kernel/arch/powerpc/platforms/pseries/hotplug-cpu.c, branch linux-5.1.y

powerpc/pseries: Perform full re-add of CPU for topology update post-migration

2019-01-30T12:28:56Z

On pseries systems, performing a partition migration can result in altering the nodes a CPU is assigned to on the destination system. For exampl, pre-migration on the source system CPUs are in node 1 and 3, post-migration on the destination system CPUs are in nodes 2 and 3. Handling the node change for a CPU can cause corruption in the slab cache if we hit a timing where a CPUs node is changed while cache_reap() is invoked. The corruption occurs because the slab cache code appears to rely on the CPU and slab cache pages being on the same node. The current dynamic updating of a CPUs node done in arch/powerpc/mm/numa.c does not prevent us from hitting this scenario. Changing the device tree property update notification handler that recognizes an affinity change for a CPU to do a full DLPAR remove and add of the CPU instead of dynamically changing its node resolves this issue. Signed-off-by: Nathan Fontenot Signed-off-by: Michael W. Bringmann Tested-by: Michael W. Bringmann Signed-off-by: Michael Ellerman

powerpc: Convert to using %pOFn instead of device_node.name

2018-10-03T05:40:01Z

In preparation to remove the node name pointer from struct device_node, convert printf users to use the %pOFn format specifier. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Michael Ellerman Cc: Arnd Bergmann Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Rob Herring Signed-off-by: Michael Ellerman

Merge branch 'topic/paca' into next

2018-03-30T22:09:36Z

Bring in yet another series that touches KVM code, and might need to be merged into the kvm-ppc branch to resolve conflicts. This required some changes in pnv_power9_force_smt4_catch/release() due to the paca array becomming an array of pointers.

powerpc/64: Use array of paca pointers and allocate pacas individually

2018-03-30T12:34:23Z

Change the paca array into an array of pointers to pacas. Allocate pacas individually. This allows flexibility in where the PACAs are allocated. Future work will allocate them node-local. Platforms that don't have address limits on PACAs would be able to defer PACA allocations until later in boot rather than allocate all possible ones up-front then freeing unused. This is slightly more overhead (one additional indirection) for cross CPU paca references, but those aren't too common. Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman

powerpc/pseries: Fix build break for SPLPAR=n and CPU hotplug

2018-02-13T11:38:39Z

Commit e67e02a544e9 ("powerpc/pseries: Fix cpu hotplug crash with memoryless nodes") adds an unconditional call to find_and_online_cpu_nid(), which is only declared if CONFIG_PPC_SPLPAR is enabled. This results in the following build error if this is not the case. arch/powerpc/platforms/pseries/hotplug-cpu.o: In function `dlpar_online_cpu': arch/powerpc/platforms/pseries/hotplug-cpu.c:369: undefined reference to `.find_and_online_cpu_nid' Follow the guideline provided by similar functions and provide a dummy function if CONFIG_PPC_SPLPAR is not enabled. This also moves the external function declaration into an include file where it should be. Fixes: e67e02a544e9 ("powerpc/pseries: Fix cpu hotplug crash with memoryless nodes") Signed-off-by: Guenter Roeck [mpe: Change subject to emphasise the build fix] Signed-off-by: Michael Ellerman

powerpc/numa: Invalidate numa_cpu_lookup_table on cpu remove

2018-02-08T12:56:10Z

When DLPAR removing a CPU, the unmapping of the cpu from a node in unmap_cpu_from_node() should also invalidate the CPUs entry in the numa_cpu_lookup_table. There is not a guarantee that on a subsequent DLPAR add of the CPU the associativity will be the same and thus could be in a different node. Invalidating the entry in the numa_cpu_lookup_table causes the associativity to be read from the device tree at the time of the add. The current behavior of not invalidating the CPUs entry in the numa_cpu_lookup_table can result in scenarios where the the topology layout of CPUs in the partition does not match the device tree or the topology reported by the HMC. This bug looks like it was introduced in 2004 in the commit titled "ppc64: cpu hotplug notifier for numa", which is 6b15e4e87e32 in the linux-fullhist tree. Hence tag it for all stable releases. Cc: stable@vger.kernel.org Signed-off-by: Nathan Fontenot Reviewed-by: Tyrel Datwyler Signed-off-by: Michael Ellerman

powerpc/pseries: Fix cpu hotplug crash with memoryless nodes

2018-01-27T09:59:45Z

On powerpc systems with shared configurations of CPUs and memory and memoryless nodes at boot, an event ordering problem was observed on a SLES12 build platforms with the hot-add of CPUs to the memoryless nodes. * The most common error occurred when the memory SLAB driver attempted to reference the memoryless node to which a CPU was being added before the kernel had finished initializing all of the data structures for the CPU and exited 'device_online' under DLPAR/hot-add. Normally the memoryless node would be initialized through the call path device_online ... arch_update_cpu_topology ... find_cpu_nid ... try_online_node. This patch ensures that the powerpc node will be initialized as early as possible, even if it was memoryless and CPU-less at the point when we are trying to hot-add a new CPU to it. Signed-off-by: Michael Bringmann Reviewed-by: Nathan Fontenot Signed-off-by: Michael Ellerman

