On a 2-socket Power9 system with 32 cores/128 threads (SMT4) and 1TB
of memory running the following guest configs:
guest A:
- 224GB of memory
- 56 VCPUs (sockets=1,cores=28,threads=2), where:
VCPUs 0-1 are pinned to CPUs 0-3,
VCPUs 2-3 are pinned to CPUs 4-7,
...
VCPUs 54-55 are pinned to CPUs 108-111
guest B:
- 4GB of memory
- 4 VCPUs (sockets=1,cores=4,threads=1)
with the following workloads (with KSM and THP enabled in all):
guest A:
stress --cpu 40 --io 20 --vm 20 --vm-bytes 512M
guest B:
stress --cpu 4 --io 4 --vm 4 --vm-bytes 512M
host:
stress --cpu 4 --io 4 --vm 2 --vm-bytes 256M
the below soft-lockup traces were observed after an hour or so and
persisted until the host was reset (this was found to be reliably
reproducible for this configuration, for kernels 4.15, 4.18, 5.0,
and 5.3-rc5):
[ 1253.183290] rcu: INFO: rcu_sched self-detected stall on CPU
[ 1253.183319] rcu: 124-....: (5250 ticks this GP) idle=10a/1/0x4000000000000002 softirq=5408/5408 fqs=1941
[ 1256.287426] watchdog: BUG: soft lockup - CPU#105 stuck for 23s! [CPU 52/KVM:19709]
[ 1264.075773] watchdog: BUG: soft lockup - CPU#24 stuck for 23s! [worker:19913]
[ 1264.079769] watchdog: BUG: soft lockup - CPU#31 stuck for 23s! [worker:20331]
[ 1264.095770] watchdog: BUG: soft lockup - CPU#45 stuck for 23s! [worker:20338]
[ 1264.131773] watchdog: BUG: soft lockup - CPU#64 stuck for 23s! [avocado:19525]
[ 1280.408480] watchdog: BUG: soft lockup - CPU#124 stuck for 22s! [ksmd:791]
[ 1316.198012] rcu: INFO: rcu_sched self-detected stall on CPU
[ 1316.198032] rcu: 124-....: (21003 ticks this GP) idle=10a/1/0x4000000000000002 softirq=5408/5408 fqs=8243
[ 1340.411024] watchdog: BUG: soft lockup - CPU#124 stuck for 22s! [ksmd:791]
[ 1379.212609] rcu: INFO: rcu_sched self-detected stall on CPU
[ 1379.212629] rcu: 124-....: (36756 ticks this GP) idle=10a/1/0x4000000000000002 softirq=5408/5408 fqs=14714
[ 1404.413615] watchdog: BUG: soft lockup - CPU#124 stuck for 22s! [ksmd:791]
[ 1442.227095] rcu: INFO: rcu_sched self-detected stall on CPU
[ 1442.227115] rcu: 124-....: (52509 ticks this GP) idle=10a/1/0x4000000000000002 softirq=5408/5408 fqs=21403
[ 1455.111787] INFO: task worker:19907 blocked for more than 120 seconds.
[ 1455.111822] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.111833] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.111884] INFO: task worker:19908 blocked for more than 120 seconds.
[ 1455.111905] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.111925] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.111966] INFO: task worker:20328 blocked for more than 120 seconds.
[ 1455.111986] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.111998] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.112048] INFO: task worker:20330 blocked for more than 120 seconds.
[ 1455.112068] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.112097] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.112138] INFO: task worker:20332 blocked for more than 120 seconds.
[ 1455.112159] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.112179] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.112210] INFO: task worker:20333 blocked for more than 120 seconds.
[ 1455.112231] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.112242] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.112282] INFO: task worker:20335 blocked for more than 120 seconds.
[ 1455.112303] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
[ 1455.112332] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1455.112372] INFO: task worker:20336 blocked for more than 120 seconds.
[ 1455.112392] Tainted: G L 5.3.0-rc5-mdr-vanilla+ #1
CPUs 45, 24, and 124 are stuck on spin locks, likely held by
CPUs 105 and 31.
