linux-kernel - Re: [PATCH] uprobes: list all active uprobes in the system

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: <f4424d36-b18e-d801-8657-85565817554b@huawei.com>
Date:   Fri, 13 Jan 2023 15:49:34 +0800
From:   Longjun Luo <luolongjun@...wei.com>
To:     Ravi Bangoria <ravi.bangoria@....com>, <peterz@...radead.org>,
        <mingo@...hat.com>, <acme@...nel.org>, <mark.rutland@....com>,
        <alexander.shishkin@...ux.intel.com>, <jolsa@...nel.org>,
        <namhyung@...nel.org>
CC:     <linux-perf-users@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <sangyan@...wei.com>, <luanjianhai@...wei.com>,
        <zhuling8@...wei.com>, <lizongwu@...wei.com>,
        <luolongjuna@...il.com>, <rostedt@...dmis.org>,
        <mhiramat@...nel.org>, <oleg@...hat.com>
Subject: Re: [PATCH] uprobes: list all active uprobes in the system


在 1/13/2023 12:03 PM, Ravi Bangoria 写道:
> +cc: Steven, Oleg, Masami
>
> On 11-Jan-23 4:18 PM, luolongjun@...wei.com wrote:
>> From: Longjun Luo <luolongjun@...wei.com>
>>
>> Since uprobes will replace instructions in the process
>> memory, it is necessary to provide one way to list
>> all active uprobes. One can access this file through
>> /sys/kernel/debug/uprobes/list.
>>
>> Output looks like this
>> =====================
>> inode+offset/ref_ctr_offset    4160760670+0x30a10/0x0
>>      ref                        1
>>      register_rwsem             (unlocked)
>>      consumer_rwsem             (unlocked)
>>      consumers-1
>>          handler                0xffffffffc03ee02e(handler) uprobe
>>          ret_handler            0x0000000000000000( )
>>          filter                 0xffffffffc03ee010(filter) uprobe
>>      consumers-2
>>          handler                0xffffffffc03e902e(handler_1) uprobe_1
>>          ret_handler            0x0000000000000000( )
>>          filter                 0xffffffffc03e9010(filter_1) uprobe_1
> You can list uprobes via /sys/kernel/tracing/uprobe_events. Any specific reason
> to introduce another way?
>
> Thanks,
> Ravi

/sys/kernel/tracing/uprobe_events only lists tracing events, not all uprobe events.
For example, if I write a kernel module to call the uprobe_register to add a new uprobe event, it will not be listed by /sys/kernel/tracing/uprobe_events.

Compare it with the kprobe is like:
/sys/kernel/tracing/uprobe_events -> /sys/kernel/tracing/kprobe_events
/sys/kernel/debug/uprobes/list -> /sys/kernel/debug/kprobes/list

Nowadays, only the trace system uses the uprobe, so /sys/kernel/tracing/uprobe_events is enough.

But we try to extend the usage of the uprobe, and we add many uprobe events by calling the uprobe_register.

For this situation, we need a approach to see all the uprobe events in the kernel.