Hi Christophe,
Ok, so let's summarise from my side.
The issue is in the PowerPC updates 5.13-1. I reverted these and after that the
issue is gone.
We know that only BookE machines are affected. Book3S machines are working with
the PowerPC updates.
I think it’s not directly an Xorg issue. It’s m
On Thursday, May 6, 2021, Chris Packham
wrote:
> The P2040/P2041 has an erratum where the normal i2c recovery mechanism
> does not work. Implement the alternative recovery mechanism documented
> in the P2040 Chip Errata Rev Q.
>
> Signed-off-by: Chris Packham
> ---
> drivers/i2c/busses/i2c-mpc.
Hi,
Le 06/05/2021 à 09:56, Christian Zigotzky a écrit :
Hi Christophe,
Ok, so let's summarise from my side.
The issue is in the PowerPC updates 5.13-1. I reverted these and after that the
issue is gone.
We know that only BookE machines are affected. Book3S machines are working with
the Power
Last user of m8260_gorom() was removed by
Commit 917f0af9e5a9 ("powerpc: Remove arch/ppc and include/asm-ppc")
removed last user of m8260_gorom().
In fact m8260_gorom() was ported to arch/powerpc/ but the
platform using it died with arch/ppc/
Remove it.
Signed-off-by: Christophe Leroy
---
arch
On Wed, May 05, 2021 at 11:34:51PM -0500, Christopher M. Riedl wrote:
> Powerpc allows for multiple CPUs to patch concurrently. When patching
> with STRICT_KERNEL_RWX a single patching_mm is allocated for use by all
> CPUs for the few times that patching occurs. Use a spinlock to protect
> the patc
kuap_save_and_lock() is only for interrupts inside kernel.
system call are only from user, calling kuap_save_and_lock()
is wrong.
Fixes: c16728835eec ("powerpc/32: Manage KUAP in C")
Signed-off-by: Christophe Leroy
---
arch/powerpc/kernel/interrupt.c | 3 ---
1 file changed, 3 deletions(-)
dif
In most cases, kuap_update_sr() will update a single segment
register.
We know that first update will always be done, if there is no
segment register to update at all, kuap_update_sr() is not
called.
Avoid recurring calculations and tests in that case.
Signed-off-by: Christophe Leroy
---
arch/
If neither KUAP nor PPC64 is selected, there is nothing in the second
union of struct pt_regs, so the alignment padding is waste of memory.
Signed-off-by: Christophe Leroy
---
arch/powerpc/include/asm/ptrace.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/inc
Move the PAGE bits into pgtable.h to be more similar to book3s/64.
Signed-off-by: Christophe Leroy
---
arch/powerpc/include/asm/book3s/32/hash.h| 45
arch/powerpc/include/asm/book3s/32/pgtable.h | 38 -
2 files changed, 37 insertions(+), 46 deletions(-)
Paolo Bonzini writes:
> On 05/05/21 14:15, Nicholas Piggin wrote:
>> Commit b1c5356e873c ("KVM: PPC: Convert to the gfn-based MMU notifier
>> callbacks") causes unmap_gfn_range and age_gfn callbacks to only work
>> on the first gfn in the range. It also makes the aging callbacks call
>> into both
Same as kuap_user_restore(), kuep_unlock() has to be called when
really returning to user, that is in interrupt_exit_user_prepare(),
not in interrupt_exit_prepare().
Fixes: b5efec00b671 ("powerpc/32s: Move KUEP locking/unlocking in C")
Signed-off-by: Christophe Leroy
---
arch/powerpc/include/asm
On 06.05.21 17:26, Zi Yan wrote:
From: Zi Yan
Hi all,
This patchset tries to remove the restriction on memory hotplug/hotremove
granularity, which is always greater or equal to memory section size[1].
With the patchset, kernel is able to online/offline memory at a size independent
of memory se
From: Zi Yan
subsection bitmap was set/cleared when a section is added/removed, but
pfn_to_online_page() uses subsection bitmap to check if the page is
online, which is not accurate. It was working when a whole section is
added/removed during memory hotplug and hotremove. When the following
patch
From: Zi Yan
Hi all,
This patchset tries to remove the restriction on memory hotplug/hotremove
granularity, which is always greater or equal to memory section size[1].