CPUs 105 and 31 are stuck in smp_call_function_many(), waiting on
target CPU 42. For instance:
# CPU 105 registers (via xmon)
R00 = c00000000020b20c R16 = 00007d1bcd800000
R01 = c00000363eaa7970 R17 = 0000000000000001
R02 = c0000000019b3a00 R18 = 000000000000006b
R03 = 000000000000002a R19 = 00007d537d7aecf0
R04 = 000000000000002a R20 = 60000000000000e0
R05 = 000000000000002a R21 = 0801000000000080
R06 = c0002073fb0caa08 R22 = 0000000000000d60
R07 = c0000000019ddd78 R23 = 0000000000000001
R08 = 000000000000002a R24 = c00000000147a700
R09 = 0000000000000001 R25 = c0002073fb0ca908
R10 = c000008ffeb4e660 R26 = 0000000000000000
R11 = c0002073fb0ca900 R27 = c0000000019e2464
R12 = c000000000050790 R28 = c0000000000812b0
R13 = c000207fff623e00 R29 = c0002073fb0ca808
R14 = 00007d1bbee00000 R30 = c0002073fb0ca800
R15 = 00007d1bcd600000 R31 = 0000000000000800
pc = c00000000020b260 smp_call_function_many+0x3d0/0x460
cfar= c00000000020b270 smp_call_function_many+0x3e0/0x460
lr = c00000000020b20c smp_call_function_many+0x37c/0x460
msr = 900000010288b033 cr = 44024824
ctr = c000000000050790 xer = 0000000000000000 trap = 100
CPU 42 is running normally, doing VCPU work:
# CPU 42 stack trace (via xmon)
[link register ] c00800001be17188 kvmppc_book3s_radix_page_fault+0x90/0x2b0 [kvm_hv]
[c000008ed3343820] c000008ed3343850 (unreliable)
[c000008ed33438d0] c00800001be11b6c kvmppc_book3s_hv_page_fault+0x264/0xe30 [kvm_hv]
[c000008ed33439d0] c00800001be0d7b4 kvmppc_vcpu_run_hv+0x8dc/0xb50 [kvm_hv]
[c000008ed3343ae0] c00800001c10891c kvmppc_vcpu_run+0x34/0x48 [kvm]
[c000008ed3343b00] c00800001c10475c kvm_arch_vcpu_ioctl_run+0x244/0x420 [kvm]
[c000008ed3343b90] c00800001c0f5a78 kvm_vcpu_ioctl+0x470/0x7c8 [kvm]
[c000008ed3343d00] c000000000475450 do_vfs_ioctl+0xe0/0xc70
[c000008ed3343db0] c0000000004760e4 ksys_ioctl+0x104/0x120
[c000008ed3343e00] c000000000476128 sys_ioctl+0x28/0x80
[c000008ed3343e20] c00000000000b388 system_call+0x5c/0x70
--- Exception: c00 (System Call) at 00007d545cfd7694
SP (7d53ff7edf50) is in userspace
It was subsequently found that ipi_message[PPC_MSG_CALL_FUNCTION]
was set for CPU 42 by at least 1 of the CPUs waiting in
smp_call_function_many(), but somehow the corresponding
call_single_queue entries were never processed by CPU 42, causing the
callers to spin in csd_lock_wait() indefinitely.
Nick Piggin suggested something similar to the following sequence as
a possible explanation (interleaving of CALL_FUNCTION/RESCHEDULE
IPI messages seems to be most common, but any mix of CALL_FUNCTION and
!CALL_FUNCTION messages could trigger it):
CPU
X: smp_muxed_ipi_set_message():
X: smp_mb()
X: message[RESCHEDULE] = 1
X: doorbell_global_ipi(42):
X: kvmppc_set_host_ipi(42, 1)
X: ppc_msgsnd_sync()/smp_mb()
X: ppc_msgsnd() -> 42
42: doorbell_exception(): // from CPU X
42: ppc_msgsync()
105: smp_muxed_ipi_set_message():
105: smb_mb()
// STORE DEFERRED DUE TO RE-ORDERING
--105: message[CALL_FUNCTION] = 1
| 105: doorbell_global_ipi(42):
| 105: kvmppc_set_host_ipi(42, 1)
| 42: kvmppc_set_host_ipi(42, 0)
| 42: smp_ipi_demux_relaxed()
| 42: // returns to executing guest
| // RE-ORDERED STORE COMPLETES
->105: message[CALL_FUNCTION] = 1
105: ppc_msgsnd_sync()/smp_mb()
105: ppc_msgsnd() -> 42
42: local_paca->kvm_hstate.host_ipi == 0 // IPI ignored
105: // hangs waiting on 42 to process messages/call_single_queue
This can be prevented with an smp_mb() at the beginning of
kvmppc_set_host_ipi(), such that stores to message[<type>] (or other
state indicated by the host_ipi flag) are ordered vs. the store to
to host_ipi.
