- 16 Dec, 2021 10 commits
-
-
Nicholas Piggin authored
Reading the CFAR register is quite costly (~20 cycles on POWER9). It is a good idea to have for most synchronous interrupts, but for async ones it is much less important. Doorbell, external, and decrementer interrupts are the important asynchronous ones. HV interrupts can't skip CFAR if KVM HV is possible, because it might be a guest exit that requires CFAR preserved. But the important pseries interrupts can avoid loading CFAR. Signed-off-by:
Nicholas Piggin <npiggin@gmail.com> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210922145452.352571-7-npiggin@gmail.com
-
Nicholas Piggin authored
Move the assertions requiring restart table searches under CONFIG_PPC_IRQ_SOFT_MASK_DEBUG. Signed-off-by:
-
Nicholas Piggin authored
Enabling MSR[EE] in interrupt handlers while interrupts are still soft masked allows PMIs to profile interrupt handlers to some degree, beyond what SIAR latching allows. When perf is not being used, this is almost useless work. It requires an extra mtmsrd in the irq handler, and it also opens the door to masked interrupts hitting and requiring replay, which is more expensive than just taking them directly. This effect can be noticable in high IRQ workloads. Avoid enabling MSR[EE] unless perf is currently in use. This saves about 60 cycles (or 8%) on a simple decrementer interrupt microbenchmark. Replayed interrupts drop from 1.4% of all interrupts taken, to 0.003%. This does prevent the soft-nmi interrupt being taken in these handlers, but that's not too reliable anyway. The SMP watchdog will continue to be the reliable way to catch lockups. Signed-off-by:
-
Nicholas Piggin authored
Interrupt code enables MSR[EE] in some irq handlers while keeping local irqs disabled via soft-mask, allowing PMI interrupts to be taken as soft-NMI to improve profiling of irq handlers. When perf is not enabled, there is no point to doing this, it's additional overhead. So provide a function that can say if PMIs should be taken promptly if possible. Signed-off-by:
-
Nicholas Piggin authored
The mtmsrd to enable MSR[RI] can be combined with the mtmsrd to enable MSR[EE] in interrupt entry code, for those interrupts which enable EE. This helps performance of important synchronous interrupts (e.g., page faults). This is similar to what commit dd152f70 ("powerpc/64s: system call avoid setting MSR[RI] until we set MSR[EE]") does for system calls. Do this by enabling EE and RI together at the beginning of the entry wrapper if PACA_IRQ_HARD_DIS is clear, and only enabling RI if it is set. Asynchronous interrupts set PACA_IRQ_HARD_DIS, but synchronous ones leave it unchanged, so by default they always get EE=1 unless they have interrupted a caller that is hard disabled. When the sync interrupt later calls interrupt_cond_local_irq_enable(), it will not require another mtmsrd because MSR[EE] was already enabled here. This avoids one mtmsrd L=1 for synchronous interrupts on 64s, which saves about 20 cycles on POWER9. And for kernel-mode interrupts, both synchronous and asynchronous, this saves an additional 40 cycles due to the mtmsrd being moved ahead of mfspr SPRN_AMR, which prevents a SPR scoreboard stall. Signed-off-by:
-
Nicholas Piggin authored
Make synchronous interrupt handler entry wrappers enable MSR[EE] if MSR[EE] was enabled in the interrupted context. IRQs are soft-disabled at this point so there is no change to high level code, but it's a masked interrupt could fire. This is a performance disadvantage for interrupts which do not later call interrupt_cond_local_irq_enable(), because an an additional mtmsrd or wrtee instruction is executed. However the important synchronous interrupts (e.g., page fault) do enable interrupts, so the performance disadvantage is mostly avoided. In the next patch, MSR[RI] enabling can be combined with MSR[EE] enabling, which mitigates the performance drop for the former and gives a performance advanage for the latter interrupts, on 64s machines. 64e is coming along for the ride for now to avoid divergences with 64s in this tricky code. Signed-off-by:
