Add of_node_put() in various paths to drop references once they are no
longer needed.
Signed-off-by: Liang He <windhl@126.com>
[mpe: Rewrite change log]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220620130221.4073228-1-windhl@126.com
of_find_node_by_path() returns a node pointer with
refcount incremented, we should use of_node_put() on it when done.
Add missing of_node_put() to avoid refcount leak.
Fixes: eac1e731b5 ("powerpc/xive: guest exploitation of the XIVE interrupt controller")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220605053225.56125-1-linmq006@gmail.com
kasan detects access beyond the end of the xibm->bitmap allocation:
BUG: KASAN: slab-out-of-bounds in _find_first_zero_bit+0x40/0x140
Read of size 8 at addr c00000001d1d0118 by task swapper/0/1
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.19.0-rc2-00001-g90df023b36dd #28
Call Trace:
[c00000001d98f770] [c0000000012baab8] dump_stack_lvl+0xac/0x108 (unreliable)
[c00000001d98f7b0] [c00000000068faac] print_report+0x37c/0x710
[c00000001d98f880] [c0000000006902c0] kasan_report+0x110/0x354
[c00000001d98f950] [c000000000692324] __asan_load8+0xa4/0xe0
[c00000001d98f970] [c0000000011c6ed0] _find_first_zero_bit+0x40/0x140
[c00000001d98f9b0] [c0000000000dbfbc] xive_spapr_get_ipi+0xcc/0x260
[c00000001d98fa70] [c0000000000d6d28] xive_setup_cpu_ipi+0x1e8/0x450
[c00000001d98fb30] [c000000004032a20] pSeries_smp_probe+0x5c/0x118
[c00000001d98fb60] [c000000004018b44] smp_prepare_cpus+0x944/0x9ac
[c00000001d98fc90] [c000000004009f9c] kernel_init_freeable+0x2d4/0x640
[c00000001d98fd90] [c0000000000131e8] kernel_init+0x28/0x1d0
[c00000001d98fe10] [c00000000000cd54] ret_from_kernel_thread+0x5c/0x64
Allocated by task 0:
kasan_save_stack+0x34/0x70
__kasan_kmalloc+0xb4/0xf0
__kmalloc+0x268/0x540
xive_spapr_init+0x4d0/0x77c
pseries_init_irq+0x40/0x27c
init_IRQ+0x44/0x84
start_kernel+0x2a4/0x538
start_here_common+0x1c/0x20
The buggy address belongs to the object at c00000001d1d0118
which belongs to the cache kmalloc-8 of size 8
The buggy address is located 0 bytes inside of
8-byte region [c00000001d1d0118, c00000001d1d0120)
The buggy address belongs to the physical page:
page:c00c000000074740 refcount:1 mapcount:0 mapping:0000000000000000 index:0xc00000001d1d0558 pfn:0x1d1d
flags: 0x7ffff000000200(slab|node=0|zone=0|lastcpupid=0x7ffff)
raw: 007ffff000000200 c00000001d0003c8 c00000001d0003c8 c00000001d010480
raw: c00000001d1d0558 0000000001e1000a 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
c00000001d1d0000: fc 00 fc fc fc fc fc fc fc fc fc fc fc fc fc fc
c00000001d1d0080: fc fc 00 fc fc fc fc fc fc fc fc fc fc fc fc fc
>c00000001d1d0100: fc fc fc 02 fc fc fc fc fc fc fc fc fc fc fc fc
^
c00000001d1d0180: fc fc fc fc 04 fc fc fc fc fc fc fc fc fc fc fc
c00000001d1d0200: fc fc fc fc fc 04 fc fc fc fc fc fc fc fc fc fc
This happens because the allocation uses the wrong unit (bits) when it
should pass (BITS_TO_LONGS(count) * sizeof(long)) or equivalent. With small
numbers of bits, the allocated object can be smaller than sizeof(long),
which results in invalid accesses.
