> > I'm not sure Yinghai's revert patch is completely equivalent to
> > a revert of the original problematic commit, by a side-by-side
> > comparison of the original commit with his recent revert patch,
> > but then I don't really know that code at all.
> >
> > In the original code there was a section (in e820_reserve_resources()):
> >
> > #ifdef CONFIG_KEXEC
> > if (crashk_res.start != crashk_res.end)
> > request_resource(res, &crashk_res);
> > #endif
> >
> > If you don't have CONFIG_KEXEC defined in your .config, which is
> > probably the case, then you would never request a crashk_res resource.
> > But in the code after the original commit, it unconditionally calls
> > (in reserve_crashkernel()):
> >
> > crashk_res.start = crash_base;
> > crashk_res.end = crash_base + crash_size - 1;
> > insert_resource(&iomem_resource, &crashk_res);
> >
> > And after Yinghai's revert patch it still does (in reserve_crashkernel()):
> >
> > crashk_res.start = crash_base;
> > crashk_res.end = crash_base + crash_size - 1;
> > crashk_res_ptr = &crashk_res;
> >
> > and (in setup_arch()):
> >
> > num_res = 3;
> > if (crashk_res_ptr) {
> > res_kernel[num_res] = crashk_res_ptr;
> > num_res++;
> > }
> > e820_reserve_resources(res_kernel, num_res);
> >
> > then (in e820_reserve_resources()):
> >
> > for (j = 0; j < nr_res_k; j++) {
> > if (!res_kernel[j])
> > continue;
> > request_resource(res, res_kernel[j]);
> > }
> >
> > which for j == 3 is:
> >
> > request_resource(res, &crashk_res);
> >
> > Now it would appear that the new:
> >
> > insert_resource(&iomem_resource, &crashk_res);
> >
> > or new:
> >
> > request_resource(res, &crashk_res);
> >
> > should be noops. But if for any reason crash_size is not zero,
> > then there could be a difference. I have no idea if this is at all
> > significant, but I thought I'd point it out just in case.
>
> why oops ?
I think he meant no-op's.
> if not valid crash kernel size etc is input, crashk_res_ptr will be null
>
> > if (crashk_res_ptr) {
> > res_kernel[num_res] = crashk_res_ptr;
> > num_res++;
> > }
>
> it that is not appended to res_kernel...
So your patch code protects against problem that Bill is mentioning
without using "#ifdef CONFIG_KEXEC", right Yinghai?
For the record: my configs, including the kernel I built with Yinghai's
revert patch, have CONFIG_KEXEC not set.
Some experiments I did last night may render these questions moot, though.
My problem is very specific to the hardware on two of my machines, and it
has something to do with setting up the system resources that
insert_resource() touches.
Dave W.
On Thu, Aug 14, 2008 at 5:03 AM, David Witbrodt <[email protected]> wrote:
>
>
>> > I'm not sure Yinghai's revert patch is completely equivalent to
>> > a revert of the original problematic commit, by a side-by-side
>> > comparison of the original commit with his recent revert patch,
>> > but then I don't really know that code at all.
>> >
>> > In the original code there was a section (in e820_reserve_resources()):
>> >
>> > #ifdef CONFIG_KEXEC
>> > if (crashk_res.start != crashk_res.end)
>> > request_resource(res, &crashk_res);
>> > #endif
>> >
>> > If you don't have CONFIG_KEXEC defined in your .config, which is
>> > probably the case, then you would never request a crashk_res resource.
>> > But in the code after the original commit, it unconditionally calls
>> > (in reserve_crashkernel()):
>> >
>> > crashk_res.start = crash_base;
>> > crashk_res.end = crash_base + crash_size - 1;
>> > insert_resource(&iomem_resource, &crashk_res);
>> >
>> > And after Yinghai's revert patch it still does (in reserve_crashkernel()):
>> >
>> > crashk_res.start = crash_base;
>> > crashk_res.end = crash_base + crash_size - 1;
>> > crashk_res_ptr = &crashk_res;
>> >
>> > and (in setup_arch()):
>> >
>> > num_res = 3;
>> > if (crashk_res_ptr) {
>> > res_kernel[num_res] = crashk_res_ptr;
>> > num_res++;
>> > }
>> > e820_reserve_resources(res_kernel, num_res);
>> >
>> > then (in e820_reserve_resources()):
>> >
>> > for (j = 0; j < nr_res_k; j++) {
>> > if (!res_kernel[j])
>> > continue;
>> > request_resource(res, res_kernel[j]);
>> > }
>> >
>> > which for j == 3 is:
>> >
>> > request_resource(res, &crashk_res);
>> >
>> > Now it would appear that the new:
>> >
>> > insert_resource(&iomem_resource, &crashk_res);
>> >
>> > or new:
>> >
>> > request_resource(res, &crashk_res);
>> >
>> > should be noops. But if for any reason crash_size is not zero,
>> > then there could be a difference. I have no idea if this is at all
>> > significant, but I thought I'd point it out just in case.
>>
>> why oops ?
>
> I think he meant no-op's.
>
>
>> if not valid crash kernel size etc is input, crashk_res_ptr will be null
>>
>> > if (crashk_res_ptr) {
>> > res_kernel[num_res] = crashk_res_ptr;
>> > num_res++;
>> > }
>>
>> it that is not appended to res_kernel...
>
> So your patch code protects against problem that Bill is mentioning
> without using "#ifdef CONFIG_KEXEC", right Yinghai?
can you try enable kexec and kdump in you .config.
it should works. my .config have config_kexec
> Some experiments I did last night may render these questions moot, though.
> My problem is very specific to the hardware on two of my machines, and it
> has something to do with setting up the system resources that
> insert_resource() touches.
please try to bisect on current tree. and every time apply the revert patch...
YH