-
Nicholas Piggin authored
KVM does not support VAS so guests always print a useless error on boot vas: HCALL(398) error -2, query_type 0, result buffer 0x57f2000 Change this to only print the message if the error is not H_FUNCTION. Signed-off-by: -
Kajol Jain authored
The code represent memory/cache level data based on PERF_MEM_LVL_* namespace, which is in the process of deprication in the favour of newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_,HOPS_} fields. Add data source encodings to represent cache/memory data based on newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_,HOPS_} fields. Add data source encodings to represent data coming from local memory/Remote memory/distant memory and remote/distant cache hits. In order to represent data coming from OpenCAPI cache/memory, we use LVLNUM "PMEM" field which is used to present persistent memory accesses. Result in power10 system with patch changes: localhost:# ./perf mem report --sort="mem,sym,dso" --stdio # Overhead Samples Memory access Symbol Shared Object # ........ ............ ........................ .......................... ................ # 29.46% 2331 L1 or L1 hit [.] __random libc-2.28.so 23.11% 2121 L1 or L1 hit [.] producer_populate_cache producer_consumer 18.56% 1758 L1 or L1 hit [.] __random_r libc-2.28.so 15.64% 1559 L2 or L2 hit [.] __random libc-2.28.so ..... 0.09% 5 Remote socket, same board Any cache hit [.] __random libc-2.28.so 0.07% 4 Remote socket, same board Any cache hit [.] __random libc-2.28.so ..... Signed-off-by:Kajol Jain <kjain@linux.ibm.com> Reviewed-by:
Madhavan Srinivasan <maddy@linux.ibm.com> Signed-off-by:
-
Kajol Jain authored
The code represent data coming from L1/L2/L3 cache hits based on PERF_MEM_LVL_* namespace, which is in the process of deprecation in the favour of newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_,HOPS_} fields. Add data source encodings to represent L1/L2/L3 cache hits based on newer composite PERF_MEM_{LVLNUM_,REMOTE_,SNOOPX_,HOPS_} fields for power10 and older platforms Result in power9 system without patch changes: localhost:# ./perf mem report --sort="mem,sym,dso" --stdio # Overhead Samples Memory access Symbol Shared Object # ........ ............ ........................ ................................. ................ # 29.51% 1 L2 hit [k] perf_event_exec [kernel.vmlinux] 27.05% 1 L1 hit [k] perf_ctx_unlock [kernel.vmlinux] 13.93% 1 L1 hit [k] vtime_delta [kernel.vmlinux] 13.11% 1 L1 hit [k] prepend_path.isra.11 [kernel.vmlinux] 8.20% 1 L1 hit [.] 00000038.plt_call.__GI_strlen libc-2.28.so 8.20% 1 L1 hit [k] perf_event_interrupt [kernel.vmlinux] Result in power9 system with patch changes: localhost:# ./perf mem report --sort="mem,sym,dso" --stdio # Overhead Samples Memory access Symbol Shared Object # ........ ............ ........................ .......................... ................ # 36.63% 1 L2 or L2 hit [k] perf_event_exec [kernel.vmlinux] 25.50% 1 L1 or L1 hit [k] vtime_delta [kernel.vmlinux] 13.12% 1 L1 or L1 hit [k] unmap_region [kernel.vmlinux] 12.62% 1 L1 or L1 hit [k] perf_sample_event_took [kernel.vmlinux] 6.93% 1 L1 or L1 hit [k] perf_ctx_unlock [kernel.vmlinux] 5.20% 1 L1 or L1 hit [.] __memcpy_power7 libc-2.28.so Signed-off-by: -
Kajol Jain authored
Add new macros for mem_hops field which can be used to represent remote-node, socket and board level details. Currently the code had macro for HOPS_0, which corresponds to data coming from another core but same node. Add new macros for HOPS_1 to HOPS_3 to represent remote-node, socket and board level data. For ex: Encodings for mem_hops fields with L2 cache: L2 - local L2 L2 | REMOTE | HOPS_0 - remote core, same node L2 L2 | REMOTE | HOPS_1 - remote node, same socket L2 L2 | REMOTE | HOPS_2 - remote socket, same board L2 L2 | REMOTE | HOPS_3 - remote board L2 Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by:
-
- 15 Dec, 2021 1 commit
-
-
Michael Ellerman authored
Bring in some more KVM commits from our KVM topic branch.
