[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-kernel
Subject:    floating point support for ptrace(2): new patches
From:       Ramón =?ISO-8859-1?Q?Garc=EDa?= =?ISO-8859-1?Q?Fern=E1ndez?=
Date:       1997-08-30 22:08:57
[Download RAW message or body]

I have updated my kernel and gdb patches. They can be found at

jl1.quim.ucm.es:/pub/ramon/

New features:
* Support of core files. Now gdb can read the floating point registers
from cores.
* The fpu is dumped to core files even if the system is emulating
the floating point unit. It is also dumped correctly in SMP
configurations (this did not work as of 2.1.51 or 2.0.30).
* The ptrace(2) inteface now works with the math emulator.
In the previous version the support of the emulator
was completely broken.
* In the 2.1.51 version of the kernel, the patch deletes the declarations
of the variable last_task_used_math in SMP configurations. This variable
is not actually maintained in SMP machines, and therefore it is not
useful. It is better a compile or link time error rather than a run
time error.
* Fixed a bug that crashed gdb when reading any core file.

I could not implement the "info float". GDB has the code for it,
but it is disabled. I cannot understand the reason that
the comments say. If someone wants to help, please read the
file i386v-nat.c

It would be good to implement a SysV style procfs interface for 
debuggers. It is much more efficient.

Ramon

[prev in list] [next in list] [prev in thread] [next in thread]