However, doing so might still allow for the following scenario (not
yet observed):
CPU
X: smp_muxed_ipi_set_message():
X: smp_mb()
X: message[RESCHEDULE] = 1
X: doorbell_global_ipi(42):
X: kvmppc_set_host_ipi(42, 1)
X: ppc_msgsnd_sync()/smp_mb()
X: ppc_msgsnd() -> 42
42: doorbell_exception(): // from CPU X
42: ppc_msgsync()
// STORE DEFERRED DUE TO RE-ORDERING
-- 42: kvmppc_set_host_ipi(42, 0)
| 42: smp_ipi_demux_relaxed()
| 105: smp_muxed_ipi_set_message():
| 105: smb_mb()
| 105: message[CALL_FUNCTION] = 1
| 105: doorbell_global_ipi(42):
| 105: kvmppc_set_host_ipi(42, 1)
| // RE-ORDERED STORE COMPLETES
-> 42: kvmppc_set_host_ipi(42, 0)
42: // returns to executing guest
105: ppc_msgsnd_sync()/smp_mb()
105: ppc_msgsnd() -> 42
42: local_paca->kvm_hstate.host_ipi == 0 // IPI ignored
105: // hangs waiting on 42 to process messages/call_single_queue
Fixing this scenario would require an smp_mb() *after* clearing
host_ipi flag in kvmppc_set_host_ipi() to order the store vs.
subsequent processing of IPI messages.
To handle both cases, this patch splits kvmppc_set_host_ipi() into
separate set/clear functions, where we execute smp_mb() prior to
setting host_ipi flag, and after clearing host_ipi flag. These
functions pair with each other to synchronize the sender and receiver
sides.
With that change in place the above workload ran for 20 hours without
triggering any lock-ups.
Fixes: 755563bc79 ("powerpc/powernv: Fixes for hypervisor doorbell handling") # v4.0
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Acked-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190911223155.16045-1-mdroth@linux.vnet.ibm.com
- Initial support for running on a system with an Ultravisor, which is software
that runs below the hypervisor and protects guests against some attacks by
the hypervisor.
- Support for building the kernel to run as a "Secure Virtual Machine", ie. as
a guest capable of running on a system with an Ultravisor.
- Some changes to our DMA code on bare metal, to allow devices with medium
sized DMA masks (> 32 && < 59 bits) to use more than 2GB of DMA space.
- Support for firmware assisted crash dumps on bare metal (powernv).
- Two series fixing bugs in and refactoring our PCI EEH code.
- A large series refactoring our exception entry code to use gas macros, both
to make it more readable and also enable some future optimisations.
As well as many cleanups and other minor features & fixups.
Thanks to:
Adam Zerella, Alexey Kardashevskiy, Alistair Popple, Andrew Donnellan, Aneesh
Kumar K.V, Anju T Sudhakar, Anshuman Khandual, Balbir Singh, Benjamin
Herrenschmidt, Cédric Le Goater, Christophe JAILLET, Christophe Leroy,
Christopher M. Riedl, Christoph Hellwig, Claudio Carvalho, Daniel Axtens,
David Gibson, David Hildenbrand, Desnes A. Nunes do Rosario, Ganesh Goudar,
Gautham R. Shenoy, Greg Kurz, Guerney Hunt, Gustavo Romero, Halil Pasic, Hari
Bathini, Joakim Tjernlund, Jonathan Neuschafer, Jordan Niethe, Leonardo Bras,
Lianbo Jiang, Madhavan Srinivasan, Mahesh Salgaonkar, Mahesh Salgaonkar,
Masahiro Yamada, Maxiwell S. Garcia, Michael Anderson, Nathan Chancellor,
Nathan Lynch, Naveen N. Rao, Nicholas Piggin, Oliver O'Halloran, Qian Cai, Ram
Pai, Ravi Bangoria, Reza Arbab, Ryan Grimm, Sam Bobroff, Santosh Sivaraj,
Segher Boessenkool, Sukadev Bhattiprolu, Thiago Bauermann, Thiago Jung
Bauermann, Thomas Gleixner, Tom Lendacky, Vasant Hegde.
-----BEGIN PGP SIGNATURE-----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=WlW+
-----END PGP SIGNATURE-----
Merge tag 'powerpc-5.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc updates from Michael Ellerman:
"This is a bit late, partly due to me travelling, and partly due to a
power outage knocking out some of my test systems *while* I was
travelling.
- Initial support for running on a system with an Ultravisor, which
is software that runs below the hypervisor and protects guests
against some attacks by the hypervisor.
- Support for building the kernel to run as a "Secure Virtual
Machine", ie. as a guest capable of running on a system with an
Ultravisor.
- Some changes to our DMA code on bare metal, to allow devices with
medium sized DMA masks (> 32 && < 59 bits) to use more than 2GB of
DMA space.
- Support for firmware assisted crash dumps on bare metal (powernv).