Use bitmap_zalloc() to allocate and initialize the irq bitmap, paired with
bitmap_free() for consistency.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220623182509.3985625-1-nathanl@linux.ibm.com
of_find_compatible_node() returns a node pointer with refcount
incremented, we should use of_node_put() on it when done.
Add missing of_node_put() to avoid refcount leak.
Fixes: eac1e731b5 ("powerpc/xive: guest exploitation of the XIVE interrupt controller")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220512090535.33397-1-linmq006@gmail.com
The kexec code paths involve code that necessarily run in real mode, as
CPUs are disabled and control is transferred to the new kernel. Disable
address sanitization for the kexec code and the functions called in real
mode on CPUs being disabled.
[paulus@ozlabs.org: combined a few work-in-progress commits of
Daniel's and wrote the commit message.]
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
[mpe: Move pseries_machine_kexec() into kexec.c so setup.c can be instrumented]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/YoTFSQ2TUSEaDdVC@cleo
Various spelling mistakes in comments.
Detected with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220430185654.5855-1-Julia.Lawall@inria.fr
__setup() handlers should return 1 to obsolete_checksetup() in
init/main.c to indicate that the boot option has been handled.
A return of 0 causes the boot option/value to be listed as an Unknown
kernel parameter and added to init's (limited) argument or environment
strings.
Also, error return codes don't mean anything to obsolete_checksetup() --
only non-zero (usually 1) or zero. So return 1 from xive_off() and
xive_store_eoi_cmdline().
Fixes: 243e25112d ("powerpc/xive: Native exploitation of the XIVE interrupt controller")
Fixes: c21ee04f11 ("powerpc/xive: Add a kernel parameter for StoreEOI")
[lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru]
Reported-by: Igor Zhbanov <i.zhbanov@omprussia.ru>:
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220313065936.4363-1-rdunlap@infradead.org
'xive_irq_bitmap_add()' can return -ENOMEM.
In this case, we should free the memory already allocated and return
'false' to the caller.
Also add an error path which undoes the 'tima = ioremap(...)'
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/564998101804886b151235c8a9f93020923bfd2c.1643718324.git.christophe.jaillet@wanadoo.fr
Cédric pointed out that XIVE IPI information exported via sysfs
(debug/powerpc/xive) display empty lines for processors which are
not online.
Switch to using for_each_online_cpu() so that information is
displayed for online-only processors.
Reported-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Sachin Sant <sachinp@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/164146703333.19039.10920919226094771665.sendpatchset@MacBook-Pro.local
Some files functions in 'arch/powerpc/sysdev' are deserving of an `__init`
macro attribute. These functions are only called by other initialization
functions and therefore should inherit the attribute.
Also, change function declarations in header files to include `__init`.
Signed-off-by: Nick Child <nick.child@ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211216220035.605465-6-nick.child@ibm.com
The automatic "save & restore" of interrupt context is a POWER10/XIVE2
feature exploited by KVM under the PowerNV platform. It is not
available under pSeries and the associated toggle should not be
exposed under the XIVE debugfs directory.
Introduce a platform handler for debugfs initialization and move the
'save-restore' entry under the native (PowerNV) backend to fix compile
when !CONFIG_PPC_POWERNV.