powerpc/hotplug: Improve responsiveness of hotplug change

2017-10-16T12:12:04Z

powerpc/hotplug: On Power systems with shared configurations of CPUs and memory, there are some issues with the association of additional CPUs and memory to nodes when hot-adding resources. During hotplug CPU operations, this patch resets the timer on topology update work function to a small value to better ensure that the CPU topology is detected and configured sooner. Signed-off-by: Michael Bringmann Signed-off-by: Michael Ellerman

powerpc/pseries: Fix "OF: ERROR: Bad of_node_put() on /cpus" during DLPAR

2017-09-21T09:30:53Z

Commit 215ee763f8cb ("powerpc: pseries: remove dlpar_attach_node dependency on full path") reworked dlpar_attach_node() to no longer look up the parent node "/cpus", but instead to have the parent node passed by the caller in the function parameter list. As a result dlpar_attach_node() is no longer responsible for freeing the reference to the parent node. However, commit 215ee763f8cb failed to remove the of_node_put(parent) call in dlpar_attach_node(), or to take into account that the reference to the parent in the caller dlpar_cpu_add() needs to be held until after dlpar_attach_node() returns. As a result doing repeated cpu add/remove dlpar operations will eventually result in the following error: OF: ERROR: Bad of_node_put() on /cpus CPU: 0 PID: 10896 Comm: drmgr Not tainted 4.13.0-autotest #1 Call Trace: dump_stack+0x15c/0x1f8 (unreliable) of_node_release+0x1a4/0x1c0 kobject_put+0x1a8/0x310 kobject_del+0xbc/0xf0 __of_detach_node_sysfs+0x144/0x210 of_detach_node+0xf0/0x180 dlpar_detach_node+0xc4/0x120 dlpar_cpu_remove+0x280/0x560 dlpar_cpu_release+0xbc/0x1b0 arch_cpu_release+0x6c/0xb0 cpu_release_store+0xa0/0x100 dev_attr_store+0x68/0xa0 sysfs_kf_write+0xa8/0xf0 kernfs_fop_write+0x2cc/0x400 __vfs_write+0x5c/0x340 vfs_write+0x1a8/0x3d0 SyS_write+0xa8/0x1a0 system_call+0x58/0x6c Fix the issue by removing the of_node_put(parent) call from dlpar_attach_node(), and ensuring that the reference to the parent node is properly held and released by the caller dlpar_cpu_add(). Fixes: 215ee763f8cb ("powerpc: pseries: remove dlpar_attach_node dependency on full path") Signed-off-by: Tyrel Datwyler Reported-by: Abdul Haleem [mpe: Add a comment in the code and frob the change log slightly] Signed-off-by: Michael Ellerman

powerpc/xive: guest exploitation of the XIVE interrupt controller

2017-09-02T11:02:35Z

This is the framework for using XIVE in a PowerVM guest. The support is very similar to the native one in a much simpler form. Each source is associated with an Event State Buffer (ESB). This is a two bit state machine which is used to trigger events. The bits are named "P" (pending) and "Q" (queued) and can be controlled by MMIO. The Guest OS registers event (or notifications) queues on which the HW will post event data for a target to notify. Instead of OPAL calls, a set of Hypervisors call are used to configure the interrupt sources and the event/notification queues of the guest: - H_INT_GET_SOURCE_INFO used to obtain the address of the MMIO page of the Event State Buffer (PQ bits) entry associated with the source. - H_INT_SET_SOURCE_CONFIG assigns a source to a "target". - H_INT_GET_SOURCE_CONFIG determines to which "target" and "priority" is assigned to a source - H_INT_GET_QUEUE_INFO returns the address of the notification management page associated with the specified "target" and "priority". - H_INT_SET_QUEUE_CONFIG sets or resets the event queue for a given "target" and "priority". It is also used to set the notification config associated with the queue, only unconditional notification for the moment. Reset is performed with a queue size of 0 and queueing is disabled in that case. - H_INT_GET_QUEUE_CONFIG returns the queue settings for a given "target" and "priority". - H_INT_RESET resets all of the partition's interrupt exploitation structures to their initial state, losing all configuration set via the hcalls H_INT_SET_SOURCE_CONFIG and H_INT_SET_QUEUE_CONFIG. - H_INT_SYNC issue a synchronisation on a source to make sure sure all notifications have reached their queue. As for XICS, the XIVE interface for the guest is described in the device tree under the "interrupt-controller" node. A couple of new properties are specific to XIVE : - "reg" contains the base address and size of the thread interrupt managnement areas (TIMA), also called rings, for the User level and for the Guest OS level. Only the Guest OS level is taken into account today. - "ibm,xive-eq-sizes" the size of the event queues. One cell per size supported, contains log2 of size, in ascending order. - "ibm,xive-lisn-ranges" the interrupt numbers ranges assigned to the guest. These are allocated using a simple bitmap. and also : - "/ibm,plat-res-int-priorities" contains a list of priorities that the hypervisor has reserved for its own use. Tested with a QEMU XIVE model for pseries and with the Power hypervisor. Signed-off-by: Cédric Le Goater Acked-by: Benjamin Herrenschmidt Signed-off-by: Michael Ellerman