devel June 2015

devel@uclibc-ng.org

4 participants
25 discussions

uClibc-ng - small C library for embedded systems branch 1.0 updated. v1.0.2-44-g7d30384
by wbx＠helium.waldemar-brodkorb.de 30 Jun '15

30 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "uClibc-ng - small C library for embedded systems". The branch, 1.0 has been updated via 7d303841dbb6e668e5acdd9218ce82231a00a967 (commit) from f347524fade67b41fb8e46283a6f0ae4aa19e6b0 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 7d303841dbb6e668e5acdd9218ce82231a00a967 Author: Vineet Gupta <Vineet.Gupta1(a)synopsys.com> Date: Tue Jun 30 17:46:56 2015 +0530 NPTL/arc: notify kernel of the TP value Native gdb makes a ptrace (GET_THREAD_AREA) which needs to return the TP. however when libc sets up TP reg (for main thread), it doesn't call arc_settls syscall so kernel doesn't know of TP register details (moreso because clone doesnt have SETTLS flag) Note that kernel doesn't know about r25 being TP etc. This commit got lost in merge of NPTL tools into arc-mainline-dev and showed up again as STAR 9000919529 (native gdb can't debug threaded apps) ------->8--------------- [ARCLinux]# gdb ./pth Reading symbols from ./pth...(no debugging symbols found)...done. (gdb) b main Breakpoint 1 at 0x106f2 (gdb) r Starting program: /pth [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/libthread_db.so.1". thread_get_info_callback: cannot get thread info: generic error (gdb) q ------->8--------------- Debugged-by: Anton Kolesov <akolesov(a)synopsys.com> Signed-off-by: Vineet Gupta <vgupta(a)synopsys.com> Signed-off-by: Vineet Gupta <vgupta(a)synopsys.com> ----------------------------------------------------------------------- Summary of changes: libpthread/nptl/sysdeps/arc/tls.h | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/libpthread/nptl/sysdeps/arc/tls.h b/libpthread/nptl/sysdeps/arc/tls.h index 63dd113..4af415a 100644 --- a/libpthread/nptl/sysdeps/arc/tls.h +++ b/libpthread/nptl/sysdeps/arc/tls.h @@ -104,8 +104,11 @@ typedef struct operation can cause a failure 'errno' must not be touched. */ # define TLS_INIT_TP(tcbp, secondcall) \ ({ \ - __builtin_set_thread_pointer(tcbp); \ - NULL; \ + long result_var; \ + __builtin_set_thread_pointer(tcbp); \ + result_var = INTERNAL_SYSCALL (arc_settls, err, 1, (tcbp)); \ + INTERNAL_SYSCALL_ERROR_P (result_var, err) \ + ? "unknown error" : NULL; \ }) /* Return the address of the dtv for the current thread. hooks/post-receive -- uClibc-ng - small C library for embedded systems

1 0

uClibc-ng - small C library for embedded systems branch master updated. 6da5676cbe02149ef64197927b8b2fbec6055fc9
by wbx＠helium.waldemar-brodkorb.de 30 Jun '15

30 Jun '15

1 0

uClibc-ng - small C library for embedded systems branch master updated. 5d793a8cfcf60e8e0bda3e6e2c1ab7a9aa30c558
by wbx＠helium.waldemar-brodkorb.de 26 Jun '15

26 Jun '15

1 0

uClibc-ng - small C library for embedded systems branch master updated. 42d180da69fd31f3a8af91c9470da6abdf91574a
by wbx＠helium.waldemar-brodkorb.de 26 Jun '15