-
- 14 Dec, 2021 1 commit
-
-
Sean Christopherson authored
Use kvm_arch_vcpu_get_wait() to get a vCPU's rcuwait object instead of using vcpu->wait directly in kvmhv_run_single_vcpu(). Functionally, this is a nop as vcpu->arch.waitp is guaranteed to point at vcpu->wait. But that is not obvious at first glance, and a future change coming in via the KVM tree, commit 510958e9 ("KVM: Force PPC to define its own rcuwait object"), will hide vcpu->wait from architectures that define __KVM_HAVE_ARCH_WQP to prevent generic KVM from attepting to wake a vCPU with the wrong rcuwait object. Reported-by:
Sachin Sant <sachinp@linux.vnet.ibm.com> Signed-off-by:
Sean Christopherson <seanjc@google.com> Tested-by:
-
- 09 Dec, 2021 28 commits
-
-
Christophe Leroy authored
Commit df1f679d ("powerpc/powermac: Add missing lockdep_register_key()") fixed a problem that was causing a WARNING. There are two other places in the same file with the same problem originating from commit 9e607f72 ("i2c_powermac: shut up lockdep warning"). Add missing lockdep_register_key() Fixes: 9e607f72 ("i2c_powermac: shut up lockdep warning") Reported-by:
Erhard Furtner <erhard_f@mailbox.org> Signed-off-by:
Christophe Leroy <christophe.leroy@csgroup.eu> Depends-on: df1f679d ("powerpc/powermac: Add missing lockdep_register_key()") Signed-off-by:
-
Hari Bathini authored
In panic path, fadump is triggered via a panic notifier function. Before calling panic notifier functions, smp_send_stop() gets called, which stops all CPUs except the panic'ing CPU. Commit 8389b37d ("powerpc: stop_this_cpu: remove the cpu from the online map.") and again commit bab26238 ("powerpc: Offline CPU in stop_this_cpu()") started marking CPUs as offline while stopping them. So, if a kernel has either of the above commits, vmcore captured with fadump via panic path would not process register data for all CPUs except the panic'ing CPU. Sample output of crash-utility with such vmcore: # crash vmlinux vmcore ... KERNEL: vmlinux DUMPFILE: vmcore [PARTIAL DUMP] CPUS: 1 DATE: Wed Nov 10 09:56:34 EST 2021 UPTIME: 00:00:42 LOAD AVERAGE: 2.27, 0.69, 0.24 TASKS: 183 NODENAME: XXXXXXXXX RELEASE: 5.15.0+ VERSION: #974 SMP Wed Nov 10 04:18:19 CST 2021 MACHINE: ppc64le (2500 Mhz) MEMORY: 8 GB PANIC: "Kernel panic - not syncing: sysrq triggered crash" PID: 3394 COMMAND: "bash" TASK: c0000000150a5f80 [THREAD_INFO: c0000000150a5f80] CPU: 1 STATE: TASK_RUNNING (PANIC) crash> p -x __cpu_online_mask __cpu_online_mask = $1 = { bits = {0x2, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0} } crash> crash> crash> p -x __cpu_active_mask __cpu_active_mask = $2 = { bits = {0xff, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0} } crash> While this has been the case since fadump was introduced, the issue was not identified for two probable reasons: - In general, the bulk of the vmcores analyzed were from crash due to exception. - The above did change since commit 8341f2f2 ("sysrq: Use panic() to force a crash") started using panic() instead of deferencing NULL pointer to force a kernel crash. But then commit de6e5d38 ("powerpc: smp_send_stop do not offline stopped CPUs") stopped marking CPUs as offline till kernel commit bab26238 ("powerpc: Offline CPU in stop_this_cpu()") reverted that change. To ensure post processing register data of all other CPUs happens as intended, let panic() function take the crash friendly path (read crash_smp_send_stop()) with the help of crash_kexec_post_notifiers option. Also, as register data for all CPUs is captured by f/w, skip IPI callbacks here for fadump, to avoid any complications in finding the right backtraces. Signed-off-by:
Hari Bathini <hbathini@linux.ibm.com> Signed-off-by:
-
Hari Bathini authored