With the patchset, kernel is able to online/offline memory at a size independent
of memory section size, as small as 2MB (the su
From: Zi Yan
For x86_64, 32bit vdso is compiled for compatibility reason and 32bit
SECTION_SIZE_BITS value is used during compilation. It causes
compilation time error when MAX_ORDER is increased in the 64bit
environment even if it is OK for 64bit SECTION_SIZE_BITS. Remove the
check during 32bit
From: Zi Yan
When we online subsections and pageblock size is larger than
a subsection, we should be able to change migratetype for partial
pageblocks. We also change the assumption that all pageblocks are in
a whole.
Signed-off-by: Zi Yan
---
include/linux/page-isolation.h | 8 ++--
mm/p
From: Zi Yan
As MAX_ORDER can be increased to larger than hugetlb page orders,
pageblock_order will be limited at HUGETLB_PAGE_ORDER. It is not
ideal for anti-fragmentation for allocating pages from buddy allocator
with MAX_ORDER > HUGETLB_PAGE_ORDER. Make pageblock_order a variable and
set it to
From: Zi Yan
Remove the section size alignment checks for memory hotplug/hotremove,
so that we can online/offline subsection memory.
Signed-off-by: Zi Yan
---
mm/memory_hotplug.c | 16 +---
mm/page_isolation.c | 4
mm/sparse.c | 17 ++---
3 files changed,
From: Zi Yan
This is only used to test onlining/offlining subsection memory in
a x86_64 system by increasing section size to 2GB and pageblock size to
1GB when MAX_ORDER is set to 20.
Signed-off-by: Zi Yan
---
arch/x86/Kconfig | 15 +++
arch/x86/include/asm/sparseme
From: Zi Yan
To enable subsection memory online/offline, we need to remove the
assumption of memory_block size being greater or equal to section size.
The following changes are made:
1. use (start_pfn, nr_pages) pair to specify memory_block size instead of
start_section_nr.
2. calculate memor
On 6 May 2021, at 11:31, David Hildenbrand wrote:
> On 06.05.21 17:26, Zi Yan wrote:
>> From: Zi Yan
>>
>> Hi all,
>>
>> This patchset tries to remove the restriction on memory hotplug/hotremove
>> granularity, which is always greater or equal to memory section size[1].
>> With the patchset, kern
On 06.05.21 17:31, David Hildenbrand wrote:
On 06.05.21 17:26, Zi Yan wrote:
From: Zi Yan
Hi all,
This patchset tries to remove the restriction on memory hotplug/hotremove
granularity, which is always greater or equal to memory section size[1].
With the patchset, kernel is able to online/offl
The last patch increases SECTION_SIZE_BITS to demonstrate the use of memory
hotplug/hotremove subsection, but is not intended to be merged as is. It is
there in case one wants to try this out and will be removed during the final
submission.
Feel free to give suggestions and comments. I am looking
On 6 May 2021, at 11:26, Zi Yan wrote:
> From: Zi Yan
>
> Hi all,
>
> This patchset tries to remove the restriction on memory hotplug/hotremove
> granularity, which is always greater or equal to memory section size[1].
> With the patchset, kernel is able to online/offline memory at a size
> inde
On 6 May 2021, at 11:40, David Hildenbrand wrote:
The last patch increases SECTION_SIZE_BITS to demonstrate the use of memory
hotplug/hotremove subsection, but is not intended to be merged as is. It is
there in case one wants to try this out and will be removed during the
fina
On 06.05.21 17:50, Zi Yan wrote:
On 6 May 2021, at 11:40, David Hildenbrand wrote:
The last patch increases SECTION_SIZE_BITS to demonstrate the use of memory
hotplug/hotremove subsection, but is not intended to be merged as is. It is
there in case one wants to try this out and will be removed
When certain PHB HW failure causes phyp to recover PHB, it marks the PE
state as temporarily unavailable. In this case, per PAPR, rtas call
ibm,read-slot-reset-state2 returns a PE state as temporarily unavailable(5)
and OS has to wait until that recovery is complete. During this state the
slot pres
On 06.05.21 17:26, Zi Yan wrote:
From: Zi Yan
subsection bitmap was set/cleared when a section is added/removed, but
pfn_to_online_page() uses subsection bitmap to check if the page is
online, which is not accurate. It was working when a whole section is
added/removed during memory hotplug and
On 6 May 2021, at 12:28, David Hildenbrand wrote:
> On 06.05.21 17:50, Zi Yan wrote:
>> On 6 May 2021, at 11:40, David Hildenbrand wrote:
>>
>> The last patch increases SECTION_SIZE_BITS to demonstrate the use of
>> memory
>> hotplug/hotremove subsection, but is not intended to be mer
On 6 May 2021, at 13:48, David Hildenbrand wrote:
> On 06.05.21 17:26, Zi Yan wrote:
>> From: Zi Yan
>>
>> subsection bitmap was set/cleared when a section is added/removed, but
>> pfn_to_online_page() uses subsection bitmap to check if the page is
>> online, which is not accurate. It was working
1. Pageblock size
There are a couple of features that rely on the pageblock size to be reasonably small to
work as expected. One example is virtio-balloon free page reporting, then there is
virtio-mem (still also glued MAX_ORDER) and we have CMA (still also glued to MAX_ORDER).