26 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "uClibc-ng - small C library for embedded systems". The branch, master has been updated via 42d180da69fd31f3a8af91c9470da6abdf91574a (commit) from 42f4f533c7a4505dd8ee53b42f7fb3db1877f520 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 42d180da69fd31f3a8af91c9470da6abdf91574a Author: Waldemar Brodkorb <wbx(a)openadk.org> Date: Wed Jun 10 10:16:24 2015 -0500 use mktemp, otherwise test is failing. reverts partially 7c721d31e4b7a0bdf6f803b8e7c38996bf60b59f ----------------------------------------------------------------------- Summary of changes: test/nptl/tst-cancel4.c | 53 ++++++++++++++++------------------------------- 1 file changed, 18 insertions(+), 35 deletions(-) diff --git a/test/nptl/tst-cancel4.c b/test/nptl/tst-cancel4.c index 4ba4045..ecaf297 100644 --- a/test/nptl/tst-cancel4.c +++ b/test/nptl/tst-cancel4.c @@ -302,7 +302,8 @@ tf_writev (void *arg) exit (1); } - +/* sleep is not early cancelable, disable for now */ +#if 0 static void * tf_sleep (void *arg) { @@ -333,7 +334,7 @@ tf_sleep (void *arg) exit (1); } - +#endif static void * tf_usleep (void *arg) @@ -996,7 +997,6 @@ tf_pause (void *arg) static void * tf_accept (void *arg) { - int tfd; struct sockaddr_un sun; /* To test a non-blocking accept call we make the call file by using a datagrame socket. */ @@ -1020,13 +1020,11 @@ tf_accept (void *arg) } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-1-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1071,7 +1069,6 @@ tf_accept (void *arg) static void * tf_send (void *arg) { - int tfd; struct sockaddr_un sun; tempfd = socket (AF_UNIX, SOCK_STREAM, 0); @@ -1087,16 +1084,15 @@ tf_send (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-2-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1155,7 +1151,6 @@ tf_send (void *arg) static void * tf_recv (void *arg) { - int tfd; struct sockaddr_un sun; tempfd = socket (AF_UNIX, SOCK_STREAM, 0); @@ -1171,16 +1166,15 @@ tf_recv (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-3-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1238,7 +1232,6 @@ tf_recv (void *arg) static void * tf_recvfrom (void *arg) { - int tfd; struct sockaddr_un sun; tempfd = socket (AF_UNIX, SOCK_DGRAM, 0); @@ -1254,16 +1247,15 @@ tf_recvfrom (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-4-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1315,7 +1307,6 @@ tf_recvfrom (void *arg) static void * tf_recvmsg (void *arg) { - int tfd; struct sockaddr_un sun; tempfd = socket (AF_UNIX, SOCK_DGRAM, 0); @@ -1331,16 +1322,15 @@ tf_recvmsg (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-5-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1693,7 +1683,6 @@ tf_msync (void *arg) static void * tf_sendto (void *arg) { - int tfd; if (arg == NULL) // XXX If somebody can provide a portable test case in which sendto() // blocks we can enable this test to run in both rounds. @@ -1714,16 +1703,15 @@ tf_sendto (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-6-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1771,7 +1759,6 @@ tf_sendto (void *arg) static void * tf_sendmsg (void *arg) { - int tfd; if (arg == NULL) // XXX If somebody can provide a portable test case in which sendmsg() // blocks we can enable this test to run in both rounds. @@ -1792,16 +1779,15 @@ tf_sendmsg (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-7-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -1893,7 +1879,6 @@ tf_creat (void *arg) static void * tf_connect (void *arg) { - int tfd; if (arg == NULL) // XXX If somebody can provide a portable test case in which connect() // blocks we can enable this test to run in both rounds. @@ -1914,16 +1899,15 @@ tf_connect (void *arg) if (++tries > 10) { printf ("%s: too many unsuccessful bind calls\n", __FUNCTION__); + exit (1); } strcpy (sun.sun_path, "/tmp/tst-cancel4-socket-2-XXXXXX"); - tfd = mkstemp(sun.sun_path); - if (tfd < 0) + if (mktemp (sun.sun_path) == NULL) { printf ("%s: cannot generate temp file name\n", __FUNCTION__); exit (1); } - close(tfd); sun.sun_family = AF_UNIX; } while (bind (tempfd, (struct sockaddr *) &sun, @@ -2135,7 +2119,6 @@ static struct ADD_TEST (ppoll, 2, 0), ADD_TEST (write, 2, 0), ADD_TEST (writev, 2, 0), - ADD_TEST (sleep, 2, 0), ADD_TEST (usleep, 2, 0), ADD_TEST (nanosleep, 2, 0), ADD_TEST (wait, 2, 0), hooks/post-receive -- uClibc-ng - small C library for embedded systems

