Re: [PATCH 2/7] ppc64el: kernel: config: little-endian powerpc64 options
On Sun, May 25, 2014 at 06:52:27PM -0300, Mauricio Faria de Oliveira wrote:
> - CRASH_DUMP is not enabled on ppc64. I don't know why, so didn't change it.
Do I really need to tell you that you can't come and work on exactly one
thing without looking left and right? If you think you need it and it
is useful, enable it for all users.
> - PPC_64K_PAGES neither. I imagine it might not be the best setting
> for most powerpc64 systems out there (powermacs, I believe), but it
> is a /must/ for POWER servers and systems expected to run ppc64el
> (currently POWER8-based). Smaller page sizes (i.e., 4k) do incur a
> significant performance hit (sorry, I don't know a number); I believe
> it's the main reason the page size is 64k on every distro running
> on POWER servers. And I think it's not an optimal setting for other
> powerpc-based systems (powermacs, others).
Lets look into Kconfig:
| config PPC_64K_PAGES
| bool "64k page size" if 44x || PPC_STD_MMU_64 || PPC_BOOK3E_64
> Perhaps if a new ppc64 flavor is created for holding the PPC_BOOK3S_64
> options, we can put it (and other options depending on it) there, and
> then include it in the ppc64el defines. I'd be happy to work on that.
There are only two types of ppc64 processors supported in linux:
| config PPC_BOOK3S_64
| bool "Server processors"
| config PPC_BOOK3E_64
| bool "Embedded processors"
There is not much choice. The first option applies to all server and
desktop processors. So I would assume BOOK3S_64 is enabled implicit
during build, but I haven't checked.
> >>+CONFIG_VSX=y
> This one should be OK for other powerpc processors, but I understand
> it adds kernel code/overhead, so I chose not to put it in ppc64.
This is a pre-condition for transactional memory.
> >>+CONFIG_NR_CPUS=2048
> This should be OK too, but not sure it incurs noticeable overhead.
It does. x86 only allows > 512 cpus with variable length cpumasks
(CPUMASK_OFFSTACK). Why does powerpc not use this?
What is the current planned cpu density?
> Unfortunately it's not marked as 'depends on' !CPU_LITTLE_ENDIAN yet;
> that would save us from that particular config line.
Can you please send patches for this upstream?
> >>+CONFIG_PPC_POWERNV=y
> >Does not depend on CPU_LITTLE_ENDIAN.
> Yes; this is again the case of 'selects' stuff not available on other
> powerpc processors/platforms (e.g., POWER7 nap, EPAPR boot).
Default "y" and depends on BOOK3S_64, so most likely activated already.
> >>+## choice: Timer frequency
> >>+CONFIG_HZ_100=y
> >>+# CONFIG_HZ_250 is not set
> >>+# CONFIG_HZ_300 is not set
> >>+# CONFIG_HZ_1000 is not set
> >>+## end choice
> >No.
> Your point is the same as for CPU_FREQ_DEFAULT_GOV above (for which I
> asked clarification), or there's something else?
We have set it to a different value in the main config. You have to
explain why this value is not appropriate for this flavour.
Also most architectures now use several types of NO_HZ, so the main
frequency is not really important anymore.
Bastian
--
Vulcans never bluff.
-- Spock, "The Doomsday Machine", stardate 4202.1
Reply to: