Re: [PATCH] mm: don't rely on system state to detect hot-plug operations

Laurent Dufour <ldufour@xxxxxxxxxxxxx> · Wed, 9 Sep 2020 11:35:23 +0200

Le 09/09/2020 à 10:31, David Hildenbrand a écrit :
On 09.09.20 10:26, Laurent Dufour wrote:
Le 08/09/2020 à 19:40, David Hildenbrand a écrit :
On 08.09.20 19:31, Greg Kroah-Hartman wrote:
On Tue, Sep 08, 2020 at 07:08:35PM +0200, Laurent Dufour wrote:
In register_mem_sect_under_node() the system_state’s value is checked to
detect whether the operation the call is made during boot time or during an
hot-plug operation. Unfortunately, that check is wrong on some
architecture, and may lead to sections being registered under multiple
nodes if node's memory ranges are interleaved.

This can be seen on PowerPC LPAR after multiple memory hot-plug and
hot-unplug operations are done. At the next reboot the node's memory ranges
can be interleaved and since the call to link_mem_sections() is made in
topology_init() while the system is in the SYSTEM_SCHEDULING state, the
node's id is not checked, and the sections registered multiple times. In
that case, the system is able to boot but later hot-plug operation may lead
to this panic because the node's links are correctly broken:

------------[ cut here ]------------
kernel BUG at /Users/laurent/src/linux-ppc/mm/memory_hotplug.c:1084!
Oops: Exception in kernel mode, sig: 5 [#1]
LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
Modules linked in: rpadlpar_io rpaphp pseries_rng rng_core vmx_crypto gf128mul binfmt_misc ip_tables x_tables xfs libcrc32c crc32c_vpmsum autofs4
CPU: 8 PID: 10256 Comm: drmgr Not tainted 5.9.0-rc1+ #25
NIP:  c000000000403f34 LR: c000000000403f2c CTR: 0000000000000000
REGS: c0000004876e3660 TRAP: 0700   Not tainted  (5.9.0-rc1+)
MSR:  800000000282b033 <SF,VEC,VSX,EE,FP,ME,IR,DR,RI,LE>  CR: 24000448  XER: 20040000
CFAR: c000000000846d20 IRQMASK: 0
GPR00: c000000000403f2c c0000004876e38f0 c0000000012f6f00 ffffffffffffffef
GPR04: 0000000000000227 c0000004805ae680 0000000000000000 00000004886f0000
GPR08: 0000000000000226 0000000000000003 0000000000000002 fffffffffffffffd
GPR12: 0000000088000484 c00000001ec96280 0000000000000000 0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000004 0000000000000003
GPR20: c00000047814ffe0 c0000007ffff7c08 0000000000000010 c0000000013332c8
GPR24: 0000000000000000 c0000000011f6cc0 0000000000000000 0000000000000000
GPR28: ffffffffffffffef 0000000000000001 0000000150000000 0000000010000000
NIP [c000000000403f34] add_memory_resource+0x244/0x340
LR [c000000000403f2c] add_memory_resource+0x23c/0x340
Call Trace:
[c0000004876e38f0] [c000000000403f2c] add_memory_resource+0x23c/0x340 (unreliable)
[c0000004876e39c0] [c00000000040408c] __add_memory+0x5c/0xf0
[c0000004876e39f0] [c0000000000e2b94] dlpar_add_lmb+0x1b4/0x500
[c0000004876e3ad0] [c0000000000e3888] dlpar_memory+0x1f8/0xb80
[c0000004876e3b60] [c0000000000dc0d0] handle_dlpar_errorlog+0xc0/0x190
[c0000004876e3bd0] [c0000000000dc398] dlpar_store+0x198/0x4a0
[c0000004876e3c90] [c00000000072e630] kobj_attr_store+0x30/0x50
[c0000004876e3cb0] [c00000000051f954] sysfs_kf_write+0x64/0x90
[c0000004876e3cd0] [c00000000051ee40] kernfs_fop_write+0x1b0/0x290
[c0000004876e3d20] [c000000000438dd8] vfs_write+0xe8/0x290
[c0000004876e3d70] [c0000000004391ac] ksys_write+0xdc/0x130
[c0000004876e3dc0] [c000000000034e40] system_call_exception+0x160/0x270
[c0000004876e3e20] [c00000000000d740] system_call_common+0xf0/0x27c
Instruction dump:
48442e35 60000000 0b030000 3cbe0001 7fa3eb78 7bc48402 38a5fffe 7ca5fa14
78a58402 48442db1 60000000 7c7c1b78 <0b030000> 7f23cb78 4bda371d 60000000
---[ end trace 562fd6c109cd0fb2 ]---

This patch addresses the root cause by not relying on the system_state
value to detect whether the call is due to a hot-plug operation or not. An
additional parameter is added to link_mem_sections() to tell the context of
the call and this parameter is propagated to register_mem_sect_under_node()
throuugh the walk_memory_blocks()'s call.