Most probably
But glimpsing at patch #2, I'd rather stop right away digging deeper into this
series :)
What is the issue of patch 2, which makes pageblock_order a variable all the
time? BTW, patch 2 fixes a bug by exporting pageblock_order, since when
HUGETLB_PAGE_SIZE_VARIABLE is set, virtio-mem will not
Dear Linux folks,
On the POWER8 system IBM S822LC, Linux 5.13+, built with USSAN, logs the
warning below.
```
[0.030091]
[0.030295] UBSAN: array-index-out-of-bounds in
arch/powerpc/kernel/legacy_serial.
[corrected subject]
Am 06.05.21 um 21:31 schrieb Paul Menzel:
Dear Linux folks,
On the POWER8 system IBM S822LC, Linux 5.13+, built with USSAN, logs the
warning below.
```
[ 0.030091]
[ 0.030295] UBSA
On Thu, May 06, 2021 at 09:10:52PM +0200, David Hildenbrand wrote:
> I have to admit that I am not really a friend of that. I still think our
> target goal should be to have gigantic THP *in addition to* ordinary THP.
> Use gigantic THP where enabled and possible, and just use ordinary THP
> everyw
On 06.05.21 21:30, Matthew Wilcox wrote:
On Thu, May 06, 2021 at 09:10:52PM +0200, David Hildenbrand wrote:
I have to admit that I am not really a friend of that. I still think our
target goal should be to have gigantic THP *in addition to* ordinary THP.
Use gigantic THP where enabled and possib
I have started bisecting again.
Link: https://forum.hyperion-entertainment.com/viewtopic.php?p=53106#p53106
> On 6. May 2021, at 10:09, Christophe Leroy
> wrote:
>
> Hi,
>
>> Le 06/05/2021 à 09:56, Christian Zigotzky a écrit :
>> Hi Christophe,
>> Ok, so let's summarise from my side.
>> The
On 6/05/21 8:03 pm, Andy Shevchenko wrote:
On Thursday, May 6, 2021, Chris Packham
mailto:chris.pack...@alliedtelesis.co.nz>>
wrote:
The P2040/P2041 has an erratum where the normal i2c recovery mechanism
does not work. Implement the alternative recovery mechanism documented
in the P2040 Chip E
The i2c controllers on the P2040/P2041 have an erratum where the
documented scheme for i2c bus recovery will not work (A-004447). A
different mechanism is needed which is documented in the P2040 Chip
Errata Rev Q (latest available at the time of writing).
Signed-off-by: Chris Packham
---
arch/po
The P2040/P2041 has an erratum where the i2c recovery scheme
documented in the reference manual (and currently implemented
in the i2c-mpc.c driver) does not work. The errata document
provides an alternative that does work. This series implements
that alternative and uses a property in the devicetre
Document the fsl,i2c-erratum-a004447 flag which indicates the presence
of an i2c erratum on some QorIQ SoCs.
Signed-off-by: Chris Packham
---
Documentation/devicetree/bindings/i2c/i2c-mpc.yaml | 7 +++
1 file changed, 7 insertions(+)
diff --git a/Documentation/devicetree/bindings/i2c/i2c-mp
The P2040/P2041 has an erratum where the normal i2c recovery mechanism
does not work. Implement the alternative recovery mechanism documented
in the P2040 Chip Errata Rev Q.