- Two series fixing bugs in and refactoring our PCI EEH code.
- A large series refactoring our exception entry code to use gas
macros, both to make it more readable and also enable some future
optimisations.
As well as many cleanups and other minor features & fixups.
Thanks to: Adam Zerella, Alexey Kardashevskiy, Alistair Popple, Andrew
Donnellan, Aneesh Kumar K.V, Anju T Sudhakar, Anshuman Khandual,
Balbir Singh, Benjamin Herrenschmidt, Cédric Le Goater, Christophe
JAILLET, Christophe Leroy, Christopher M. Riedl, Christoph Hellwig,
Claudio Carvalho, Daniel Axtens, David Gibson, David Hildenbrand,
Desnes A. Nunes do Rosario, Ganesh Goudar, Gautham R. Shenoy, Greg
Kurz, Guerney Hunt, Gustavo Romero, Halil Pasic, Hari Bathini, Joakim
Tjernlund, Jonathan Neuschafer, Jordan Niethe, Leonardo Bras, Lianbo
Jiang, Madhavan Srinivasan, Mahesh Salgaonkar, Mahesh Salgaonkar,
Masahiro Yamada, Maxiwell S. Garcia, Michael Anderson, Nathan
Chancellor, Nathan Lynch, Naveen N. Rao, Nicholas Piggin, Oliver
O'Halloran, Qian Cai, Ram Pai, Ravi Bangoria, Reza Arbab, Ryan Grimm,
Sam Bobroff, Santosh Sivaraj, Segher Boessenkool, Sukadev Bhattiprolu,
Thiago Bauermann, Thiago Jung Bauermann, Thomas Gleixner, Tom
Lendacky, Vasant Hegde"
* tag 'powerpc-5.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (264 commits)
powerpc/mm/mce: Keep irqs disabled during lockless page table walk
powerpc: Use ftrace_graph_ret_addr() when unwinding
powerpc/ftrace: Enable HAVE_FUNCTION_GRAPH_RET_ADDR_PTR
ftrace: Look up the address of return_to_handler() using helpers
powerpc: dump kernel log before carrying out fadump or kdump
docs: powerpc: Add missing documentation reference
powerpc/xmon: Fix output of XIVE IPI
powerpc/xmon: Improve output of XIVE interrupts
powerpc/mm/radix: remove useless kernel messages
powerpc/fadump: support holes in kernel boot memory area
powerpc/fadump: remove RMA_START and RMA_END macros
powerpc/fadump: update documentation about option to release opalcore
powerpc/fadump: consider f/w load area
powerpc/opalcore: provide an option to invalidate /sys/firmware/opal/core file
powerpc/opalcore: export /sys/firmware/opal/core for analysing opal crashes
powerpc/fadump: update documentation about CONFIG_PRESERVE_FA_DUMP
powerpc/fadump: add support to preserve crash data on FADUMP disabled kernel
powerpc/fadump: improve how crashed kernel's memory is reserved
powerpc/fadump: consider reserved ranges while releasing memory
powerpc/fadump: make crash memory ranges array allocation generic
...
With support to copy multiple kernel boot memory regions owing to copy
size limitation, also handle holes in the memory area to be preserved.
Support as many as 128 kernel boot memory regions. This allows having
an adequate FADump capture kernel size for different scenarios.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821385448.5656.6124791213910877759.stgit@hbathini.in.ibm.com
RMA_START is defined as '0' and there is even a BUILD_BUG_ON() to
make sure it is never anything else. Remove this macro and use '0'
instead as code change is needed anyway when it has to be something
else. Also, remove unused RMA_END macro.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821384096.5656.15026984053970204652.stgit@hbathini.in.ibm.com
OPAL loads kernel & initrd at 512MB offset (256MB size), also exported
as ibm,opal/dump/fw-load-area. So, if boot memory size of FADump is
less than 768MB, kernel memory to be exported as '/proc/vmcore' would
be overwritten by f/w while loading kernel & initrd. To avoid such a
scenario, enforce a minimum boot memory size of 768MB on OPAL platform
and skip using FADump if a newer F/W version loads kernel & initrd
above 768MB.
Also, irrespective of RMA size, set the minimum boot memory size
expected on pseries platform at 320MB. This is to avoid inflating the
minimum memory requirements on systems with 512M/1024M RMA size.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821381414.5656.1592867278535469652.stgit@hbathini.in.ibm.com
Export /sys/firmware/opal/core file to analyze opal crashes. Since OPAL
core can be generated independent of CONFIG_FA_DUMP support in kernel,
add this support under a new kernel config option CONFIG_OPAL_CORE.