Fixes: 1e7684dc4f ("powerpc/xive: Add a debugfs toggle for save-restore")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211201165418.1041842-1-clg@kaod.org
On POWER10, the automatic "save & restore" of interrupt context is
always available. Provide a way to deactivate it for tests or
performance.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-11-clg@kaod.org
StoreEOI is activated by default on platforms supporting the feature
(POWER10) and will be used as soon as firmware advertises its
availability. The kernel parameter provides a way to deactivate its
use. It can be still be reactivated through debugfs.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-10-clg@kaod.org
It can be used to deactivate temporarily StoreEOI for tests or
performance on platforms supporting the feature (POWER10)
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-9-clg@kaod.org
The XIVE driver under Linux uses a single interrupt priority and only
one event queue is configured per CPU. Expose the contents under
a 'xive/eqs/cpuX' debugfs file.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-8-clg@kaod.org
and remove the EQ entries output which is not very useful since only
the next two events of the queue are taken into account. We will
improve the dump of the EQ in the next patches.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-7-clg@kaod.org
and fix some compile issues when !CONFIG_DEBUG_FS.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
[mpe: Add empty stub to fix !CONFIG_DEBUG_FS build]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-5-clg@kaod.org
StoreEOI (the capability to EOI with a store) requires load-after-store
ordering in some cases to be reliable. P10 introduced a new offset for
load operations to enforce correct ordering and the XIVE driver has
the required support since kernel 5.8, commit b1f9be9392
("powerpc/xive: Enforce load-after-store ordering when StoreEOI is active")
Since skiboot v7, StoreEOI support is advertised on P10 with a new flag
on the PowerNV platform. See skiboot commit 4bd7d84afe46 ("xive/p10:
Introduce a new OPAL_XIVE_IRQ_STORE_EOI2 flag"). When detected,
activate the feature.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-4-clg@kaod.org
and extend output of debugfs and xmon with addresses of the ESB
management and trigger pages.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-3-clg@kaod.org
These routines are not on hot code paths and pr_debug() is easier to
activate. Also add a '0x' prefix to hex printed values (HW IRQ number).
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211105102636.1016378-2-clg@kaod.org
Commit 4f86a06e2d ("irqdomain: Make normal and nomap irqdomains
exclusive") introduced an IRQ_DOMAIN_FLAG_NO_MAP flag to isolate the
'nomap' domains still in use under the powerpc arch. With this new
flag, the revmap_tree of the IRQ domain is not used anymore. This
change broke the support of shared LSIs [1] in the XIVE driver because
it was relying on a lookup in the revmap_tree to query previously
mapped interrupts. Linux now creates two distinct IRQ mappings on the
same HW IRQ which can lead to unexpected behavior in the drivers.
The XIVE IRQ domain is not a direct mapping domain and its HW IRQ
interrupt number space is rather large : 1M/socket on POWER9 and
POWER10, change the XIVE driver to use a 'tree' domain type instead.
[1] For instance, a linux KVM guest with virtio-rng and virtio-balloon
devices.
Fixes: 4f86a06e2d ("irqdomain: Make normal and nomap irqdomains exclusive")
Cc: stable@vger.kernel.org # v5.14+
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Greg Kurz <groug@kaod.org>
Acked-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211116134022.420412-1-clg@kaod.org
When an interrupt is passed through, the KVM XIVE device calls the
set_vcpu_affinity() handler which raises the P bit to mask the
interrupt and to catch any in-flight interrupts while routing the
interrupt to the guest.
On the guest side, drivers (like some Intels) can request at probe
time some MSIs and call synchronize_irq() to check that there are no
in flight interrupts. This will call the XIVE get_irqchip_state()
handler which will always return true as the interrupt P bit has been
set on the host side and lock the CPU in an infinite loop.
Fix that by discarding disabled interrupts in get_irqchip_state().
Fixes: da15c03b04 ("powerpc/xive: Implement get_irqchip_state method for XIVE to fix shutdown race")
Cc: stable@vger.kernel.org #v5.4+
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Tested-by: seeteena <s1seetee@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20211011070203.99726-1-clg@kaod.org
Merge our fixes branch into next.
That lets us resolve a conflict in arch/powerpc/sysdev/xive/common.c.
Between cbc06f051c ("powerpc/xive: Do not skip CPU-less nodes when
creating the IPIs"), which moved request_irq() out of xive_init_ipis(),
and 17df41fec5 ("powerpc: use IRQF_NO_DEBUG for IPIs") which added
IRQF_NO_DEBUG to that request_irq() call, which has now moved.
Compiling ppc64le_defconfig with clang-14 shows a modpost warning:
WARNING: modpost: vmlinux.o(.text+0xa74e0): Section mismatch in
reference from the function xive_setup_cpu_ipi() to the function
.init.text:xive_request_ipi()
The function xive_setup_cpu_ipi() references
the function __init xive_request_ipi().