Signed-off-by: Chris Packham
---
Notes:
Changes in v2:
- Use readb_poll_timeout instead of open-coded loop
drive
On Fri, May 7, 2021 at 3:43 AM Mahesh Salgaonkar wrote:
>
> When certain PHB HW failure causes phyp to recover PHB, it marks the PE
> state as temporarily unavailable. In this case, per PAPR, rtas call
> ibm,read-slot-reset-state2 returns a PE state as temporarily unavailable(5)
> and OS has to wa
Hi Christian,
On 4/30/21 06:59, Christian Zigotzky wrote:
> Hello,
>
> The Nemo board (A-EON AmigaOne X1000) [1] and the FSL P5040 Cyrus+ board
> (A-EON AmigaOne X5000) [2] with installed AMD Radeon HD6970 NI graphics cards
> (Cayman
> XT) [3] don't boot with the latest git kernel anymore after
The SW LRU is in an MMU feature section. When not used, that's a
dozen of NOPs to fetch for nothing.
Define an ALT section that does the few remaining operations.
That also avoids a double read on SRR1 in the SW LRU case.
Signed-off-by: Christophe Leroy
---
arch/powerpc/kernel/head_book3s_32.S
Le 06/05/2021 à 04:34, Jordan Niethe a écrit :
From: Russell Currey
Add the arch specific insn page allocator for powerpc. This allocates
ROX pages if STRICT_KERNEL_RWX is enabled. These pages are only written
to with patch_instruction() which is able to write RO pages.
Reviewed-by: Daniel
Le 06/05/2021 à 04:34, Jordan Niethe a écrit :
From: Russell Currey
To enable strict module RWX on powerpc, set:
CONFIG_STRICT_MODULE_RWX=y
You should also have CONFIG_STRICT_KERNEL_RWX=y set to have any real
security benefit.
ARCH_HAS_STRICT_MODULE_RWX is set to require ARCH_HAS_STR
mmu_has_feature(MMU_FTR_TYPE_RADIX) can be evaluated regardless of
CONFIG_PPC_RADIX_MMU.
When CONFIG_PPC_RADIX_MMU is not set, mmu_has_feature(MMU_FTR_TYPE_RADIX)
will evaluate to 'false' at build time because MMU_FTR_TYPE_RADIX
wont be included in MMU_FTRS_POSSIBLE.
Signed-off-by: Christophe Ler
Le 06/05/2021 à 04:34, Jordan Niethe a écrit :
setup_text_poke_area() is a late init call so it runs before
mark_rodata_ro() and after the init calls. This lets all the init code
patching simply write to their locations. In the future, kprobes is
going to allocate its instruction pages RO whic
Le 06/05/2021 à 04:34, Jordan Niethe a écrit :
Commit 74451e66d516 ("bpf: make jited programs visible in traces") added
a default bpf_jit_free() implementation. Powerpc did not use the default
bpf_jit_free() as powerpc did not set the images read-only. The default
bpf_jit_free() called bpf_jit
Le 06/05/2021 à 04:34, Jordan Niethe a écrit :
Add the necessary call to bpf_jit_binary_lock_ro() to remove write and
add exec permissions to the JIT image after it has finished being
written.
Without CONFIG_STRICT_MODULE_RWX the image will be writable and
executable until the call to bpf_jit
Commit 39c8bf2b3cc1 ("powerpc: Retire e200 core (mpc555x processor)")
removed CONFIG_E200.
Commit f9158d58a4e1 ("powerpc/mm: Add mask of always present MMU
features") was merged in the same cycle and added a new use of
CONFIG_E200.
Remove that use.
Fixes: f9158d58a4e1 ("powerpc/mm: Add mask of al
We recently discovered some of our mitigation patching was not safe
against other CPUs running concurrently.
Add a test which enable/disables all mitigations in a tight loop while
also running some stress load. On an unpatched system this almost always
leads to an oops and panic/reboot, but we als
Hi Gustavo,
Great! I will test it. Many thanks for your help.
Cheers,
Christian
> On 7. May 2021, at 01:55, Gustavo A. R. Silva wrote:
>
> Hi Christian,
>
>> On 4/30/21 06:59, Christian Zigotzky wrote:
>> Hello,
>>
>> The Nemo board (A-EON AmigaOne X1000) [1] and the FSL P5040 Cyrus+ board
54 matches
Mail list logo