1 0

uClibc-ng - small C library for embedded systems branch master updated. 42f4f533c7a4505dd8ee53b42f7fb3db1877f520
by wbx＠helium.waldemar-brodkorb.de 24 Jun '15

24 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "uClibc-ng - small C library for embedded systems". The branch, master has been updated via 42f4f533c7a4505dd8ee53b42f7fb3db1877f520 (commit) from 8869e46f53d66c447fd32086da6c508c03bd1b19 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 42f4f533c7a4505dd8ee53b42f7fb3db1877f520 Author: Waldemar Brodkorb <wbx(a)openadk.org> Date: Wed Jun 24 15:54:28 2015 -0500 mips64: patches from OpenWrt Add these changes to get mips64 with n32 ABI working. ----------------------------------------------------------------------- Summary of changes: extra/Configs/Config.in | 1 + libc/sysdeps/linux/mips/sysdep.h | 20 ++- .../unix/sysv/linux/mips/mips64/sysdep-cancel.h | 182 ++++++++++++++++++++ 3 files changed, 202 insertions(+), 1 deletion(-) create mode 100644 libpthread/nptl/sysdeps/unix/sysv/linux/mips/mips64/sysdep-cancel.h diff --git a/extra/Configs/Config.in b/extra/Configs/Config.in index 2f64b79..4fbd398 100644 --- a/extra/Configs/Config.in +++ b/extra/Configs/Config.in @@ -226,6 +226,7 @@ config TARGET_SUBARCH default "i586" if CONFIG_586 default "i686" if CONFIG_686 default "arcv2" if CONFIG_ARC_CPU_HS + default "mips64" if CONFIG_MIPS_N64_ABI default "" source "extra/Configs/Config.in.arch" diff --git a/libc/sysdeps/linux/mips/sysdep.h b/libc/sysdeps/linux/mips/sysdep.h index 86873ac..0ed0500 100644 --- a/libc/sysdeps/linux/mips/sysdep.h +++ b/libc/sysdeps/linux/mips/sysdep.h @@ -93,7 +93,8 @@ backwards into the previous fn. */ #ifdef __PIC__ -#define PSEUDO(name, syscall_name, args) \ +# if _MIPS_SIM == _ABIO32 +# define PSEUDO(name, syscall_name, args) \ .align 2; \ 99: move a0, v0; \ la t9,__syscall_error; \ @@ -106,6 +107,23 @@ .set reorder; \ bne a3, zero, 99b; \ L(syse1): +# else +# define PSEUDO(name, syscall_name, args) \ + .align 2; \ + 99: \ + .set noat; \ + .cpsetup t9, $1, name; \ + .set at; \ + move a0, v0; \ + dla t9,__syscall_error; \ + .cpreturn; \ + jr t9; \ + ENTRY(name) \ + li v0, SYS_ify(syscall_name); \ + syscall; \ + bne a3, zero, 99b; \ +L(syse1): +# endif #else #define PSEUDO(name, syscall_name, args) \ .set noreorder; \ diff --git a/libpthread/nptl/sysdeps/unix/sysv/linux/mips/mips64/sysdep-cancel.h b/libpthread/nptl/sysdeps/unix/sysv/linux/mips/mips64/sysdep-cancel.h new file mode 100644 index 0000000..74f3db9 --- /dev/null +++ b/libpthread/nptl/sysdeps/unix/sysv/linux/mips/mips64/sysdep-cancel.h @@ -0,0 +1,182 @@ +/* Copyright (C) 2003, 2004, 2005, 2006 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 + 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 + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library. If not, see + <http://www.gnu.org/licenses/>. */ + +#include <sysdep.h> +#include <tls.h> +#ifndef __ASSEMBLER__ +# include <pthreadP.h> +#endif +#include <sys/asm.h> + +/* Gas will put the initial save of $gp into the CIE, because it appears to + happen before any instructions. So we use cfi_same_value instead of + cfi_restore. */ + +#if !defined NOT_IN_libc || defined IS_IN_libpthread || defined IS_IN_librt + +#ifdef __PIC__ +# undef PSEUDO +# define PSEUDO(name, syscall_name, args) \ + .align 2; \ + L(pseudo_start): \ + cfi_startproc; \ + cfi_adjust_cfa_offset (STKSPACE); \ + cfi_rel_offset (gp, STKOFF_GP); \ + 99: move a0, v0; \ + PTR_LA t9,__syscall_error; \ + /* manual cpreturn */ \ + REG_L gp, STKOFF_GP(sp); \ + cfi_same_value (gp); \ + RESTORESTK; \ + jr t9; \ + .type __##syscall_name##_nocancel, @function; \ + .globl __##syscall_name##_nocancel; \ + __##syscall_name##_nocancel: \ + SAVESTK; \ + .cpsetup t9, STKOFF_GP, name; \ + cfi_rel_offset (gp, STKOFF_GP); \ + li v0, SYS_ify(syscall_name); \ + syscall; \ + bne a3, zero, SYSCALL_ERROR_LABEL; \ + /* manual cpreturn */ \ + REG_L gp, STKOFF_GP(sp); \ + cfi_same_value (gp); \ + RESTORESTK; \ + ret; \ + .size __##syscall_name##_nocancel,.-__##syscall_name##_nocancel; \ + ENTRY (name) \ + SAVESTK; \ + .cpsetup t9, STKOFF_GP, name; \ + cfi_rel_offset (gp, STKOFF_GP); \ + SINGLE_THREAD_P(v1); \ + bne zero, v1, L(pseudo_cancel); \ + .set noreorder; \ + li v0, SYS_ify(syscall_name); \ + syscall; \ + .set reorder; \ + bne a3, zero, SYSCALL_ERROR_LABEL; \ + /* manual cpreturn */ \ + REG_L gp, STKOFF_GP(sp); \ + cfi_same_value (gp); \ + RESTORESTK; \ + ret; \ + L(pseudo_cancel): \ + cfi_adjust_cfa_offset (STKSPACE); \ + cfi_rel_offset (gp, STKOFF_GP); \ + REG_S ra, STKOFF_RA(sp); \ + cfi_rel_offset (ra, STKOFF_RA); \ + PUSHARGS_##args; /* save syscall args */ \ + CENABLE; \ + REG_S v0, STKOFF_SVMSK(sp); /* save mask */ \ + POPARGS_##args; /* restore syscall args */ \ + .set noreorder; \ + li v0, SYS_ify (syscall_name); \ + syscall; \ + .set reorder; \ + REG_S v0, STKOFF_SC_V0(sp); /* save syscall result */ \ + REG_S a3, STKOFF_SC_ERR(sp); /* save syscall error flag */ \ + REG_L a0, STKOFF_SVMSK(sp); /* pass mask as arg1 */ \ + CDISABLE; \ + REG_L a3, STKOFF_SC_ERR(sp); /* restore syscall error flag */ \ + REG_L ra, STKOFF_RA(sp); /* restore return address */ \ + REG_L v0, STKOFF_SC_V0(sp); /* restore syscall result */ \ + bne a3, zero, SYSCALL_ERROR_LABEL; \ + /* manual cpreturn */ \ + REG_L gp, STKOFF_GP(sp); \ + cfi_same_value (gp); \ + RESTORESTK; \ + L(pseudo_end): + + +# undef PSEUDO_END +# define PSEUDO_END(sym) cfi_endproc; .end sym; .size sym,.