, size); if (mem != NULL) { // only set when success *memptr = mem; return 0; } return ENOMEM; } musl [6]: int posix_memalign(void **res, size_t align, size_t len) { if (align < sizeof(void *)) return EINVAL; void *mem = __memalign(align, len); if (!mem) return errno; *res = mem; // only set when success return 0; } [1] http://man7.org/linux/man-pages/man3/posix_memalign.3.html [2] http://pubs.opengroup.org/onlinepubs/009695399/ functions/posix_memalign.html [3] https://linux.die.net/man/3/posix_memalign [4] https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin. git;a=blob;f=newlib/libc/sys/linux/malloc.c;h=25007e8896a1292d6ae821aa35d7b8 973ea99bad;hb=HEAD#l4938 [5] https://sourceware.org/git/?p=glibc.git;a=blob;f=malloc/malloc.c;h= 488579390578e09a1a232ac5161daee086dbbd6b;hb=HEAD#l5086 [6] http://git.musl-libc.org/cgit/musl/tree/src/malloc/posix_memalign.c [7] https://github.com/gcc-mirror/gcc/blob/master/gcc/testsuite/ gcc.dg/torture/pr60092.c

2 2

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

devel March 2017