search for: sigact

...libs/klibc/klibc.git;a=commit;h=77c7ccc993d605eb7f6bef511017d28c1bb2d7e7 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Fri, 28 Aug 2020 19:21:47 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Fri, 28 Aug 2020 19:52:24 +0100 [klibc] signal: Move rt_sigaction() argument mangling to arch directories On alpha, sparc, and sparc64, the rt_sigaction() system call takes an additional parameter. Define another wrapper function on those architectures to handle this, instead of doing it directly in sigaction(). This is preparation for fixing signal handlin...

...commit;h=755b07d826618779a2232199cb53e2a44809ad89 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Fri, 28 Aug 2020 19:37:19 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Sat, 29 Aug 2020 17:22:38 +0100 [klibc] signal: Add config flag for additional sigaction fixup On ia64, sigaction() needs to make further changes to the given struct sigaction. Add a config flag for this, that forces copying of struct sigaction. Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/include/klibc/sysconfig.h | 11 +++++++++++ usr/klibc/sigaction.c...

...Thus to fix this regression we must construct descriptors solely for passing to/from the kernel. Despite this, usr/dash/static/sh -c "usr/utils/static/true; exit" was still seen to hang during the test suite. This is because sa_mask and sa_flags were round the wrong way (matching the old sigaction syscall rather than the newer rt_sigaction syscall, despite the former not existing on ia64), causing the kernel to think SA_NOCLDWAIT was set. This bug appears to have been present since the ia64 port was first committed. --- usr/include/arch/ia64/klibc/archsignal.h | 3 ++- usr/klibc/sigacti...

...org/?p=libs/klibc/klibc.git;a=commit;h=ae17951ae73ec52348c77259269637b854f0c236 Author: Jessica Clarke <jrtc27 at jrtc27.com> AuthorDate: Fri, 28 Aug 2020 19:48:32 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Sat, 29 Aug 2020 17:23:08 +0100 [klibc] ia64: Fix sigaction function implementation Commit 8418552 ("[klibc] ia64: Fix shared build") switched to the function descriptor-less ABI. However, when passing a function pointer to the kernel for signal handling, it always expects a descriptor, and -mno-pic does not set any ELF flags that the kernel c...

...org/?p=libs/klibc/klibc.git;a=commit;h=454901a0c733986bee36da111a08fe7df0ff0b98 Author: Jessica Clarke <jrtc27 at jrtc27.com> AuthorDate: Fri, 28 Aug 2020 19:48:32 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Fri, 28 Aug 2020 20:22:50 +0100 [klibc] ia64: Fix sigaction struct layout and function implementation Commit 8418552 ("[klibc] ia64: Fix shared build") switched to the function descriptor-less ABI. However, when passing a function pointer to the kernel for signal handling, it always expects a descriptor, and -mno-pic does not set any ELF flags...

...ae6250cf55b16fa65bdc490 Author: Thorsten Glaser <tg at mirbsd.org> AuthorDate: Sun, 27 May 2012 17:16:48 +0000 Committer: maximilian attems <max at stro.at> CommitDate: Tue, 29 May 2012 19:07:09 +0200 [klibc] alpha: fix signal handler setup on DEC Alpha We need a five-argument rt_sigaction syscall form with a hidden argument but that can be NULL as the kernel sets it up for us better than we can (easily) do, at speed cost. Signed-off-by: Thorsten Glaser <tg at mirbsd.org> [ Mark alpha as working. -maks ] Signed-off-by: maximilian attems <max at stro.at> --- usr/klib...

Commit 8418552 ("[klibc] ia64: Fix shared build") missed this use of the GP register, although the code appears to have been dubious anyway, assuming the address of errno was the first thing pointed to by GP. --- usr/klibc/arch/ia64/vfork.S | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/usr/klibc/arch/ia64/vfork.S b/usr/klibc/arch/ia64/vfork.S index

Hi, is there a DD-accessible porterbox somewhere (slow would be ok, as this is smallish software) with an up-to-date sid (enough to install the recently-built libklibc-dev 2.0~rc5-1 and all other B-D of mksh 40.9.20120518-1, as well as strace and gdb-minimal)? Similarily to http://www.zytor.com/pipermail/klibc/2012-May/003229.html I found klibc-compiled programmes on Alpha to fail (SIGSEGV

...sr/home/jfeise/source/openssh-2.5.1p1/openbsd-compat' gcc -g -O2 -Wall -I/usr/local/openssl -I/usr/local/openssl -I. -I.. -I. -I./.. -DHAVE_CONFIG_H -c bsd-arc4random.c In file included from ../includes.h:39, from bsd-arc4random.c:25: /usr/include/signal.h:217: warning: `struct sigaction' declared inside parameter list /usr/include/signal.h:217: warning: its scope is only this definition or declaration, /usr/include/signal.h:217: warning: which is probably not what you want. /usr/include/signal.h:219: warning: `struct sigaction' declared inside parameter list And this i...

...commit;h=a1556c691981c0651d8179db53665a6cf1be7953 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Fri, 28 Aug 2020 19:37:19 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Fri, 28 Aug 2020 20:22:50 +0100 [klibc] signal: Add config flag for additional sigaction fixup On ia64, sigaction() needs to make further changes to the given struct sigaction. Add a config flag for this, that forces copying of struct sigaction. Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/include/klibc/sysconfig.h | 11 +++++++++++ usr/klibc/sigaction.c...

