Current-Users archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: go vs. NetBSD ELF
On Nov 1, 11:54pm, kre%munnari.OZ.AU@localhost (Robert Elz) wrote:
-- Subject: Re: go vs. NetBSD ELF
| Is it really rational to keep adding cases for every new note that
| gets discovered (to ignore them), just so that the kernel can print a
| message (and otherwise ignore) notes that it hasn't yet been told
| about ?
It is just for DIAGNOSTIC; production kernels don't print anything.
Now if we want DIAGNOSTIC to print less stuff we can change it to
DEBUG_ELF_NOTES or something.
| Does the message really accomplish anything, other than annoying people?
It makes sure that we don't break our own notes and notifies about new
ones.
| For what it's worth, I see that message from a binary, which readelf
| decodes as ...
|
| andromeda$ readelf -n ~/bin/cv
|
| Notes at offset 0x0000010c with length 0x00000034:
| Owner Data size Description
| NetBSD 0x00000004 IDENT 199905 (0.0T)
| NetBSD 0x00000007 Unknown note type: (0x00000002)
|
| Notes at offset 0x000010ac with length 0x00000050:
| Owner Data size Description
| 01.01 0x00000000 NT_VERSION (version)
| 01.01 0x00000000 NT_VERSION (version)
| 01.01 0x00000000 NT_VERSION (version)
| 01.01 0x00000000 NT_VERSION (version)
|
| (I'm pretty that was compiled on at least 1.3, maybe even 1.6 NetBSD,
| I have no idea what 0.0T was!)
|
| Every time I run it, I get on the console (and in messages) ...
|
| /home/kre/bin/cv: Unknown elf note type 1 (NetBSD tag): [namesz=8, descsz=0 name=01.01 ]
|
| The program (continues to) work fine, including running this old 32 bit i386
| binary on an amd64 system, with no particular 32 bit support (other
| than /emul/linux32) added.
|
| andromeda$ file ~/bin/cv
| /home/kre/bin/cv: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for NetBSD, not stripped
Now I know what it is I can fix it.
christos
Home |
Main Index |
Thread Index |
Old Index