linux-kernel - Re: perf/tracepoint: another fuzzer generated lockup

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]

Message-ID: <20131109002508.GG14606@localhost.localdomain>
Date:	Sat, 9 Nov 2013 01:25:10 +0100
From:	Frederic Weisbecker <fweisbec@...il.com>
To:	Vince Weaver <vincent.weaver@...ne.edu>
Cc:	Steven Rostedt <rostedt@...dmis.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Ingo Molnar <mingo@...nel.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Dave Jones <davej@...hat.com>
Subject: Re: perf/tracepoint: another fuzzer generated lockup

On Fri, Nov 08, 2013 at 04:15:21PM -0500, Vince Weaver wrote:
> 
> int main(int argc, char **argv) {
> /* 1 */
> /* fd = 82 */
> 
> 	memset(&pe[82],0,sizeof(struct perf_event_attr));
> 	pe[82].type=PERF_TYPE_TRACEPOINT;
> 	pe[82].size=80;
> 	pe[82].config=0x18;

I did some more testing and 0x18 on my setup is the irq_work_exit event. I tried a few
other tracepoints (tried the timer interrupt for example) but they did not trigger the issue.

And since the irq work handles the pending disable and wakeup, maybe that fiddles with some
perf state and make the NMI handler run into some buggy loop.

Perhaps that's triggered with some help of the disable ioctl.

I don't know, my bed is calling so I'll test that further tomorrow or so.

Thanks.

> 	pe[82].sample_period=0xdd95cac7b947b610;
> 	pe[82].sample_type=PERF_SAMPLE_IP|PERF_SAMPLE_TIME|PERF_SAMPLE_READ|PERF_SAMPLE_CALLCHAIN|PERF_SAMPLE_PERIOD|PERF_SAMPLE_DATA_SRC;
> //|PERF_SAMPLE_IDENTIFIER; /* 18135 */
> 	pe[82].read_format=PERF_FORMAT_TOTAL_TIME_ENABLED|PERF_FORMAT_TOTAL_TIME_RUNNING|PERF_FORMAT_ID|PERF_FORMAT_GROUP; /* f */
> 	pe[82].exclude_hv=1;
> 	pe[82].mmap=1;
> 	pe[82].comm=1;
> 	pe[82].watermark=1;
> 	pe[82].precise_ip=1; /* constant skid */
> 	pe[82].mmap_data=1;
> 	pe[82].exclude_host=1;
> 	pe[82].exclude_guest=1;
> 	pe[82].exclude_callchain_user=1;
> 	pe[82].wakeup_watermark=232;
> 	pe[82].bp_type=HW_BREAKPOINT_EMPTY;
> 	pe[82].config1=0x9;
> 	pe[82].branch_sample_type=PERF_SAMPLE_BRANCH_USER|PERF_SAMPLE_BRANCH_KERNEL|PERF_SAMPLE_BRANCH_HV|PERF_SAMPLE_BRANCH_ANY|PERF_SAMPLE_BRANCH_ANY_CALL|PERF_SAMPLE_BRANCH_ANY_RETURN|PERF_SAMPLE_BRANCH_IND_CALL;
> //|PERF_SAMPLE_BRANCH_ABORT_TX;
> 	pe[82].sample_regs_user=39;
> 	pe[82].sample_stack_user=37552;
> 
> 	fd[82]=perf_event_open(&pe[82],
> 				0, /* current thread */
> 				-1, /* all cpus */
> 				-1, /* New Group Leader */
> 				PERF_FLAG_FD_OUTPUT /*2*/ );
> 
> 
> /* 2 */
> 	mmap_result[82]=mmap(NULL, 20480,PROT_READ|PROT_WRITE, MAP_SHARED,fd[82], 0);
> /* 3 */
> 	memset(&sa, 0, sizeof(struct sigaction));
> 	sa.sa_sigaction = our_handler;
> 	sa.sa_flags = SA_SIGINFO;
> 	if (sigaction( SIGRTMIN+2, &sa, NULL) < 0) {
> 		printf("Error setting up signal handler\n");
> 	}
> 	fcntl(fd[82], F_SETFL, O_RDWR|O_NONBLOCK|O_ASYNC);
> 	fcntl(fd[82], F_SETSIG, SIGRTMIN+2);
> 	fcntl(fd[82], F_SETOWN,getpid());
> 	/* Replayed 3 syscalls */
> 	return 0;
> }

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/