Bug#1009712: sv: fails to control the service on Hurd

2022-12-21 Thread João
Hello, Coming back to this bug. I noticed that supervise/ok that sv is trying to open is a named pipe not a file. I don't know if this is telling in some way. I attach a more detailed debug log. Problems seem to occur in __dir_lookup. Best wishes, João Breakpoint 1 at 0x17a0: file

Bug#1009712: sv: fails to control the service on Hurd

2022-07-05 Thread Samuel Thibault
João Pedro Malhado, le mar. 05 juil. 2022 19:37:16 +0200, a ecrit: > However errnno is already 1073741902 before the function call, so does not > really change after the function call. Perhaps change the source code to set it to 0, just to be sure. > open64 (__oflag=10, __path=0x803caa0

Bug#1009712: sv: fails to control the service on Hurd

2022-07-05 Thread João Pedro Malhado
Hello Samuel, As you suggested the open64() function being used seems to be that defined in sysdeps/mach/hurd/open.c However errnno is already 1073741902 before the function call, so does not really change after the function call. On the other hand the variable fd that stores the value returned

Bug#1009712: sv: fails to control the service on Hurd

2022-07-03 Thread Samuel Thibault
João Pedro Malhado, le lun. 04 juil. 2022 00:14:43 +0200, a ecrit: > Below is a step by step debug trace with libc symbols. At the last step gdb > (or > the whole machine?) locks up. I would say strange, but maybe I shouldn't ;) That's also expected: you're stepping into very critical sections,

Bug#1009712: sv: fails to control the service on Hurd

2022-07-03 Thread João Pedro Malhado
Hello Samuel, On Sun, Jul 03, 2022 at 10:09:04PM +0200, Samuel Thibault wrote: > João Pedro Malhado, le dim. 03 juil. 2022 21:41:05 +0200, a ecrit: > > > This looks strange and I cannot interpret what is going on. > > This is not strange, see the values of E* macros I'm glad that you don't find

Bug#1009712: sv: fails to control the service on Hurd

2022-07-03 Thread Samuel Thibault
João Pedro Malhado, le dim. 03 juil. 2022 21:41:05 +0200, a ecrit: > On Sun, Jul 03, 2022 at 06:37:46PM +0200, Samuel Thibault wrote: > > João, le sam. 02 juil. 2022 20:53:55 +0200, a ecrit: > > > I don't know why open64() is returning -1. > > > > That's usually given in errno. > > Trying to

Bug#1009712: sv: fails to control the service on Hurd

2022-07-03 Thread João Pedro Malhado
X-Debbugs-CC: debian-h...@lists.debian.org Hello Samuel, On Sun, Jul 03, 2022 at 06:37:46PM +0200, Samuel Thibault wrote: > João, le sam. 02 juil. 2022 20:53:55 +0200, a ecrit: > > I don't know why open64() is returning -1. > > That's usually given in errno. Trying to print the errno on gdb

Bug#1009712: sv: fails to control the service on Hurd

2022-07-03 Thread Samuel Thibault
Hello, João, le sam. 02 juil. 2022 20:53:55 +0200, a ecrit: > I have put sv on a debugger and seems that the failure is happening when > trying > to open the supervise/ok file (the file exists, and is a 0 bytes file), > open_write() returns -1. > I don't know why open64() is returning -1.

Bug#1009712: sv: fails to control the service on Hurd

2022-07-02 Thread João
X-Debbugs-CC: debian-h...@lists.debian.org Hello Lorenzo, I have put sv on a debugger and seems that the failure is happening when trying to open the supervise/ok file (the file exists, and is a 0 bytes file), open_write() returns -1. Here is a backtrace. #0 open64 (__oflag=10,