linux-kernel - Re: [PATCH V9 10/10] arm64/perf: Implement branch records save on PMU IRQ

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 23 May 2023 15:39:01 +0100
From:   James Clark <james.clark@....com>
To:     Anshuman Khandual <anshuman.khandual@....com>,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        will@...nel.org, catalin.marinas@....com, mark.rutland@....com
Cc:     Mark Brown <broonie@...nel.org>, Rob Herring <robh@...nel.org>,
        Marc Zyngier <maz@...nel.org>,
        Suzuki Poulose <suzuki.poulose@....com>,
        Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...hat.com>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        linux-perf-users@...r.kernel.org
Subject: Re: [PATCH V9 10/10] arm64/perf: Implement branch records save on PMU
 IRQ



On 15/03/2023 05:14, Anshuman Khandual wrote:
> This modifies armv8pmu_branch_read() to concatenate live entries along with
> task context stored entries and then process the resultant buffer to create
> perf branch entry array for perf_sample_data. It follows the same principle
> like task sched out.
> 
> Cc: Catalin Marinas <catalin.marinas@....com>
> Cc: Will Deacon <will@...nel.org>
> Cc: Mark Rutland <mark.rutland@....com>
> Cc: linux-arm-kernel@...ts.infradead.org
> Cc: linux-kernel@...r.kernel.org
> Signed-off-by: Anshuman Khandual <anshuman.khandual@....com>
> ---

[...]

>  void armv8pmu_branch_read(struct pmu_hw_events *cpuc, struct perf_event *event)
>  {
>  	struct brbe_hw_attr *brbe_attr = (struct brbe_hw_attr *)cpuc->percpu_pmu->private;
> +	struct arm64_perf_task_context *task_ctx = event->pmu_ctx->task_ctx_data;
> +	struct brbe_regset live[BRBE_MAX_ENTRIES];
> +	int nr_live, nr_store;
>  	u64 brbfcr, brbcr;
> -	int idx, loop1_idx1, loop1_idx2, loop2_idx1, loop2_idx2, count;
>  
>  	brbcr = read_sysreg_s(SYS_BRBCR_EL1);
>  	brbfcr = read_sysreg_s(SYS_BRBFCR_EL1);
> @@ -739,36 +743,13 @@ void armv8pmu_branch_read(struct pmu_hw_events *cpuc, struct perf_event *event)
>  	write_sysreg_s(brbfcr | BRBFCR_EL1_PAUSED, SYS_BRBFCR_EL1);
>  	isb();
>  
> -	/* Determine the indices for each loop */
> -	loop1_idx1 = BRBE_BANK0_IDX_MIN;
> -	if (brbe_attr->brbe_nr <= BRBE_BANK_MAX_ENTRIES) {
> -		loop1_idx2 = brbe_attr->brbe_nr - 1;
> -		loop2_idx1 = BRBE_BANK1_IDX_MIN;
> -		loop2_idx2 = BRBE_BANK0_IDX_MAX;
> -	} else {
> -		loop1_idx2 = BRBE_BANK0_IDX_MAX;
> -		loop2_idx1 = BRBE_BANK1_IDX_MIN;
> -		loop2_idx2 = brbe_attr->brbe_nr - 1;
> -	}
> -
> -	/* Loop through bank 0 */
> -	select_brbe_bank(BRBE_BANK_IDX_0);
> -	for (idx = 0, count = loop1_idx1; count <= loop1_idx2; idx++, count++) {
> -		if (!capture_branch_entry(cpuc, event, idx))
> -			goto skip_bank_1;
> -	}
> -
> -	/* Loop through bank 1 */
> -	select_brbe_bank(BRBE_BANK_IDX_1);
> -	for (count = loop2_idx1; count <= loop2_idx2; idx++, count++) {
> -		if (!capture_branch_entry(cpuc, event, idx))
> -			break;
> -	}
> -
> -skip_bank_1:
> -	cpuc->branches->branch_stack.nr = idx;
> -	cpuc->branches->branch_stack.hw_idx = -1ULL;
> +	nr_live = capture_brbe_regset(brbe_attr, live);
> +	nr_store = task_ctx->nr_brbe_records;
> +	nr_store = stitch_stored_live_entries(task_ctx->store, live, nr_store,
> +					      nr_live, brbe_attr->brbe_nr);
> +	process_branch_entries(cpuc, event, task_ctx->store, nr_store);

Hi Anshuman,

With the following command I get a crash:

  perf record --branch-filter any,save_type -a -- ls

[  101.171822] Unable to handle kernel NULL pointer dereference at
virtual address 0000000000000600
...
[145380.414654] Call trace:
[145380.414739]  armv8pmu_branch_read+0x7c/0x578
[145380.414895]  armv8pmu_handle_irq+0x104/0x1c0
[145380.415043]  armpmu_dispatch_irq+0x38/0x70
[145380.415209]  __handle_irq_event_percpu+0x124/0x3b8
[145380.415392]  handle_irq_event+0x54/0xc8
[145380.415567]  handle_fasteoi_irq+0x100/0x1e0
[145380.415718]  generic_handle_domain_irq+0x38/0x58
[145380.415895]  gic_handle_irq+0x5c/0x130
[145380.416025]  call_on_irq_stack+0x24/0x58
[145380.416173]  el1_interrupt+0x74/0xc0
[145380.416321]  el1h_64_irq_handler+0x18/0x28
[145380.416475]  el1h_64_irq+0x64/0x68
[145380.416604]  smp_call_function_single+0xe8/0x1f0
[145380.416745]  event_function_call+0xbc/0x1c8
[145380.416919]  _perf_event_enable+0x84/0xa0
[145380.417069]  perf_ioctl+0xe8/0xd68
[145380.417204]  __arm64_sys_ioctl+0x9c/0xe0
[145380.417353]  invoke_syscall+0x4c/0x120
[145380.417523]  el0_svc_common+0xd0/0x120
[145380.417693]  do_el0_svc+0x3c/0xb8
[145380.417859]  el0_svc+0x50/0xc0
[145380.418004]  el0t_64_sync_handler+0x84/0xf0
[145380.418160]  el0t_64_sync+0x190/0x198

When using --branch-filter any,u without -a it seems to be fine so could
be that task_ctx is null in per-cpu mode, or something to do with the
userspace only flag?

I'm also wondering if it's possible to collapse some of the last 5
commits? They seem to mostly modify things in brbe.c which is a new file
so the history probably isn't important at this point it just makes it a
bit harder to review.

>  	process_branch_aborts(cpuc);
> +	task_ctx->nr_brbe_records = 0;
>  
>  	/* Unpause the buffer */
>  	write_sysreg_s(brbfcr & ~BRBFCR_EL1_PAUSED, SYS_BRBFCR_EL1);