<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<tt></tt>
<br><tt>- Enabled CONFIG_SPARSEMEM support for ia64 kernels; tested
on</tt>
<br><tt> RHEL5-alpha (2.6.17-1.2519.4.5.el5). Without
this fix, crash</tt>
<br><tt> would fail during initialization with error message
indicating:</tt>
<br><tt> "crash: CONFIG_SPARSEMEM kernels not supported for
this architecture"</tt>
<br><tt> (anderson@redhat.com, dwilder@us.ibm.com)</tt><tt></tt>
<p><tt>- Moved read_in_kernel_config() to just after the internal
gdb</tt>
<br><tt> module gets initialized. Without this fix, xen
kernels built</tt>
<br><tt> with CONFIG_IKCONFIG would fail during initialization
indicating:</tt>
<br><tt> "crash: gdb_interface: gdb not initialized?"</tt>
<br><tt> (anderson@redhat.com, moriwaka@valinux.co.jp)</tt><tt></tt>
<p><tt>- Implemented new s390/s390x command "s390dbf" command to
print out</tt>
<br><tt> kernel traces from the s390 debug feature (s390dbf).
The debug</tt>
<br><tt> feature is an s390 kernel trace API which uses wraparound
buffers</tt>
<br><tt> to store trace records in memory. Many of the
s390 device drivers</tt>
<br><tt> use this feature. There is some documentation
of the s390dbf in</tt>
<br><tt> the kernel sources under /Documentation/s390/s390dbf.txt.</tt>
<br><tt> (holzheu@de.ibm.com)</tt><tt></tt>
<p><tt>- RHEL5-alpha kernel modules (only x86_64 confirmed) may possibly</tt>
<br><tt> fail to be loaded with the "mod" command due to dwarf2
errors</tt>
<br><tt> associated with the the split module.ko/module.ko.debug
debuginfo</tt>
<br><tt> facility used by RHEL kernels. Bugzillas have
been filed to</tt>
<br><tt> address those problems, but the crash utility's error-reporting</tt>
<br><tt> mechanism has beem modified to properly reflect that
the internal</tt>
<br><tt> gdb module has failed to load the kernel module's
debug data.</tt>
<br><tt> Without this fix, the "mod -[sS]" commands would silently
return</tt>
<br><tt> without loading the module data because the "add-symbol-file"</tt>
<br><tt> operation inside the gdb module failed, did a longjmp(),
and ended</tt>
<br><tt> up back at the crash prompt. That behaviour
has been changed</tt>
<br><tt> to report the module name and the gdb error like so:</tt><tt></tt>
<p><tt> crash> mod -S</tt>
<br><tt> mod: /lib/modules/2.6.17-1.2564.1/kernel/drivers/scsi/scsi_mod.ko</tt>
<br><tt> gdb add-symbol-file command failed</tt>
<br><tt> crash></tt><tt></tt>
<p><tt> Note that this problem occurs in all post-RHEL4 kernels,
i.e.,</tt>
<br><tt> FC4, FC5, and now FC6 and RHEL5.</tt>
<br><tt> (anderson@redhat.com)</tt><tt></tt>
<p><tt>- Fix for runaway unkillable "repeat" command output that
can happen</tt>
<br><tt> when scrolling is turned off and the command that
was entered is</tt>
<br><tt> bogus. (anderson@redhat.com)</tt><tt></tt>
<p><tt>- Fix for "struct structure.member address" output when the
member</tt>
<br><tt> is an array; additional members beyond the array contents
would</tt>
<br><tt> get displayed. (anderson@redhat.com)</tt><tt></tt>
<p><tt>- Fix to internal gdb module to properly handle relocatable
kernel</tt>
<br><tt> virtual addresses; this will be required for upcoming
relocatable</tt>
<br><tt> RHEL5 kernels required for the kexec/kdump facility.</tt>
<br><tt> (anderson@redhat.com)</tt><tt></tt>
<p><tt>- Combined kernel_init(PRE_GDB) and kernel_init(POST_GDB)
into a</tt>
<br><tt> single call to kernel_init() that is done after gdb
is initialized;</tt>
<br><tt> verify_version() now called by kernel_init().
This is just a code</tt>
<br><tt> re-work, and does not change any functionality. (anderson@redhat.com)</tt>
<br><tt></tt> <tt></tt>
<p><tt>Download from: <A HREF="http://people.redhat.com/anderson">http://people.redhat.com/anderson</A></tt>
<br><tt></tt> </html>