Kdump can be triggered after panic_notifers since commit f06e5153 ("kernel/panic.c: add "crash_kexec_post_notifiers" option for kdump after panic_notifers") introduced crash_kexec_post_notifiers option. But using this option would mean smp_send_stop(), that marks all other CPUs as offline, gets called before kdump is triggered. As a result, kdump routines fail to save other CPUs' registers. To fix this, kdump friendly crash_smp_send_stop() function was introduced with kernel commit 0ee59413 ("x86/panic: replace smp_send_stop() with kdump friendly version in panic path"). Override this kdump friendly weak function to handle crash_kexec_post_notifiers option appropriately on powerpc. Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
-
Thadeu Lima de Souza Cascardo authored
A mis-match between reported and actual mitigation is not restricted to the Vulnerable case. The guest might also report the mitigation as "Software count cache flush" and the host will still mitigate with branch cache disabled. So, instead of skipping depending on the detected mitigation, simply skip whenever the detected miss_percent is the expected one for a fully mitigated system, that is, above 95%. Signed-off-by:
Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Signed-off-by:
-
Anders Roxell authored
Clang warns: arch/powerpc/platforms/cell/pervasive.c:81:2: error: unannotated fall-through between switch labels case SRR1_WAKEEE: ^ arch/powerpc/platforms/cell/pervasive.c:81:2: note: insert 'break;' to avoid fall-through case SRR1_WAKEEE: ^ break; 1 error generated. Clang is more pedantic than GCC, which does not warn when failing through to a case that is just break or return. Clang's version is more in line with the kernel's own stance in deprecated.rst. Add athe missing break to silence the warning. Fixes: 6e83985b ("powerpc/cbe: Do not process external or decremeter interrupts from sreset") Reported-by:Naresh Kamboju <naresh.kamboju@linaro.org> Signed-off-by:
Anders Roxell <anders.roxell@linaro.org> Reviewed-by:
Nathan Chancellor <nathan@kernel.org> Reviewed-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Xiang wangx authored
struct of_device_id should normally be const. Signed-off-by:
Xiang wangx <wangxiang@cdjrlc.com> Signed-off-by:
-
Christophe Leroy authored
copy_inst_from_kernel_nofault() uses copy_from_kernel_nofault() to copy one or two 32bits words. This means calling an out-of-line function which itself calls back copy_from_kernel_nofault_allowed() then performs a generic copy with loops. Rewrite copy_inst_from_kernel_nofault() to do everything at a single place and use __get_kernel_nofault() directly to perform single accesses without loops. Allthough the generic function uses pagefault_disable(), it is not required on powerpc because do_page_fault() bails earlier when a kernel mode fault happens on a kernel address. As the function has now become very small, inline it. With this change, on an 8xx the time spent in the loop in ftrace_replace_code() is reduced by 23% at function tracer activation and 27% at nop tracer activation. The overall time to activate function tracer (measured with shell command 'time') is 570ms before the patch and 470ms after the patch. Even vmlinux size is reduced (by 152 instruction). Before the patch: 00000018 <copy_inst_from_kernel_nofault>: 18: 94 21 ff e0 stwu r1,-32(r1) 1c: 7c 08 02 a6 mflr r0 20: 38 a0 00 04 li r5,4 24: 93 e1 00 1c stw r31,28(r1) 28: 7c 7f 1b 78 mr r31,r3 2c: 38 61 00 08 addi r3,r1,8 30: 90 01 00 24 stw r0,36(r1) 34: 48 00 00 01 bl 34 <copy_inst_from_kernel_nofault+0x1c> 34: R_PPC_REL24 copy_from_kernel_nofault 38: 2c 03 00 00 cmpwi r3,0 3c: 40 82 00 0c bne 48 <copy_inst_from_kernel_nofault+0x30> 40: 81 21 00 08 lwz r9,8(r1) 44: 91 3f 00 00 stw r9,0(r31) 48: 80 01 00 24 lwz r0,36(r1) 4c: 83 e1 00 1c lwz r31,28(r1) 50: 38 21 00 20 addi r1,r1,32 54: 7c 08 03 a6 mtlr r0 58: 4e 80 00 20 blr After the patch (before inlining): 00000018 <copy_inst_from_kernel_nofault>: 18: 3d 20 b0 00 lis r9,-20480 1c: 7c 04 48 40 cmplw r4,r9 20: 7c 69 1b 78 mr r9,r3 24: 41 80 00 14 blt 38 <copy_inst_from_kernel_nofault+0x20> 28: 81 44 00 00 lwz r10,0(r4) 2c: 38 60 00 00 li r3,0 30: 91 49 00 00 stw r10,0(r9) 34: 4e 80 00 20 blr 38: 38 60 ff de li r3,-34 3c: 4e 80 00 20 blr 40: 38 60 ff f2 li r3,-14 44: 4e 80 00 20 blr Signed-off-by:
-
Christophe Leroy authored
Because of circular inclusion of asm/hw_breakpoint.h, we need to move definition of asm/reg.h outside of inst.h so that asm/hw_breakpoint.h gets it without including asm/inst.h Also remove asm/inst.h from asm/uprobes.h as it's not needed anymore. Signed-off-by:
-
Christophe Leroy authored
Unlike PPC64 ABI, PPC32 uses the stack to pass a parameter defined as a struct, even when the struct has a single simple element. To avoid that, define ppc_inst_t as u32 on PPC32. Keep it as 'struct ppc_inst' when __CHECKER__ is defined so that sparse can perform type checking. Also revert commit 511eea5e ("powerpc/kprobes: Fix Oops by passing ppc_inst as a pointer to emulate_step() on ppc32") as now the instruction to be emulated is passed as a register to emulate_step(). Signed-off-by:
-
Christophe Leroy authored
In order to stop using 'struct ppc_inst' on PPC32, define a ppc_inst_t typedef. Signed-off-by:
-
Christophe Leroy authored
PPC64 version of ___get_user_instr() can be used for PPC32 as well, by simply disabling the suffix part with IS_ENABLED(CONFIG_PPC64). Signed-off-by:
-
Christophe Leroy authored
Today we have the following IBATs allocated: ---[ Instruction Block Address Translation ]--- 0: 0xc0000000-0xc03fffff 0x00000000 4M Kernel x m 1: 0xc0400000-0xc05fffff 0x00400000 2M Kernel x m 2: 0xc0600000-0xc06fffff 0x00600000 1M Kernel x m 3: 0xc0700000-0xc077ffff 0x00700000 512K Kernel x m 4: 0xc0780000-0xc079ffff 0x00780000 128K Kernel x m 5: 0xc07a0000-0xc07bffff 0x007a0000 128K Kernel x m 6: - 7: - The two 128K should be a single 256K instead. When _etext is not aligned to 128Kbytes, the system will allocate all necessary BATs to the lower 128Kbytes boundary, then allocate an additional 128Kbytes BAT for the remaining block. Instead, align the top to 128Kbytes so that the function directly allocates a 256Kbytes last block: ---[ Instruction Block Address Translation ]--- 0: 0xc0000000-0xc03fffff 0x00000000 4M Kernel x m 1: 0xc0400000-0xc05fffff 0x00400000 2M Kernel x m 2: 0xc0600000-0xc06fffff 0x00600000 1M Kernel x m 3: 0xc0700000-0xc077ffff 0x00700000 512K Kernel x m 4: 0xc0780000-0xc07bffff 0x00780000 256K Kernel x m 5: - 6: - 7: - Signed-off-by:
-
Christophe Leroy authored
All platforms now have KUAP and KUEP so remove CONFIG_PPC_HAVE_KUAP and CONFIG_PPC_HAVE_KUEP. Signed-off-by:
-
Christophe Leroy authored
This adds KUAP support to book3e/64. This is done by reading the content of SPRN_MAS1 and checking the TID at the time user pgtable is loaded. Signed-off-by:
-
Christophe Leroy authored
This adds KUAP support to 85xx in 32 bits mode. This is done by reading the content of SPRN_MAS1 and checking the TID at the time user pgtable is loaded. Signed-off-by:
-
Christophe Leroy authored
This adds KUAP support to 40x. This is done by checking the content of SPRN_PID at the time user pgtable is loaded. 40x doesn't have KUEP, but KUAP implies KUEP because when the PID doesn't match the page's PID, the page cannot be read nor executed. So KUEP is now automatically selected when KUAP is selected and disabled when KUAP is disabled. Signed-off-by:
-
Christophe Leroy authored
This adds KUAP support to 44x. This is done by checking the content of SPRN_PID at the time it is read and written into SPRN_MMUCR. Signed-off-by:
-
Christophe Leroy authored
