On 04/13/16 at 02:47pm, Pratyush Anand wrote:
Hi Dave,
On 13/04/2016:05:04:23 PM, Dave Young wrote:
> Hi, Pratyush
>
> The patchset looks good to me except the quotation in 3/3.
>
> I think we can improve it more like below, but it can also be done later:
> * The original checking files timestamp logic can be another function
> and we can call it in your is_system_modified()
>
> What do you think?
In the function is_system_modified(), we track modification in hardware
resources. So, probably it would be better to have another function
is_files_modified() where all those logic can be moved.
They sounds somehow identical, files should belong to system?
However, I should do another modification in V3. Currently I am setting
force_rebuild variable if system is modified. Meaning of force_rebuild is
something different, so probably it would be better to keep another variable like
system_modified for this purpose.
Yes, it just works, but force_rebuild should mean about only *force* rebuild
for things like below:
1) one specify force_rebuild in /etc/kdump.conf
2) for cases we are not sure, such as one add extra_modules in /etc/kdump.conf
possible reason is we do not know the dependency of extra_modules so that we
are not sure we should check what modules about the timestamps thus just simply
force-rebuild the inird
But for things we can determine, another variable should be even better, but I
am not sure if the code will be complexer.
Thanks
Dave