xref: /OK3568_Linux_fs/kernel/Documentation/ABI/testing/sysfs-memory-page-offline (revision 4882a59341e53eb6f0b4789bf948001014eff981)
1*4882a593SmuzhiyunWhat:		/sys/devices/system/memory/soft_offline_page
2*4882a593SmuzhiyunDate:		Sep 2009
3*4882a593SmuzhiyunKernelVersion:	2.6.33
4*4882a593SmuzhiyunContact:	andi@firstfloor.org
5*4882a593SmuzhiyunDescription:
6*4882a593Smuzhiyun		Soft-offline the memory page containing the physical address
7*4882a593Smuzhiyun		written into this file. Input is a hex number specifying the
8*4882a593Smuzhiyun		physical address of the page. The kernel will then attempt
9*4882a593Smuzhiyun		to soft-offline it, by moving the contents elsewhere or
10*4882a593Smuzhiyun		dropping it if possible. The kernel will then be placed
11*4882a593Smuzhiyun		on the bad page list and never be reused.
12*4882a593Smuzhiyun
13*4882a593Smuzhiyun		The offlining is done in kernel specific granuality.
14*4882a593Smuzhiyun		Normally it's the base page size of the kernel, but
15*4882a593Smuzhiyun		this might change.
16*4882a593Smuzhiyun
17*4882a593Smuzhiyun		The page must be still accessible, not poisoned. The
18*4882a593Smuzhiyun		kernel will never kill anything for this, but rather
19*4882a593Smuzhiyun		fail the offline.  Return value is the size of the
20*4882a593Smuzhiyun		number, or a error when the offlining failed.  Reading
21*4882a593Smuzhiyun		the file is not allowed.
22*4882a593Smuzhiyun
23*4882a593SmuzhiyunWhat:		/sys/devices/system/memory/hard_offline_page
24*4882a593SmuzhiyunDate:		Sep 2009
25*4882a593SmuzhiyunKernelVersion:	2.6.33
26*4882a593SmuzhiyunContact:	andi@firstfloor.org
27*4882a593SmuzhiyunDescription:
28*4882a593Smuzhiyun		Hard-offline the memory page containing the physical
29*4882a593Smuzhiyun		address written into this file. Input is a hex number
30*4882a593Smuzhiyun		specifying the physical address of the page. The
31*4882a593Smuzhiyun		kernel will then attempt to hard-offline the page, by
32*4882a593Smuzhiyun		trying to drop the page or killing any owner or
33*4882a593Smuzhiyun		triggering IO errors if needed.  Note this may kill
34*4882a593Smuzhiyun		any processes owning the page. The kernel will avoid
35*4882a593Smuzhiyun		to access this page assuming it's poisoned by the
36*4882a593Smuzhiyun		hardware.
37*4882a593Smuzhiyun
38*4882a593Smuzhiyun		The offlining is done in kernel specific granuality.
39*4882a593Smuzhiyun		Normally it's the base page size of the kernel, but
40*4882a593Smuzhiyun		this might change.
41*4882a593Smuzhiyun
42*4882a593Smuzhiyun		Return value is the size of the number, or a error when
43*4882a593Smuzhiyun		the offlining failed.
44*4882a593Smuzhiyun		Reading the file is not allowed.
45