On booke/40x we don't have segments like book3s/32. On booke/40x we don't have access protection groups like 8xx. Use the PID register to provide user access protection. Kernel address space can be accessed with any PID. User address space has to be accessed with the PID of the user. User PID is always not null. Everytime the kernel is entered, set PID register to 0 and restore PID register when returning to user. Everytime kernel needs to access user data, PID is restored for the access. In TLB miss handlers, check the PID and bail out to data storage exception when PID is 0 and accessed address is in user space. Note that also forbids execution of user text by kernel except when user access is unlocked. But this shouldn't be a problem as the kernel is not supposed to ever run user text. This patch prepares the infrastructure but the real activation of KUAP is done by following patches for each processor type one by one. Signed-off-by:
-
Christophe Leroy authored
PPC_KUAP_DEBUG is supported by all platforms doing PPC_KUAP, it doesn't depend on Radix on book3s/64. This will avoid adding one more dependency when implementing KUAP on book3e/64. Signed-off-by:
-
Christophe Leroy authored
Also call kuap_lock() and kuap_save_and_lock() from interrupt functions with CONFIG_PPC64. For book3s/64 we keep them empty as it is done in assembly. Also do the locked assert when switching task unless it is book3s/64. Signed-off-by:
-
Christophe Leroy authored
We have many functionnalities common to 40x and BOOKE, it leads to many places with #if defined(CONFIG_BOOKE) || defined(CONFIG_40x). We are going to add a few more with KUAP for booke/40x, so create a new symbol which is defined when either BOOKE or 40x is defined. Signed-off-by:
-
Christophe Leroy authored
In order to reuse it on booke/4xx, move KUAP setup routine out of 8xx.c Make them usable on SMP by removing the __init tag as it is called for each CPU. And use __prevent_user_access() instead of hard coding initial lock. Signed-off-by:
-
Christophe Leroy authored
Add kuap_lock() and call it when entering interrupts from user. It is called kuap_lock() as it is similar to kuap_save_and_lock() without the save. However book3s/32 already have a kuap_lock(). Rename it kuap_lock_addr(). Signed-off-by:
-
Christophe Leroy authored
__kuap_assert_locked() is redundant with __kuap_get_and_assert_locked(). Move the verification of CONFIG_PPC_KUAP_DEBUG in kuap_assert_locked() and make it call __kuap_get_and_assert_locked() directly. Signed-off-by:
-
Christophe Leroy authored
Today, every platform checks that KUAP is not de-activated before doing the real job. Move the verification out of platform specific functions. Signed-off-by:
-
Christophe Leroy authored
Make the following functions generic to all platforms. - bad_kuap_fault() - kuap_assert_locked() - kuap_save_and_lock() (PPC32 only) - kuap_kernel_restore() - kuap_get_and_assert_locked() And for all platforms except book3s/64 - allow_user_access() - prevent_user_access() - prevent_user_access_return() - restore_user_access() Prepend __ in front of the name of platform specific ones. For now the generic just calls the platform specific, but next patch will move redundant parts of specific functions into the generic one. Signed-off-by:
-
Christophe Leroy authored
Deactivating KUEP at boot time is unrelevant for PPC32 and BOOK3E/64. Remove it. It allows to refactor setup_kuep() via a __weak function that only PPC64s will overide for now. Signed-off-by:
-
Christophe Leroy authored
Calling 'mfsr' to get the content of segment registers is heavy, in addition it requires clearing of the 'reserved' bits. In order to avoid this operation, save it in mm context and in thread struct. The saved sr0 is the one used by kernel, this means that on locking entry it can be used as is. For unlocking, the only thing to do is to clear SR_NX. This improves null_syscall selftest by 12 cycles, ie 4%. Signed-off-by:
-
