Re: [PATCH v2 2/2] memblock: do not start bottom-up allocations with kernel_end

Mike Rapoport [off-list ref] writes:

quoted

Signed-off-by: Roman Gushchin <redacted>

Reviewed-by: Mike Rapoport <redacted>

I've seen a couple of spurious triggers of the WARN_ONCE() removed by this
patch. This happens on some ppc64le bare metal (powernv) server machines with
CONFIG_SWIOTLB=y and crashkernel=4G, as described in a candidate patch I posted
to solve this issue in a different way:

https://lore.kernel.org/linuxppc-dev/20201218062103.76102-1-bauerman@linux.ibm.com/

Since this patch solves that problem, is it possible to include it in the next
feasible v5.11-rcX, with the following tag?

Fixes: 8fabc623238e ("powerpc: Ensure that swiotlb buffer is allocated from low memory")

This is because reverting the commit above also solves the problem on the
machines where I've seen this issue.

-- 
Thiago Jung Bauermann
IBM Linux Technology Center

On Fri, 22 Jan 2021 01:37:14 -0300 Thiago Jung Bauermann [off-list ref] wrote:

Mike Rapoport [off-list ref] writes:

quoted

quoted

Signed-off-by: Roman Gushchin <redacted>

Reviewed-by: Mike Rapoport <redacted>

I've seen a couple of spurious triggers of the WARN_ONCE() removed by this
patch. This happens on some ppc64le bare metal (powernv) server machines with
CONFIG_SWIOTLB=y and crashkernel=4G, as described in a candidate patch I posted
to solve this issue in a different way:

https://lore.kernel.org/linuxppc-dev/20201218062103.76102-1-bauerman@linux.ibm.com/

Since this patch solves that problem, is it possible to include it in the next
feasible v5.11-rcX, with the following tag?

We could do this, if we're confident that this patch doesn't depend on
[1/2] "mm: cma: allocate cma areas bottom-up"?  I think it is...

Fixes: 8fabc623238e ("powerpc: Ensure that swiotlb buffer is allocated from low memory")

I added that.

On Sat, Jan 23, 2021 at 06:09:11PM -0800, Andrew Morton wrote:

On Fri, 22 Jan 2021 01:37:14 -0300 Thiago Jung Bauermann [off-list ref] wrote:

quoted

Mike Rapoport [off-list ref] writes:

quoted

quoted

Signed-off-by: Roman Gushchin <redacted>

Reviewed-by: Mike Rapoport <redacted>

I've seen a couple of spurious triggers of the WARN_ONCE() removed by this
patch. This happens on some ppc64le bare metal (powernv) server machines with
CONFIG_SWIOTLB=y and crashkernel=4G, as described in a candidate patch I posted
to solve this issue in a different way:

https://lore.kernel.org/linuxppc-dev/20201218062103.76102-1-bauerman@linux.ibm.com/

Since this patch solves that problem, is it possible to include it in the next
feasible v5.11-rcX, with the following tag?

We could do this, if we're confident that this patch doesn't depend on
[1/2] "mm: cma: allocate cma areas bottom-up"?  I think it is...

A think it does not depend on cma bottom-up allocation, it's rather the other
way around: without this CMA bottom-up allocation could fail with KASLR
enabled.

Still, this patch may need updates to the way x86 does early reservations:

https://lore.kernel.org/lkml/20210115083255.12744-1-rppt@kernel.org
 

quoted

Fixes: 8fabc623238e ("powerpc: Ensure that swiotlb buffer is allocated from low memory")

I added that.

-- 
Sincerely yours,
Mike.

Mike Rapoport [off-list ref] writes:

On Sat, Jan 23, 2021 at 06:09:11PM -0800, Andrew Morton wrote:

quoted

On Fri, 22 Jan 2021 01:37:14 -0300 Thiago Jung Bauermann [off-list ref] wrote:

quoted

Mike Rapoport [off-list ref] writes:

quoted

quoted

Signed-off-by: Roman Gushchin <redacted>

Reviewed-by: Mike Rapoport <redacted>

I've seen a couple of spurious triggers of the WARN_ONCE() removed by this
patch. This happens on some ppc64le bare metal (powernv) server machines with
CONFIG_SWIOTLB=y and crashkernel=4G, as described in a candidate patch I posted
to solve this issue in a different way:

https://lore.kernel.org/linuxppc-dev/20201218062103.76102-1-bauerman@linux.ibm.com/

Since this patch solves that problem, is it possible to include it in the next
feasible v5.11-rcX, with the following tag?

We could do this,

Thanks!

quoted

if we're confident that this patch doesn't depend on
[1/2] "mm: cma: allocate cma areas bottom-up"?  I think it is...

A think it does not depend on cma bottom-up allocation, it's rather the other
way around: without this CMA bottom-up allocation could fail with KASLR
enabled.

I agree. Conceptually, this could have been patch 1 in this series.

Still, this patch may need updates to the way x86 does early reservations:

https://lore.kernel.org/lkml/20210115083255.12744-1-rppt@kernel.org

Ah, I wasn't aware of this. Thanks for fixing those issues. That series
seems to be well accepted.

quoted

quoted

Fixes: 8fabc623238e ("powerpc: Ensure that swiotlb buffer is allocated from low memory")

I added that.

Thanks!
-- 
Thiago Jung Bauermann
IBM Linux Technology Center

Mike Rapoport [off-list ref] writes:

On Sat, Jan 23, 2021 at 06:09:11PM -0800, Andrew Morton wrote:

quoted

On Fri, 22 Jan 2021 01:37:14 -0300 Thiago Jung Bauermann [off-list ref] wrote:

quoted

Mike Rapoport [off-list ref] writes:

quoted

quoted

Signed-off-by: Roman Gushchin <redacted>

Reviewed-by: Mike Rapoport <redacted>

I've seen a couple of spurious triggers of the WARN_ONCE() removed by this
patch. This happens on some ppc64le bare metal (powernv) server machines with
CONFIG_SWIOTLB=y and crashkernel=4G, as described in a candidate patch I posted
to solve this issue in a different way:

https://lore.kernel.org/linuxppc-dev/20201218062103.76102-1-bauerman@linux.ibm.com/

Since this patch solves that problem, is it possible to include it in the next
feasible v5.11-rcX, with the following tag?

We could do this, if we're confident that this patch doesn't depend on
[1/2] "mm: cma: allocate cma areas bottom-up"?  I think it is...

A think it does not depend on cma bottom-up allocation, it's rather the other
way around: without this CMA bottom-up allocation could fail with KASLR
enabled.

I noticed that this patch is now upstream as:

2dcb39645441 memblock: do not start bottom-up allocations with kernel_end

Still, this patch may need updates to the way x86 does early reservations:

https://lore.kernel.org/lkml/20210115083255.12744-1-rppt@kernel.org

... but the patches from this link still aren't. Isn't this a potential
problem for x86?

The patch series on the link above is now superseded by v2:

https://lore.kernel.org/linux-mm/20210128105711.10428-1-rppt@kernel.org/

-- 
Thiago Jung Bauermann
IBM Linux Technology Center