This is often because xive_setup_cpu_ipi lacks a __init
annotation or the annotation of xive_request_ipi is wrong.
xive_request_ipi() is called from xive_setup_cpu_ipi(), which is not
__init, so xive_request_ipi() should not be marked __init. Remove the
attribute so there is no more warning.
Fixes: cbc06f051c ("powerpc/xive: Do not skip CPU-less nodes when creating the IPIs")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210816185711.21563-1-nathan@kernel.org
No functional change in this patch. arch_debugfs_dir is the generic kernel
name declared in linux/debugfs.h for arch-specific debugfs directory.
Architectures like x86/s390 already use the name. Rename powerpc
specific powerpc_debugfs_root to arch_debugfs_dir.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210812132831.233794-2-aneesh.kumar@linux.ibm.com
On PowerVM, CPU-less nodes can be populated with hot-plugged CPUs at
runtime. Today, the IPI is not created for such nodes, and hot-plugged
CPUs use a bogus IPI, which leads to soft lockups.
We can not directly allocate and request the IPI on demand because
bringup_up() is called under the IRQ sparse lock. The alternative is
to allocate the IPIs for all possible nodes at startup and to request
the mapping on demand when the first CPU of a node is brought up.
Fixes: 7dcc37b3ef ("powerpc/xive: Map one IPI interrupt per node")
Cc: stable@vger.kernel.org # v5.13
Reported-by: Geetika Moolchandani <Geetika.Moolchandani1@ibm.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Tested-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210807072057.184698-1-clg@kaod.org
On P10, the feature doing an automatic "save & restore" of a VCPU
interrupt context is set by default in OPAL. When a VP context is
pulled out, the state of the interrupt registers are saved by the XIVE
interrupt controller under the internal NVP structure representing the
VP. This saves a costly store/load in guest entries and exits.
If OPAL advertises the "save & restore" feature in the device tree,
it should also have set the 'H' bit in the CAM line. Check that when
vCPUs are connected to their ICP in KVM before going any further.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210720134209.256133-3-clg@kaod.org
There is no need to use the lockup detector ("noirqdebug") for IPIs.
The ipistorm benchmark measures a ~10% improvement on high systems
when this flag is set.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210719130614.195886-1-clg@kaod.org
The default domain of the PCI/MSIs is not the XIVE domain anymore. To
list the IRQ mappings under XMON and debugfs, query the IRQ data from
the low level XIVE domain.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-32-clg@kaod.org
The MSI domain clears the IRQ with msi_domain_free(), which calls
irq_domain_free_irqs_top(), which clears the handler data. This is a
problem for the XIVE controller since we need to unmap MMIO pages and
free a specific XIVE structure.
The 'msi_free()' handler is called before irq_domain_free_irqs_top()
when the handler data is still available. Use that to clear the XIVE
controller data.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-10-clg@kaod.org
In the early days of XIVE support, commit cffb717ceb ("powerpc/xive:
Ensure active irqd when setting affinity") tried to fix an issue
related to interrupt migration. If the root cause was related to CPU
unplug, it should have been fixed and there is no reason to keep the
irqd_is_started() check. This test is also breaking affinity setting
of MSIs which can set before starting the associated IRQ.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-8-clg@kaod.org
That was a workaround in the XIVE domain because of the lack of MSI
domain. This is now handled.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-7-clg@kaod.org
pr_debug() is easier to activate and it helps to know how the kernel
configures the HW when tweaking the IRQ subsystem.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-5-clg@kaod.org
This adds handlers to allocate/free IRQs in a domain hierarchy. We
could try to use xive_irq_domain_map() in xive_irq_domain_alloc() but
we rely on xive_irq_alloc_data() to set the IRQ handler data and
duplicating the code is simpler.
xive_irq_free_data() needs to be called when IRQ are freed to clear
the MMIO mappings and free the XIVE handler data, xive_irq_data
structure. This is going to be a problem with MSI domains which we
will address later.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-4-clg@kaod.org
This is a smatch warning:
arch/powerpc/sysdev/xive/common.c:1161 xive_request_ipi() warn: unsigned 'xid->irq' is never less than zero.