Fixes: 4fbce633910e ("mm/memory_hotplug.c: make register_mem_sect_under_node() a callback of walk_memory_range()")
Signed-off-by: Laurent Dufour <ldufour@xxxxxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Cc: "Rafael J. Wysocki" <rafael@xxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
   drivers/base/node.c  | 20 +++++++++++++++-----
   include/linux/node.h |  6 +++---
   mm/memory_hotplug.c  |  3 ++-
   3 files changed, 20 insertions(+), 9 deletions(-)

diff --git a/drivers/base/node.c b/drivers/base/node.c
index 508b80f6329b..27f828eeb531 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -762,14 +762,19 @@ static int __ref get_nid_for_pfn(unsigned long pfn)
   }
   
   /* register memory section under specified node if it spans that node */
+struct rmsun_args {
+	int	nid;
+	bool	hotadd;
+};
   static int register_mem_sect_under_node(struct memory_block *mem_blk,
-					 void *arg)
+					void *args)
   {
   	unsigned long memory_block_pfns = memory_block_size_bytes() / PAGE_SIZE;
   	unsigned long start_pfn = section_nr_to_pfn(mem_blk->start_section_nr);
   	unsigned long end_pfn = start_pfn + memory_block_pfns - 1;
-	int ret, nid = *(int *)arg;
+	int ret, nid = ((struct rmsun_args *)args)->nid;
   	unsigned long pfn;
+	bool hotadd = ((struct rmsun_args *)args)->hotadd;
   
   	for (pfn = start_pfn; pfn <= end_pfn; pfn++) {
   		int page_nid;
@@ -789,7 +794,7 @@ static int register_mem_sect_under_node(struct memory_block *mem_blk,
   		 * case, during hotplug we know that all pages in the memory
   		 * block belong to the same node.
   		 */
-		if (system_state == SYSTEM_BOOTING) {
+		if (!hotadd) {
   			page_nid = get_nid_for_pfn(pfn);
   			if (page_nid < 0)
   				continue;
@@ -832,10 +837,15 @@ void unregister_memory_block_under_nodes(struct memory_block *mem_blk)
   			  kobject_name(&node_devices[mem_blk->nid]->dev.kobj));
   }
   
-int link_mem_sections(int nid, unsigned long start_pfn, unsigned long end_pfn)
+int link_mem_sections(int nid, unsigned long start_pfn, unsigned long end_pfn,
+		      bool hotadd)
   {
+	struct rmsun_args args;
+
+	args.nid = nid;
+	args.hotadd = hotadd;
   	return walk_memory_blocks(PFN_PHYS(start_pfn),
-				  PFN_PHYS(end_pfn - start_pfn), (void *)&nid,
+				  PFN_PHYS(end_pfn - start_pfn), (void *)&args,
   				  register_mem_sect_under_node);
   }
   
diff --git a/include/linux/node.h b/include/linux/node.h
index 4866f32a02d8..6df9a4548650 100644
--- a/include/linux/node.h
+++ b/include/linux/node.h
@@ -100,10 +100,10 @@ typedef  void (*node_registration_func_t)(struct node *);
   
   #if defined(CONFIG_MEMORY_HOTPLUG_SPARSE) && defined(CONFIG_NUMA)
   extern int link_mem_sections(int nid, unsigned long start_pfn,
-			     unsigned long end_pfn);
+			     unsigned long end_pfn, bool hotadd);
   #else
   static inline int link_mem_sections(int nid, unsigned long start_pfn,
-				    unsigned long end_pfn)
+				    unsigned long end_pfn, bool hotadd)

Adding a random boolean flag to a function is a horrible way to do
anything.

Now you need to look up what that flag means every time you run across a
caller, breaking your reading of what is happening.

Make this two different functions please, that describe what they do,
and have them call a common "helper function" that does the work with
the flag if you really want to do this type of thing.

link_mem_sections() and link_mem_sections_hotadd()?

But not this way, please no.

For memmap_init_zone() we solved it via an enum with MEMMAP_HOTPLUG vs
MEMMAP_EARLY. Maybe we can generalize, because it tries to tackle
roughly the same thing.

Thanks Dacid, I like this idea.

However, I think I'll not reuse the memmap_context enum, but introduce a new one
specific to the link operation, to be more explicit, something like:

/*
   * When a hotplug operation is done, all pages in the memory block belong to the
   * same node, so there is no need to do such a check in that case.
   */
enum linkmem_context {
	LINKMEM_NO_CHECK_NODE_ID,
	LINKMEM_CHECK_NODE_ID,
}

I'm bad at naming so feel free to argue.


"context" does not really fit the two cases that rather tell you what to
do (like a single flag).

I agree "linkmem_option" might be better, isn't it?

I would have renamed "enum memmap_context" to something like "enum
mp_context" ("memory plug") and used

MP_CONTEXT_EARLY / MP_CONTEXT_HOTPLUG

Instead of using fairly specific "LINKMEM_*_CHECK_NODE_ID" ...

My point was to explicitly show by the enum name, why there is a different 
processing done. Showing that we are at boot or hot-plug time is not telling the 
caller that the node id check will be by-passed and that the assumption is made 
that all pages belong to the same node.
I think it may be good to be explicit about that.

I am, also bad at naming, so ... :)

So I'm arguing, but ... ;)