-sym + +#endif + +# define PUSHARGS_0 /* nothing to do */ +# define PUSHARGS_1 PUSHARGS_0 REG_S a0, STKOFF_A0(sp); cfi_rel_offset (a0, STKOFF_A0); +# define PUSHARGS_2 PUSHARGS_1 REG_S a1, STKOFF_A1(sp); cfi_rel_offset (a1, STKOFF_A1); +# define PUSHARGS_3 PUSHARGS_2 REG_S a2, STKOFF_A2(sp); cfi_rel_offset (a2, STKOFF_A2); +# define PUSHARGS_4 PUSHARGS_3 REG_S a3, STKOFF_A3(sp); cfi_rel_offset (a3, STKOFF_A3); +# define PUSHARGS_5 PUSHARGS_4 REG_S a4, STKOFF_A4(sp); cfi_rel_offset (a3, STKOFF_A4); +# define PUSHARGS_6 PUSHARGS_5 REG_S a5, STKOFF_A5(sp); cfi_rel_offset (a3, STKOFF_A5); + +# define POPARGS_0 /* nothing to do */ +# define POPARGS_1 POPARGS_0 REG_L a0, STKOFF_A0(sp); +# define POPARGS_2 POPARGS_1 REG_L a1, STKOFF_A1(sp); +# define POPARGS_3 POPARGS_2 REG_L a2, STKOFF_A2(sp); +# define POPARGS_4 POPARGS_3 REG_L a3, STKOFF_A3(sp); +# define POPARGS_5 POPARGS_4 REG_L a4, STKOFF_A4(sp); +# define POPARGS_6 POPARGS_5 REG_L a5, STKOFF_A5(sp); + +/* Save an even number of slots. Should be 0 if an even number of slots + are used below, or SZREG if an odd number are used. */ +# define STK_PAD SZREG + +/* Place values that we are more likely to use later in this sequence, i.e. + closer to the SP at function entry. If you do that, the are more + likely to already be in your d-cache. */ +# define STKOFF_A5 (STK_PAD) +# define STKOFF_A4 (STKOFF_A5 + SZREG) +# define STKOFF_A3 (STKOFF_A4 + SZREG) +# define STKOFF_A2 (STKOFF_A3 + SZREG) /* MT and more args. */ +# define STKOFF_A1 (STKOFF_A2 + SZREG) /* MT and 2 args. */ +# define STKOFF_A0 (STKOFF_A1 + SZREG) /* MT and 1 arg. */ +# define STKOFF_RA (STKOFF_A0 + SZREG) /* Used if MT. */ +# define STKOFF_SC_V0 (STKOFF_RA + SZREG) /* Used if MT. */ +# define STKOFF_SC_ERR (STKOFF_SC_V0 + SZREG) /* Used if MT. */ +# define STKOFF_SVMSK (STKOFF_SC_ERR + SZREG) /* Used if MT. */ +# define STKOFF_GP (STKOFF_SVMSK + SZREG) /* Always used. */ + +# define STKSPACE (STKOFF_GP + SZREG) +# define SAVESTK PTR_SUBU sp, STKSPACE; cfi_adjust_cfa_offset(STKSPACE) +# define RESTORESTK PTR_ADDU sp, STKSPACE; cfi_adjust_cfa_offset(-STKSPACE) + +# ifdef IS_IN_libpthread +# define CENABLE PTR_LA t9, __pthread_enable_asynccancel; jalr t9 +# define CDISABLE PTR_LA t9, __pthread_disable_asynccancel; jalr t9 +# elif defined IS_IN_librt +# define CENABLE PTR_LA t9, __librt_enable_asynccancel; jalr t9 +# define CDISABLE PTR_LA t9, __librt_disable_asynccancel; jalr t9 +# else +# define CENABLE PTR_LA t9, __libc_enable_asynccancel; jalr t9 +# define CDISABLE PTR_LA t9, __libc_disable_asynccancel; jalr t9 +# endif + +# ifndef __ASSEMBLER__ +# define SINGLE_THREAD_P \ + __builtin_expect (THREAD_GETMEM (THREAD_SELF, \ + header.multiple_threads) \ + == 0, 1) +# else +# define SINGLE_THREAD_P(reg) \ + READ_THREAD_POINTER(reg); \ + lw reg, MULTIPLE_THREADS_OFFSET(reg) +#endif + +#elif !defined __ASSEMBLER__ + +# define SINGLE_THREAD_P 1 +# define NO_CANCELLATION 1 + +#endif + +#ifndef __ASSEMBLER__ +# define RTLD_SINGLE_THREAD_P \ + __builtin_expect (THREAD_GETMEM (THREAD_SELF, \ + header.multiple_threads) == 0, 1) +#endif hooks/post-receive -- uClibc-ng - small C library for embedded systems

