Re: [PATCH v3 3/3] acpi,memory-hotplug : add memory offline code to acpi_memory_device_remove()

On Mon, 2012-10-29 at 06:16 +0000, Wen Congyang wrote:

At 10/27/2012 01:14 AM, Toshi Kani Wrote:

quoted

On Fri, 2012-10-26 at 18:31 +0800, wency@cn.fujitsu.com wrote:

quoted

From: Yasuaki Ishimatsu <redacted>

The memory device can be removed by 2 ways:
1. send eject request by SCI
2. echo 1 >/sys/bus/pci/devices/PNP0C80:XX/eject

In the 1st case, acpi_memory_disable_device() will be called.
In the 2nd case, acpi_memory_device_remove() will be called.

Hi Yasuaki, Wen,

Why do you need to have separate code design & implementation for the
two cases?  In other words, can the 1st case simply use the same code
path of the 2nd case, just like I did for the CPU hot-remove patch
below?  It will simplify the code and make the memory notify handler
more consistent with other handlers.
https://lkml.org/lkml/2012/10/19/456

Yes, the 1st case can simply reuse the same code of the 2nd case.
It is another issue. The memory is not offlined and removed in 2nd
case. This patchset tries to fix this problem. After doing this,
we can merge the codes for the two cases.

But there is some bug in the code for 2nd case:
If offlining memory failed, we don't know such error in 2nd case, and
the kernel will in a dangerous state: the memory device is poweroffed
but the kernel is using it.

We should fix this bug before merging them.

Hi Wen,

Sounds good.  Thanks for the clarification!

-Toshi

Thanks
Wen Congyang

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>