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1"> <title></title> </head> <body text="#330000" bgcolor="#ffffff"> <tt><font size="+1">I am attempting to run the sun webserver 7

...el.org/?p=libs/klibc/klibc.git;a=commit;h=623b03fb63f0c0473b378cf61c77258fd450b5a1 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Fri, 28 Aug 2020 19:59:28 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Fri, 28 Aug 2020 20:22:50 +0100 [klibc] ia64: sigaction: Make signal updates atomic * If rt_sigaction() fails, undo the descriptor update * Mask the signal while we're (potentially) updating its descriptor Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/klibc/arch/ia64/sigaction.c | 14 ++++++++++++++ 1 file changed, 14 in...

...el.org/?p=libs/klibc/klibc.git;a=commit;h=69fb0c9b936dd59a7f7c9fab40f9206b80837b5b Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Fri, 28 Aug 2020 19:59:28 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Sat, 29 Aug 2020 17:23:08 +0100 [klibc] ia64: sigaction: Make signal updates atomic * If rt_sigaction() fails, undo the descriptor update * Mask the signal while we're (potentially) updating its descriptor Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/klibc/arch/ia64/sigaction.c | 14 ++++++++++++++ 1 file changed, 14 in...

..." r s y n c : c o n n e".., 63) = 63 poll(0xFFBEE0E8, 1, 60000) = 1 write(1, " ?\0\0\b r s y n c : c".., 67) Err#32 EPIPE Received signal #13, SIGPIPE [ignored] write(2, " r s y n c : w r i t e".., 81) Err#32 EPIPE Received signal #13, SIGPIPE [ignored] sigaction(SIGUSR1, 0xFFBEDAF8, 0xFFBEDB78) = 0 sigaction(SIGUSR2, 0xFFBEDAF8, 0xFFBEDB78) = 0 getpid() = 7970 [7969] Received signal #18, SIGCLD [caught] siginfo: SIGCLD CLD_EXITED pid=7971 status=0x000C waitid(P_ALL, 0, 0xFFBED6D8, WEXITED|WTRAPPED|WNOHANG) = 0 waitid(P_ALL, 0, 0xFFBED6D8,...

...klibc.git;a=commit;h=570ed1e207cbe38ed487c722f8ac7db68e664a94 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Sun, 23 Aug 2020 15:18:19 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Thu, 27 Aug 2020 15:00:33 +0100 [klibc] alpha: Pass restorer to rt_sigaction() and disable executable stack alpha does not support the SA_RESTORER flag, but allows specifiying a restorer callback as an additional parameter to rt_sigaction(). We should do this to avoid needing an executable stack. * Force the SA_SIGINFO flag on for all signal handlers, so that we can...

...at decadent.org.uk> CommitDate: Thu, 27 Aug 2020 15:00:33 +0100 [klibc] signal: Add sysconfig setting to force SA_SIGINFO on On alpha, arm, i386, m68k, powerpc, s390, sh, and sparc (32-bit), the kernel sets up the signal stack frame differently depending on the SA_SIGINFO flag, not whether the sigaction() or rt_sigaction() system call was used to install the handler. On alpha and sparc, we are going to start providing our own restorer that will call rt_sigaction(), so will need to ensure this flag is always set. Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/include/kli...

...enssh/bin/ssh\" -D_PATH_SSH_ASKPASS_DEFAULT=\"/opt/openssh/libexec/ssh-askpass\" -D_PATH_SFTP_SERVER=\"/opt/openssh/libexec/sftp-server\" -DHAVE_CONFIG_H -c cli.c In file included from includes.h:39, from cli.c:1: /usr/include/signal.h:216: warning: `struct sigaction' declared inside parameter list /usr/include/signal.h:216: warning: its scope is only this definition or declaration, /usr/include/signal.h:216: warning: which is probably not what you want. /usr/include/signal.h:218: warning: `struct sigaction' declared inside parameter list cli.c: In f...

...0 03:51:16 +0100 [klibc] signal: Add compile-time check on signal types If we are going to use the RT signal API, we must have the right type definitions. Check for signs that we have the old API definitions: * _NSIG less than 64 * sigset_t too small for _NSIG * sa_mask not at the end of struct sigaction Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/klibc/sigaction.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/usr/klibc/sigaction.c b/usr/klibc/sigaction.c index 19a8a54b..30ef39ef 100644 --- a/usr/klibc/sigaction.c +++ b/usr/klibc/sigaction.c @@ -3,6 +3,7...

...= 1 21056: write(1, " s", 1) = 1 21056: read(0, "\r", 1) = 1 21056: write(1, "\n", 1) = 1 21056: ioctl(0, TCSETSW, 0xFF345C28) = 0 21056: sigaction(SIGINT, 0xFFBED4C0, 0xFFBED5C0) = 0 21056: sigaction(SIGTERM, 0xFFBED4C0, 0xFFBED5C0) = 0 21056: sigaction(SIGQUIT, 0xFFBED4C0, 0xFFBED5C0) = 0 21056: sigaction(SIGALRM, 0xFFBED4C0, 0xFFBED5C0) = 0 21056: sigaction(SIGTSTP, 0xFFBED4C0, 0xFFBED5C0) = 0 21056: sigact...

...it;h=1cd11aaed1dece773c6b1ce2e99a0fe98b51321e Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Thu, 27 Aug 2020 01:58:19 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Thu, 27 Aug 2020 03:51:11 +0100 [klibc] alpha: Fix definitions of _NSIG and struct sigaction We use the RT signals API, but include the kernel UAPI header that defines _NSIG and struct sigaction for the old API. Copy over all the definitions and fix those two. Signed-off-by: Ben Hutchings <ben at decadent.org.uk> --- usr/include/arch/alpha/klibc/archsignal.h | 78 ++++++++++++...