Also, avoid code duplication by moving common code used while exporting
/proc/vmcore and/or /sys/firmware/opal/core file(s).
Signed-off-by: Hari Bathini <hbathini@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821378503.5656.3693769384945087756.stgit@hbathini.in.ibm.com
Add a new kernel config option, CONFIG_PRESERVE_FA_DUMP that ensures
that crash data, from previously crash'ed kernel, is preserved. This
helps in cases where FADump is not enabled but the subsequent memory
preserving kernel boot is likely to process this crash data. One
typical usecase for this config option is petitboot kernel.
As OPAL allows registering address with it in the first kernel and
retrieving it after MPIPL, use it to store the top of boot memory.
A kernel that intends to preserve crash data retrieves it and avoids
using memory beyond this address.
Move arch_reserved_kernel_pages() function as it is needed for both
FA_DUMP and PRESERVE_FA_DUMP configurations.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821375751.5656.11459483669542541602.stgit@hbathini.in.ibm.com
Firmware provides architected register state data at the time of crash.
Process this data and build CPU notes to append to ELF core. In case
this data is missing or in unsupported format, at least append crashing
CPU's register data, to have something to work with in the vmcore file.
Signed-off-by: Hari Bathini <hbathini@linux.vnet.ibm.com>
Signed-off-by: Vasant Hegde <hegdevasant@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821367702.5656.5546683836236508389.stgit@hbathini.in.ibm.com
Make OPAL call to indicate that the dump is processed and the metadata
area in OPAL can be cleared/released. Also, setup/initialize FADump
for re-registration.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821356046.5656.12270927048195494911.stgit@hbathini.in.ibm.com
If all kernel boot memory regions are not registered for MPIPL before
system crashes, try processing the partial crashdump but warn the user
before proceeding.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821352793.5656.1734051341024721407.stgit@hbathini.in.ibm.com
Add support in the kernel to process the crash'ed kernel's memory
preserved during MPIPL and export it as /proc/vmcore file for the
userland scripts to filter and analyze it later.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821351482.5656.6255805804744333073.stgit@hbathini.in.ibm.com
Firmware uses a 32-bit field for size while copying/backing-up memory
during MPIPL. So, the maximum value that could be represented with
a PAGE_SIZE aligned 32-bit field will be the maximum copy size for a
region but FADump capture kernel usually needs more memory than that
to be preserved to avoid running into out of memory errors.
So, request firmware to copy multiple kernel boot memory regions
instead of just one (which worked fine for pseries as 64-bit field
was used for size there).
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821350193.5656.3664853158523582019.stgit@hbathini.in.ibm.com
During kexec boot, metadata address needs to be reset to avoid running
into errors interpreting stale metadata address, in case the kexec'ed
kernel crashes before metadata address could be setup again.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821346629.5656.10783321582005237813.stgit@hbathini.in.ibm.com
OPAL allows registering address with it in the first kernel and
retrieving it after MPIPL. Setup kernel metadata and register its
address with OPAL to use it for processing the crash dump.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821345011.5656.13567765019032928471.stgit@hbathini.in.ibm.com
MPIPL is Memory Preserving IPL supported from POWER9. This enables the
kernel to reset the system with memory 'preserved'. Also, it supports
copying memory from a source address to some destination address during
MPIPL boot. Add MPIPL interface definitions here to leverage these f/w
features in adding FADump support for PowerNV platform.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821340710.5656.10071829040515662624.stgit@hbathini.in.ibm.com
Move platform specific register/un-register code, the RTAS calls, to
register/un-register callback functions. This would also mean moving
code that initializes and prints the platform specific FADump data.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Reviewed-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821332856.5656.16380417702046411631.stgit@hbathini.in.ibm.com
Introduce callback functions for platform specific operations like
register, unregister, invalidate & such. Also, define place-holders
for the same on pSeries platform.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821330286.5656.15538934400074110770.stgit@hbathini.in.ibm.com
Currently, FADump is only supported on pSeries but that is going to
change soon with FADump support being added on PowerNV platform. So,
move rtas specific definitions to platform code to allow FADump
to have multiple platforms support.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821328494.5656.16219929140866195511.stgit@hbathini.in.ibm.com
The builds breaks when IOMMU_API=n, eg. skiroot_defconfig:
arch/powerpc/platforms/powernv/npu-dma.c:96:28: error: 'get_gpu_pci_dev_and_pe' defined but not used
arch/powerpc/platforms/powernv/npu-dma.c:126:13: error: 'pnv_npu_set_window' defined but not used
Fixes: b4d37a7b69 ("powerpc/powernv: Remove unused pnv_npu_try_dma_set_bypass() function")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
There's a bug in skiboot that causes the OPAL_XIVE_ALLOCATE_IRQ call
to return the 32-bit value 0xffffffff when OPAL has run out of IRQs.
