RE: [PATCH 1/2] x86: Add Isolated Memory Regions for Quark X1000

["Ong, Boon Leong" <boon.leong.ong@intel.com>]

Hi Bryan,

These two patches requires another patch "x86: Re-enable IO-APIC for non-SMP X86_32" 
to be included to make sure we have the right build for Quark.

I tested v6 on Galileo Gen v2 just now. I am happy with it.
Many thanks.

Below are the logs:

root@quark:/proc# dmesg | grep imr
[    3.727884] imr: protecting kernel .text - .rodata: 9240 KiB (c1000000 - c1906000)
[    3.736421] imr_selftest: self_test pass zero sized IMR
[    3.742413] imr_selftest: self_test pass overlapped IMR @ (0xc1000000 - 0xc1906000)
[    3.751164] imr_selftest: self_test pass overlapped IMR @ (0xc1905c00 - 0xc220bc00)
[    3.759916] imr_selftest: self_test pass overlapped IMR @ (0xc0fff400 - 0xc1905400)
[    3.768566] imr_selftest: self_test pass 1KiB IMR @ 0x00000000 - access-all
[    3.776542] imr_selftest: self_test pass 1KiB IMR @ 0x00000000 - cpu access
[    3.784442] imr_selftest: self_test pass index-teardown - cpu access
[    3.791690] imr_selftest: self_test pass 2KiB IMR @ 0x00000000
[    3.798396] imr_selftest: self_test pass index-teardown 2KiB
[    3.804957] imr_selftest: self_test pass 2KiB IMR @ 0x00000000
[    3.811616] imr_selftest: self_test pass addr-teardown 2KiB
root@quark:/proc# cat /sys/kernel/debug/imr_state
imr00: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr01: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr02: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr03: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr04: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr05: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr06: base=0x00000000, end=0x00000000, size=0x00000000 rmask=0xbfffffff, wmask=0xffffffff, disabled, unlocked
imr07: base=0x01000000, end=0x01905fff, size=0x00905fff rmask=0x00000001, wmask=0x00000001, enabled , locked

 

>-----Original Message-----
>From: Ong, Boon Leong
>Sent: Thursday, January 29, 2015 9:47 PM
>To: 'Bryan O'Donoghue'
>Cc: Thomas Gleixner; Ingo Molnar; H. Peter Anvin; x86@kernel.org;
>dvhart@infradead.org; linux-kernel@vger.kernel.org; Andy Shevchenko; Ingo
>Molnar
>Subject: RE: [PATCH 1/2] x86: Add Isolated Memory Regions for Quark X1000
>
>Bryan,
>
>Once you have the next revision ready, I would like to test it on my end across
>both Galileo Gen v1 & v2.
>
>Cheers,
>BL
>
>>-----Original Message-----
>>From: Bryan O'Donoghue [mailto:pure.logic@nexus-software.ie]
>>Sent: Thursday, January 29, 2015 6:12 PM
>>To: Andy Shevchenko; Ingo Molnar
>>Cc: Thomas Gleixner; Ingo Molnar; H. Peter Anvin; x86@kernel.org;
>>dvhart@infradead.org; Ong, Boon Leong; linux-kernel@vger.kernel.org
>>Subject: Re: [PATCH 1/2] x86: Add Isolated Memory Regions for Quark
>>X1000
>>
>>On 29/01/15 10:08, Andy Shevchenko wrote:
>>> On Thu, Jan 29, 2015 at 9:44 AM, Ingo Molnar <mingo@kernel.org> wrote:
>>>> * Bryan O'Donoghue <pure.logic@nexus-software.ie> wrote:
>>>
>>> []
>>>
>>>> In any case, I don't see any major problems with this code, so if
>>>> it's fixed it could go into v3.20.
>>>
>>> Brian, it would be really nice to have it in 3.20 since we have
>>> several drivers already in upstream or be ready for 3.20.
>>> It would be then first upstream kernel with minimal Quark support.
>>
>>NP lads - I'll get a patch out today