Fixes: fd6db2892e ("powerpc/xive: Modernize XIVE-IPI domain with an 'alloc' handler")
Cc: stable@vger.kernel.org # v5.13
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701152412.1507612-1-clg@kaod.org
Only a handful of old PPC systems are still using the old 'nomap'
variant of the irqdomain library. Move the associated definitions
behind a configuration option, which will allow us to make some
more radical changes.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Merge misc updates from Andrew Morton:
"A few misc subsystems and some of MM.
175 patches.
Subsystems affected by this patch series: ia64, kbuild, scripts, sh,
ocfs2, kfifo, vfs, kernel/watchdog, and mm (slab-generic, slub,
kmemleak, debug, pagecache, msync, gup, memremap, memcg, pagemap,
mremap, dma, sparsemem, vmalloc, documentation, kasan, initialization,
pagealloc, and memory-failure)"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (175 commits)
mm/memory-failure: unnecessary amount of unmapping
mm/mmzone.h: fix existing kernel-doc comments and link them to core-api
mm: page_alloc: ignore init_on_free=1 for debug_pagealloc=1
net: page_pool: use alloc_pages_bulk in refill code path
net: page_pool: refactor dma_map into own function page_pool_dma_map
SUNRPC: refresh rq_pages using a bulk page allocator
SUNRPC: set rq_page_end differently
mm/page_alloc: inline __rmqueue_pcplist
mm/page_alloc: optimize code layout for __alloc_pages_bulk
mm/page_alloc: add an array-based interface to the bulk page allocator
mm/page_alloc: add a bulk page allocator
mm/page_alloc: rename alloced to allocated
mm/page_alloc: duplicate include linux/vmalloc.h
mm, page_alloc: avoid page_to_pfn() in move_freepages()
mm/Kconfig: remove default DISCONTIGMEM_MANUAL
mm: page_alloc: dump migrate-failed pages
mm/mempolicy: fix mpol_misplaced kernel-doc
mm/mempolicy: rewrite alloc_pages_vma documentation
mm/mempolicy: rewrite alloc_pages documentation
mm/mempolicy: rename alloc_pages_current to alloc_pages
...
The 'chip_id' field of the XIVE CPU structure is used to choose a
target for a source located on the same chip. For that, the XIVE
driver queries the chip identifier from the "ibm,chip-id" property
and compares it to a 'src_chip' field identifying the chip of a
source. This information is only available on the PowerNV platform,
'src_chip' being assigned to XIVE_INVALID_CHIP_ID under pSeries.
The "ibm,chip-id" property is also not available on all platforms. It
was first introduced on PowerNV and later, under QEMU for pSeries/KVM.
However, the property is not part of PAPR and does not exist under
pSeries/PowerVM.
Assign 'chip_id' to XIVE_INVALID_CHIP_ID by default and let the
PowerNV platform override the value with the "ibm,chip-id" property.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210413130352.1183267-1-clg@kaod.org
Instead of calling irq_create_mapping() to map the IPI for a node,
introduce an 'alloc' handler. This is usually an extension to support
hierarchy irq_domains which is not exactly the case for XIVE-IPI
domain. However, we can now use the irq_domain_alloc_irqs() routine
which allocates the IRQ descriptor on the specified node, even better
for cache performance on multi node machines.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-10-clg@kaod.org
ipistorm [*] can be used to benchmark the raw interrupt rate of an
interrupt controller by measuring the number of IPIs a system can
sustain. When applied to the XIVE interrupt controller of POWER9 and
POWER10 systems, a significant drop of the interrupt rate can be
observed when crossing the second node boundary.
This is due to the fact that a single IPI interrupt is used for all
CPUs of the system. The structure is shared and the cache line updates
impact greatly the traffic between nodes and the overall IPI
performance.
As a workaround, the impact can be reduced by deactivating the IRQ
lockup detector ("noirqdebug") which does a lot of accounting in the
Linux IRQ descriptor structure and is responsible for most of the
performance penalty.
As a fix, this proposal allocates an IPI interrupt per node, to be
shared by all CPUs of that node. It solves the scaling issue, the IRQ
lockup detector still has an impact but the XIVE interrupt rate scales
linearly. It also improves the "noirqdebug" case as showed in the
tables below.
* P9 DD2.2 - 2s * 64 threads
"noirqdebug"
Mint/s Mint/s
chips cpus IPI/sys IPI/chip IPI/chip IPI/sys
--------------------------------------------------------------
1 0-15 4.984023 4.875405 4.996536 5.048892
0-31 10.879164 10.544040 10.757632 11.037859
0-47 15.345301 14.688764 14.926520 15.310053
0-63 17.064907 17.066812 17.613416 17.874511
2 0-79 11.768764 21.650749 22.689120 22.566508
0-95 10.616812 26.878789 28.434703 28.320324
0-111 10.151693 31.397803 31.771773 32.388122
0-127 9.948502 33.139336 34.875716 35.224548
* P10 DD1 - 4s (not homogeneous) 352 threads
"noirqdebug"
Mint/s Mint/s
chips cpus IPI/sys IPI/chip IPI/chip IPI/sys
--------------------------------------------------------------
1 0-15 2.409402 2.364108 2.383303 2.395091
0-31 6.028325 6.046075 6.089999 6.073750
0-47 8.655178 8.644531 8.712830 8.724702
0-63 11.629652 11.735953 12.088203 12.055979
0-79 14.392321 14.729959 14.986701 14.973073
0-95 12.604158 13.004034 17.528748 17.568095
2 0-111 9.767753 13.719831 19.968606 20.024218
0-127 6.744566 16.418854 22.898066 22.995110
0-143 6.005699 19.174421 25.425622 25.417541
0-159 5.649719 21.938836 27.952662 28.059603
0-175 5.441410 24.109484 31.133915 31.127996
3 0-191 5.318341 24.405322 33.999221 33.775354
0-207 5.191382 26.449769 36.050161 35.867307
0-223 5.102790 29.356943 39.544135 39.508169
0-239 5.035295 31.933051 42.135075 42.071975
0-255 4.969209 34.477367 44.655395 44.757074
4 0-271 4.907652 35.887016 47.080545 47.318537
0-287 4.839581 38.076137 50.464307 50.636219
0-303 4.786031 40.881319 53.478684 53.310759
0-319 4.743750 43.448424 56.388102 55.973969
0-335 4.709936 45.623532 59.400930 58.926857
0-351 4.681413 45.646151 62.035804 61.830057
[*] https://github.com/antonblanchard/ipistorm
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-9-clg@kaod.org
When under xmon, the "dxi" command dumps the state of the XIVE
interrupts. If an interrupt number is specified, only the state of
the associated XIVE interrupt is dumped. This form of the command
lacks an irq_data parameter which is nevertheless used by
xmon_xive_get_irq_config(), leading to an xmon crash.
Fix that by doing a lookup in the system IRQ mapping to query the IRQ
descriptor data. Invalid interrupt numbers, or not belonging to the
XIVE IRQ domain, OPAL event interrupt number for instance, should be
caught by the previous query done at the firmware level.
Fixes: 97ef275077 ("powerpc/xive: Fix xmon support on the PowerNV platform")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-8-clg@kaod.org
Move the xmon routine under XIVE subsystem and rework the loop on the
interrupts taking into account the xive_irq_domain to filter out IPIs.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-7-clg@kaod.org
When looping on IRQ descriptor, irq_data is always valid.
Fixes: 930914b7d5 ("powerpc/xive: Add a debugfs file to dump internal XIVE state")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-6-clg@kaod.org
Now that the IPI interrupt has its own domain, the checks on the HW
interrupt number XIVE_IPI_HW_IRQ and on the chip can be replaced by a
check on the domain.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-5-clg@kaod.org
The IPI interrupt has its own domain now. Testing the HW interrupt
number is not needed anymore.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210331144514.892250-4-clg@kaod.org