Unfortunatelty, OPAL return values are signed 64-bit entities and
errors are supposed to be negative. If that happens, the linux code
confusingly treats 0xffffffff as a valid IRQ number and panics at some
point.
A fix was recently merged in skiboot:
e97391ae2bb5 ("xive: fix return value of opal_xive_allocate_irq()")
but we need a workaround anyway to support older skiboots already
in the field.
Internally convert 0xffffffff to OPAL_RESOURCE which is the usual error
returned upon resource exhaustion.
Cc: stable@vger.kernel.org # v4.12+
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/156821713818.1985334.14123187368108582810.stgit@bahia.lan
prep_irq_for_idle() is intended to be called before entering
H_CEDE (and it is used by the pseries cpuidle driver). However the
default pseries idle routine does not call it, leading to mismanaged
lazy irq state when the cpuidle driver isn't in use. Manifestations of
this include:
* Dropped IPIs in the time immediately after a cpu comes
online (before it has installed the cpuidle handler), making the
online operation block indefinitely waiting for the new cpu to
respond.
* Hitting this WARN_ON in arch_local_irq_restore():
/*
* We should already be hard disabled here. We had bugs
* where that wasn't the case so let's dbl check it and
* warn if we are wrong. Only do that when IRQ tracing
* is enabled as mfmsr() can be costly.
*/
if (WARN_ON_ONCE(mfmsr() & MSR_EE))
__hard_irq_disable();
Call prep_irq_for_idle() from pseries_lpar_idle() and honor its
result.
Fixes: 363edbe261 ("powerpc: Default arch idle could cede processor on pseries")
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190910225244.25056-1-nathanl@linux.ibm.com
We have OPAL_MSG_PRD message type to pass prd related messages from
OPAL to `opal-prd`. It can handle messages upto 64 bytes. We have a
requirement to send bigger than 64 bytes of data from OPAL to
`opal-prd`. Lets add new message type (OPAL_MSG_PRD2) to pass bigger
data.
Signed-off-by: Vasant Hegde <hegdevasant@linux.vnet.ibm.com>
[mpe: Make the error string clear that it's the PRD2 event that failed]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190826065701.8853-2-hegdevasant@linux.vnet.ibm.com
Use "opal-msg-size" device tree property to allocate memory for
"opal_msg".
Signed-off-by: Vasant Hegde <hegdevasant@linux.vnet.ibm.com>
[mpe: s/uint32_t/u32/ and mark opal_msg_size as __ro_after_init]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190826065701.8853-1-hegdevasant@linux.vnet.ibm.com
Neither pnv_npu_try_dma_set_bypass() nor the pnv_npu_dma_set_32() and
pnv_npu_dma_set_bypass() helpers called by it are used anywhere in the
kernel tree, so remove them.
mpe: They're unused since 2d6ad41b2c ("powerpc/powernv: use the
generic iommu bypass code") removed the last usage.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190903165147.11099-1-hch@lst.de
Commit <684d984038aa> ('powerpc/powernv: Add debugfs interface for
imc-mode and imc') added debugfs interface for the nest imc pmu
devices to support changing of different ucode modes. Primarily adding
this capability for debug. But when doing so, the code did not
consider the case of cpu-less nodes. So when reading the _cmd_ or
_mode_ file of a cpu-less node will create this crash.
Faulting instruction address: 0xc0000000000d0d58
Oops: Kernel access of bad area, sig: 11 [#1]
...
CPU: 67 PID: 5301 Comm: cat Not tainted 5.2.0-rc6-next-20190627+ #19
NIP: c0000000000d0d58 LR: c00000000049aa18 CTR:c0000000000d0d50
REGS: c00020194548f9e0 TRAP: 0300 Not tainted (5.2.0-rc6-next-20190627+)
MSR: 9000000000009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR:28022822 XER: 00000000
CFAR: c00000000049aa14 DAR: 000000000003fc08 DSISR:40000000 IRQMASK: 0
...
NIP imc_mem_get+0x8/0x20
LR simple_attr_read+0x118/0x170
Call Trace:
simple_attr_read+0x70/0x170 (unreliable)
debugfs_attr_read+0x6c/0xb0
__vfs_read+0x3c/0x70
vfs_read+0xbc/0x1a0
ksys_read+0x7c/0x140
system_call+0x5c/0x70
Patch fixes the issue with a more robust check for vbase to NULL.
