[Bug 1202385] New: kernel: debuginfo for vdso*.so should be in a separate file
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385 Bug ID: 1202385 Summary: kernel: debuginfo for vdso*.so should be in a separate file Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Basesystem Assignee: screening-team-bugs@suse.de Reporter: martin.liska@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- As mentioned here: https://sourceware.org/bugzilla/show_bug.cgi?id=29478 and here: https://bugzilla.redhat.com/show_bug.cgi?id=1970578 debuginfod faces serious troubles when it comes to serving a VDSO debug info that is part of every executable. Thus a suggested fix can be putting the vdso a separate kernel-source-userspace package or so. Doing that, debuginfod won't have to search for the file in a huge kernel-default-debuginfo-5.19.0-1.1.x86_64.rpm file compressed with an even slower bzip2 algorithm. -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
Martin Li��ka
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c1
Jiri Slaby
Thus a suggested fix can be putting the vdso a separate kernel-source-userspace package or so.
It'd have to be in kernel-%{flavor}-vdso instead -- it's different for different flavors. In that case kernel-%{flavor}-vdso-debug* should be created automatically. I think it's not that hard to do that on the kernel side. The questions is: do we also need "Require: kernel-%{flavor}-vdso" in kernel-%{flavor} to drag it automatically in? I.e. what tools read vdso.so if any at all? -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c2
--- Comment #2 from Martin Li��ka
(In reply to Martin Li��ka from comment #0)
Thus a suggested fix can be putting the vdso a separate kernel-source-userspace package or so.
It'd have to be in kernel-%{flavor}-vdso instead -- it's different for different flavors. In that case kernel-%{flavor}-vdso-debug* should be created automatically.
Ah, I see.
I think it's not that hard to do that on the kernel side. The questions is: do we also need "Require: kernel-%{flavor}-vdso" in kernel-%{flavor} to drag it automatically in?
I guess so ...
I.e. what tools read vdso.so if any at all?
Because dynamic linker loads it if I'm correct: $ ldd `which echo` linux-vdso.so.1 (0x00007ffff7fc1000) libc.so.6 => /lib64/libc.so.6 (0x00007ffff7c00000) /lib64/ld-linux-x86-64.so.2 (0x00007ffff7fc3000) -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c3
--- Comment #3 from Jiri Slaby
I.e. what tools read vdso.so if any at all?
Because dynamic linker loads it if I'm correct:
$ ldd `which echo` linux-vdso.so.1 (0x00007ffff7fc1000) libc.so.6 => /lib64/libc.so.6 (0x00007ffff7c00000) /lib64/ld-linux-x86-64.so.2 (0x00007ffff7fc3000)
Nope, IMO it's supposed to be loaded by the kernel from the built-in image (not from disk). By load_elf_fdpic_binary() -> arch_setup_additional_pages() -> map_vdso_randomized() -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c4
--- Comment #4 from Michal Suchanek
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c5
--- Comment #5 from Martin Li��ka
(In reply to Martin Li��ka from comment #2)
I.e. what tools read vdso.so if any at all?
Because dynamic linker loads it if I'm correct:
$ ldd `which echo` linux-vdso.so.1 (0x00007ffff7fc1000) libc.so.6 => /lib64/libc.so.6 (0x00007ffff7c00000) /lib64/ld-linux-x86-64.so.2 (0x00007ffff7fc3000)
Nope, IMO it's supposed to be loaded by the kernel from the built-in image (not from disk). By load_elf_fdpic_binary() -> arch_setup_additional_pages() -> map_vdso_randomized()
Ah, thanks for the explanation! -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c6
--- Comment #6 from Martin Li��ka
The question is how do we pull in the vdso debuginfo?
In my case it's loaded by gdb via debuginfod: $ gdb `which make` ... Downloading 0.02 MB separate debug info for system-supplied DSO at 0x7ffff7fc1000 ...
Also ideally both the .so and the .so.debug would be in the same package but I don't know if rpm even supports that.
They are automatically separated to a foobar-debug{info,source} packages. -- You are receiving this mail because: You are on the CC list for the bug.
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c7
Jiri Slaby
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c8
Martin Li��ka
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c9
Michal Suchanek
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c10
Jiri Slaby
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c14
--- Comment #14 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c15
--- Comment #15 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c31
--- Comment #31 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c32
--- Comment #32 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c33
--- Comment #33 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c36
--- Comment #36 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c37
--- Comment #37 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c39
--- Comment #39 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c40
--- Comment #40 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c41
--- Comment #41 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c42
--- Comment #42 from Swamp Workflow Management
![](https://seccdn.libravatar.org/avatar/a895f78a81a109471893519443e4d933.jpg?s=120&d=mm&r=g)
https://bugzilla.suse.com/show_bug.cgi?id=1202385
https://bugzilla.suse.com/show_bug.cgi?id=1202385#c47
--- Comment #47 from Swamp Workflow Management
participants (1)
-
bugzilla_noreply@suse.com