1 0

[PATCH 0/2] xtensa: fix failing tst-mqueue* tests
by Max Filippov 23 Jun '15

23 Jun '15

Hi Waldemar, these two patches fix tst-mqueue* tests that are currently failing on xtensa. Max Filippov (2): xtensa: add ret_ERRVAL definition xtensa: fix stack frame size for NPTL libc/sysdeps/linux/xtensa/sysdep.h | 5 +++++ libpthread/nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h | 10 +++++++++- 2 files changed, 14 insertions(+), 1 deletion(-) -- 1.8.1.4

2 3

uClibc-ng - small C library for embedded systems branch 1.0 updated. v1.0.2-43-gf347524
by wbx＠helium.waldemar-brodkorb.de 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "uClibc-ng - small C library for embedded systems". The branch, 1.0 has been updated via f347524fade67b41fb8e46283a6f0ae4aa19e6b0 (commit) via f3a8267dabab20b095abe36e4a2be14ccc34b38e (commit) from d551a039ad443e1ba1c29461b43ca865785fbb68 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit f347524fade67b41fb8e46283a6f0ae4aa19e6b0 Author: Max Filippov <jcmvbkbc(a)gmail.com> Date: Mon Jun 22 04:10:54 2015 +0300 xtensa: add ret_ERRVAL definition ret_ERRVAL is used by mq_timedsend and mq_timedreceive, it needs to be defined to retw, otherwise error return from those functions segfaults. This fixes the following testsuite failures: .... tst-mqueue1 FAIL tst-mqueue1 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' .... tst-mqueue2 FAIL tst-mqueue2 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' .... tst-mqueue3 FAIL tst-mqueue3 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' .... tst-mqueue4 FAIL tst-mqueue4 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' .... tst-mqueue5 FAIL tst-mqueue5 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' .... tst-mqueue6 FAIL tst-mqueue6 got 1 expected 0 Didn't expect signal from child: got `Segmentation fault' Signed-off-by: Max Filippov <jcmvbkbc(a)gmail.com> commit f3a8267dabab20b095abe36e4a2be14ccc34b38e Author: Max Filippov <jcmvbkbc(a)gmail.com> Date: Mon Jun 22 04:10:55 2015 +0300 xtensa: fix stack frame size for NPTL Cancellable syscalls use call8 to call functions that enable/disable cancellation, thus they cannot use the default FRAMESIZE. Redefine FRAMESIZE for such syscalls. This fixes the following testsuite failure: .... tst-mqueue8 FAIL tst-mqueue8 got 1 expected 0 going to cancel mq_receive in-time in-time mq_receive cancellation succeeded going to cancel mq_receive early Didn't expect signal from child: got `Segmentation fault' Signed-off-by: Max Filippov <jcmvbkbc(a)gmail.com> ----------------------------------------------------------------------- Summary of changes: libc/sysdeps/linux/xtensa/sysdep.h | 5 +++++ .../nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h | 10 +++++++++- 2 files changed, 14 insertions(+), 1 deletion(-) diff --git a/libc/sysdeps/linux/xtensa/sysdep.h b/libc/sysdeps/linux/xtensa/sysdep.h index cab4a2f..4873a41 100644 --- a/libc/sysdeps/linux/xtensa/sysdep.h +++ b/libc/sysdeps/linux/xtensa/sysdep.h @@ -72,7 +72,9 @@ #define JUMPTARGET(name) name #endif +#ifndef FRAMESIZE #define FRAMESIZE 16 +#endif #define CALL_MCOUNT /* Do nothing. */ @@ -130,6 +132,9 @@ #define PSEUDO_END_ERRVAL(name) \ END (name) +#undef ret_ERRVAL +#define ret_ERRVAL retw + #if defined _LIBC_REENTRANT # if defined USE___THREAD # ifndef NOT_IN_libc diff --git a/libpthread/nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h b/libpthread/nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h index 6bb2aeb..f82957d 100644 --- a/libpthread/nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h +++ b/libpthread/nptl/sysdeps/unix/sysv/linux/xtensa/sysdep-cancel.h @@ -24,7 +24,15 @@ #endif #if !defined NOT_IN_libc || defined IS_IN_libpthread || defined IS_IN_librt -// FIXME: ENTRY includes an entry instruction, here we'd want entry sp, 48! + +/* CENABLE/CDISABLE in PSEUDO below use call8, stack frame size must be + * at least 32. + */ +#if FRAMESIZE < 32 +#undef FRAMESIZE +#define FRAMESIZE 32 +#endif + # undef PSEUDO # define PSEUDO(name, syscall_name, args) \ .text; \ hooks/post-receive -- uClibc-ng - small C library for embedded systems