Before patch, ls output for the debugfs imc directory
# ls /sys/kernel/debug/powerpc/imc/
imc_cmd_0 imc_cmd_251 imc_cmd_253 imc_cmd_255 imc_mode_0 imc_mode_251 imc_mode_253 imc_mode_255
imc_cmd_250 imc_cmd_252 imc_cmd_254 imc_cmd_8 imc_mode_250 imc_mode_252 imc_mode_254 imc_mode_8
After patch, ls output for the debugfs imc directory
# ls /sys/kernel/debug/powerpc/imc/
imc_cmd_0 imc_cmd_8 imc_mode_0 imc_mode_8
Actual bug here is that, we have two loops with potentially different
loop counts. That is, in imc_get_mem_addr_nest(), loop count is
obtained from the dt entries. But in case of export_imc_mode_and_cmd(),
loop was based on for_each_nid() count. Patch fixes the loop count in
latter based on the struct mem_info. Ideally it would be better to
have array size in struct imc_pmu.
Fixes: 684d984038 ('powerpc/powernv: Add debugfs interface for imc-mode and imc')
Reported-by: Qian Cai <cai@lca.pw>
Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190827101635.6942-1-maddy@linux.vnet.ibm.com
The various translation structure invalidations performed in early boot
when the MMU is off are not required, because everything is invalidated
immediately before a CPU first enables its MMU (see early_init_mmu
and early_init_mmu_secondary).
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190902152931.17840-6-npiggin@gmail.com
This callback is only required because the partition table init comes
before process table allocation on powernv (aka bare metal aka native).
Change the order to allocate the process table first, and remove the
callback.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190902152931.17840-2-npiggin@gmail.com
Currently we check that an IODA2 compatible PHB is upstream of this slot.
This is mainly to avoid pnv_php creating slots for the various "virtual
PHBs" that we create for NVLink. There's no real need for this restriction
so allow it on IODA3.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190903101605.2890-10-oohall@gmail.com
When we reset PCI devices managed by a hotplug driver the reset may
generate spurious hotplug events that cause the PCI device we're resetting
to be torn down accidently. This is a problem for EEH (when the driver is
EEH aware) since we want to leave the OS PCI device state intact so that
the device can be re-set without losing any resources (network, disks,
etc) provided by the driver.
Generic PCI code provides the pci_bus_error_reset() function to handle
resetting a PCI Device (or bus) by using the reset method provided by the
hotplug slot driver. We can use this function if the EEH core has
requested a hot reset (common case) without tripping over the hotplug
driver.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190903101605.2890-8-oohall@gmail.com
Support for switching CAPI cards into and out of CAPI mode was removed a
while ago. Drop the comment since it's no longer relevant.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190903101605.2890-7-oohall@gmail.com
Most dma_map_ops instances are IOMMUs that work perfectly fine in 32-bits
of IOVA space, and the generic direct mapping code already provides its
own routines that is intelligent based on the amount of memory actually
present. Wire up the dma-direct routine for the ARM direct mapping code
as well, and otherwise default to the constant 32-bit mask. This way
we only need to override it for the occasional odd IOMMU that requires
64-bit IOVA support, or IOMMU drivers that are more efficient if they
can fall back to the direct mapping.
Signed-off-by: Christoph Hellwig <hch@lst.de>
While the default ->mmap and ->get_sgtable implementations work for the
majority of our dma_map_ops impementations they are inherently safe
for others that don't use the page allocator or CMA and/or use their
own way of remapping not covered by the common code. So remove the
defaults if these methods are not wired up, but instead wire up the
default implementations for all safe instances.
Fixes: e1c7e32453 ("dma-mapping: always provide the dma_map_ops based implementation")
Signed-off-by: Christoph Hellwig <hch@lst.de>
The common machine_check_event data structures and queues are mostly
platform independent, with powernv decoding SRR1/DSISR/etc., into
machine_check_event objects.
This patch converts pseries to use this infrastructure by decoding
fwnmi/rtas data into machine_check_event objects.
This allows queueing to be used by a subsequent change to delay the
virtual mode handling of machine checks that occur in kernel space
where it is unsafe to switch immediately to virtual mode, similarly
to powernv.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[mpe: Fix implicit fallthrough warnings in mce_handle_error()]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190802105709.27696-10-npiggin@gmail.com
Re-use the code introduced in pseries to save and dump the contents
of the SLB in the case of an SLB involved machine check exception.
This patch also avoids allocating the SLB save array on pseries radix.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190802105709.27696-9-npiggin@gmail.com
SWIOTLB checks range of incoming CPU addresses to be bounced and sees if
the device can access it through its DMA window without requiring bouncing.
