> after a quick glance i'm happy to conclude that all but one of Andi's
At least the weird in_interrupt_stack() code seems to be still
in there.
I also still see a lot of open coded hex manipulation.
The synchronization code looks as bad as it was before.
The getthread magic pids are also stil
Andi Kleen wrote:
..
You should probably use simple_strtoul() instead of inventing an
own hex parser in kgdb.c. And sprintf instead of an own hex writer.
In general more use sprintf would probably shorten a lot of the parser
code.
..
Speaking of which.. the kernel implementation of snprintf() s
On Mon, Feb 11, 2008 at 10:32:46AM -0500, Frank Ch. Eigler wrote:
> Andi Kleen <[EMAIL PROTECTED]> writes:
>
> > [...] The kgdb code in git-x86#mm is right now is totally broken of
> > course because the CFI annotations in assembler code are gone now,
> > but they are needed for gdb use. [...]
>
Andi Kleen <[EMAIL PROTECTED]> writes:
> [...] The kgdb code in git-x86#mm is right now is totally broken of
> course because the CFI annotations in assembler code are gone now,
> but they are needed for gdb use. [...]
This loss of CFI data is unfortunate for other potential consumers
too, such
I sent Jan & Jason earlier a quick review of the kgdb code currently in
git-x86#mm.
The kgdb code in git-x86#mm is right now is totally broken of course because
the CFI
annotations in assembler code are gone now, but they are needed for gdb use.
And the bogus fault handling code is still in th
5 matches
Mail list logo