1 0

uClibc-ng - small C library for embedded systems branch master updated. 8869e46f53d66c447fd32086da6c508c03bd1b19
by wbx＠helium.waldemar-brodkorb.de 23 Jun '15

23 Jun '15

1 0

uClibc-ng - small C library for embedded systems branch 1.0 updated. v1.0.2-41-gd551a03
by wbx＠helium.waldemar-brodkorb.de 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "uClibc-ng - small C library for embedded systems". The branch, 1.0 has been updated via d551a039ad443e1ba1c29461b43ca865785fbb68 (commit) from afad51b5f4b0093937304ca99fb8b679ab47eb56 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit d551a039ad443e1ba1c29461b43ca865785fbb68 Author: Waldemar Brodkorb <wbx(a)openadk.org> Date: Tue Jun 23 21:18:49 2015 +0200 add patch from ldso-future branch Fixes segfaults when gcc 5.1 is used for x86. http://git.uclibc.org/uClibc/commit/ldso/ldso/i386/dl-sysdep.h?h=ldso-futur… Mentioned in #uclibc. ----------------------------------------------------------------------- Summary of changes: ldso/ldso/i386/dl-sysdep.h | 30 +++++++++++++----------------- 1 file changed, 13 insertions(+), 17 deletions(-) diff --git a/ldso/ldso/i386/dl-sysdep.h b/ldso/ldso/i386/dl-sysdep.h index a66c802..8efa1a8 100644 --- a/ldso/ldso/i386/dl-sysdep.h +++ b/ldso/ldso/i386/dl-sysdep.h @@ -37,31 +37,27 @@ extern unsigned long _dl_linux_resolver(struct elf_resolve * tpnt, int reloc_ent | (((type) == R_386_COPY) * ELF_RTYPE_CLASS_COPY)) /* Return the link-time address of _DYNAMIC. Conveniently, this is the - first element of the GOT. This must be inlined in a function which - uses global data. */ -static __always_inline Elf32_Addr elf_machine_dynamic (void) attribute_unused; -static __always_inline Elf32_Addr + first element of the GOT, a special entry that is never relocated. */ +extern const Elf32_Addr _GLOBAL_OFFSET_TABLE_[] attribute_hidden; +static __always_inline Elf32_Addr __attribute__ ((unused, const)) elf_machine_dynamic (void) { - register Elf32_Addr *got __asm__ ("%ebx"); - return *got; + /* This produces a GOTOFF reloc that resolves to zero at link time, so in + fact just loads from the GOT register directly. By doing it without + an asm we can let the compiler choose any register. */ + return _GLOBAL_OFFSET_TABLE_[0]; } +extern Elf32_Dyn bygotoff[] __asm__ ("_DYNAMIC") attribute_hidden; /* Return the run-time load address of the shared object. */ -static __always_inline Elf32_Addr elf_machine_load_address (void) attribute_unused; -static __always_inline Elf32_Addr +static __always_inline Elf32_Addr attribute_unused elf_machine_load_address (void) { - /* It doesn't matter what variable this is, the reference never makes - it to assembly. We need a dummy reference to some global variable - via the GOT to make sure the compiler initialized %ebx in time. */ - Elf32_Addr addr; - int tmp; - __asm__ ("leal _dl_start@GOTOFF(%%ebx), %0\n" - "subl _dl_start@GOT(%%ebx), %0" - : "=r" (addr) : "m" (tmp) : "cc"); - return addr; + /* Compute the difference between the runtime address of _DYNAMIC as seen + by a GOTOFF reference, and the link-time address found in the special + unrelocated first GOT entry. */ + return (Elf32_Addr) &bygotoff - elf_machine_dynamic (); } static __always_inline void hooks/post-receive -- uClibc-ng - small C library for embedded systems

1 0

uClibc-ng - small C library for embedded systems branch master updated. 1d81dc5276b78a6f893c35ffd80bbf331d4f0833
by wbx＠helium.waldemar-brodkorb.de 23 Jun '15

23 Jun '15

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

devel June 2015