In such cases it just chooses to skip bouncing. But for cases like secure
guests on powerpc platform all addresses need to be bounced into the shared
pool of memory because the host cannot access it otherwise. Hence the need
to do the bouncing is not related to device's DMA window and use of bounce
buffers is forced by setting swiotlb_force.
Also, connect the shared memory conversion functions into the
ARCH_HAS_MEM_ENCRYPT hooks and call swiotlb_update_mem_attributes() to
convert SWIOTLB's memory pool to shared memory.
Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
[ bauerman: Use ARCH_HAS_MEM_ENCRYPT hooks to share swiotlb memory pool. ]
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190820021326.6884-15-bauerman@linux.ibm.com
Secure guest memory is inacessible to devices so regular DMA isn't
possible.
In that case set devices' dma_map_ops to NULL so that the generic
DMA code path will use SWIOTLB to bounce buffers for DMA.
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190820021326.6884-14-bauerman@linux.ibm.com
Normally, the HV emulates some instructions like MSGSNDP, MSGCLRP
from a KVM guest. To emulate the instructions, it must first read
the instruction from the guest's memory and decode its parameters.
However for a secure guest (aka SVM), the page containing the
instruction is in secure memory and the HV cannot access directly.
It would need the Ultravisor (UV) to facilitate accessing the
instruction and parameters but the UV currently does not have
the support for such accesses.
Until the UV has such support, disable doorbells in SVMs. This might
incur a performance hit but that is yet to be quantified.
With this patch applied (needed only in SVMs not needed for HV) we
are able to launch SVM guests with multi-core support. Eg:
qemu -smp sockets=2,cores=2,threads=2.
Fix suggested by Benjamin Herrenschmidt. Thanks to input from
Paul Mackerras, Ram Pai and Michael Anderson.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190820021326.6884-13-bauerman@linux.ibm.com
Secure guests need to share the DTL buffers with the hypervisor. To that
end, use a kmem_cache constructor which converts the underlying buddy
allocated SLUB cache pages into shared memory.
Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190820021326.6884-10-bauerman@linux.ibm.com
Introduce CONFIG_PPC_SVM to control support for secure guests and include
Ultravisor-related helpers when it is selected
Signed-off-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190820021326.6884-3-bauerman@linux.ibm.com
The ultravisor (UV) provides an in-memory console which follows the
OPAL in-memory console structure.
This patch extends the OPAL msglog code to initialize the UV memory
console and provide the "/sys/firmware/ultravisor/msglog" interface
for userspace to view the UV message log.
Signed-off-by: Claudio Carvalho <cclaudio@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Tested-by: Claudio Carvalho <cclaudio@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190828130521.26764-2-mpe@ellerman.id.au
This patch refactors the code in opal-msglog that operates on the OPAL
memory console in order to make it cleaner and also allow the reuse of
the new memcons_* functions.
Signed-off-by: Claudio Carvalho <cclaudio@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Tested-by: Claudio Carvalho <cclaudio@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190828130521.26764-1-mpe@ellerman.id.au
LDBAR is a per-thread SPR populated and used by the thread-imc pmu
driver to dump the data counter into memory. It contains memory along
with few other configuration bits. LDBAR is populated and enabled only
when any of the thread imc pmu events are monitored.
In ultravisor enabled systems, LDBAR becomes ultravisor privileged and
an attempt to write to it will cause a Hypervisor Emulation Assistance
interrupt.
In ultravisor enabled systems, the ultravisor is responsible to maintain
the LDBAR (e.g. save and restore it).
This restricts LDBAR access to only when ultravisor is disabled.
Signed-off-by: Claudio Carvalho <cclaudio@linux.ibm.com>
Reviewed-by: Ram Pai <linuxram@us.ibm.com>
Reviewed-by: Ryan Grimm <grimm@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190822034838.27876-7-cclaudio@linux.ibm.com
In PEF enabled systems, some of the resources which were previously
hypervisor privileged are now ultravisor privileged and controlled by
the ultravisor firmware.
This adds FW_FEATURE_ULTRAVISOR to indicate if PEF is enabled.
The host kernel can use FW_FEATURE_ULTRAVISOR, for instance, to skip
accessing resources (e.g. PTCR and LDBAR) in case PEF is enabled.
Signed-off-by: Claudio Carvalho <cclaudio@linux.ibm.com>
[ andmike: Device node name to "ibm,ultravisor" ]
Signed-off-by: Michael Anderson <andmike@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190822034838.27876-4-cclaudio@linux.ibm.com
As now we have xchg_no_kill/tce_kill, these are not used anymore so
remove them.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190829085252.72370-6-aik@ozlabs.ru
