From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4D3193A4502 for ; Mon, 13 Apr 2026 08:19:29 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776068374; cv=none; b=DmaNxlXDokdKlOPEDb3+T03zW5bgXk+KTVA3L1bxqJA31r+rhdmPbuYQrJxnL1m2YmGz4qnmcI7OafMSu1oNQ7SWRsqqOcnSiu+50fz31OkpWSbsX1rTaMXb/461YwSTQIoo1H485x4Ta5rdiL1DdFAHRlbHyOuzVnx5JrEvgI0= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776068374; c=relaxed/simple; bh=V60/M9mc6y8UfqWngghQt4RJRH0uYNAYylg6qSekgww=; h=Message-ID:Subject:From:To:Cc:Date:In-Reply-To:References: Content-Type:MIME-Version; b=r1puvkPIjLJlqJmKL1jZB644Gyv2OU2dcB0J29jpxXq/ySFfQNsiD61vZQIE9ng4NKvkgKjyBQu0Dbitk76kOt29u9iMRVhF08ug7I1GfXuqS0UFjA0jYTJEM7guAbyuTzkb9+1AOUG4LTxsxC8GeHvuJVUNWc22MyZnlwpW100= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=N8xSzzJx; dkim=pass (2048-bit key) header.d=redhat.com header.i=@redhat.com header.b=HbWMDctV; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="N8xSzzJx"; dkim=pass (2048-bit key) header.d=redhat.com header.i=@redhat.com header.b="HbWMDctV" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1776068368; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:autocrypt:autocrypt; bh=NcrSdL+6/tRRW05NAZ6inc2WHBwAEQDecnXM8gaacXk=; b=N8xSzzJxlJd5ZiRFkhcl/FB1Xa2Efs4s/tJM2OHrBNZuyDIPd1BiWLZ56nOUvsYBPM98l8 TqjuQdh+aonT6N44SOiZ5StsdfDCr4bc7pfygse/LuwYxZCdFQN2SQg/8GCW6bNCjceVfD kEWd/TNFlL64eRDxhQuQWT+5MXc+UuQ= Received: from mail-wr1-f69.google.com (mail-wr1-f69.google.com [209.85.221.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-595-xLLAkxdKM62eLfQLeDz_AQ-1; Mon, 13 Apr 2026 04:19:26 -0400 X-MC-Unique: xLLAkxdKM62eLfQLeDz_AQ-1 X-Mimecast-MFC-AGG-ID: xLLAkxdKM62eLfQLeDz_AQ_1776068365 Received: by mail-wr1-f69.google.com with SMTP id ffacd0b85a97d-43d1dea12aaso2987068f8f.1 for ; Mon, 13 Apr 2026 01:19:26 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=google; t=1776068365; x=1776673165; darn=vger.kernel.org; h=mime-version:user-agent:content-transfer-encoding:autocrypt :references:in-reply-to:date:cc:to:from:subject:message-id:from:to :cc:subject:date:message-id:reply-to; bh=NcrSdL+6/tRRW05NAZ6inc2WHBwAEQDecnXM8gaacXk=; b=HbWMDctVm20iIzIBQcTSk93hL6fnLjQqBcWXD8J+AGlQQ69XUbh8Qm1RCdGWWUiQLr 8akP+NcEXKhQAUBHi23ANvH8a4UBJmWtNqZb9fSXYQ8ZOLByKbmJBOM8jd8uVWmbdb2/ Oao0csEO922/58GjSh9clbgG4z9Kr6Csf41iFZJzC6qgUxyu1y2bZG8tBktgHUNm2CsT F7wBwtJTCojtgmsp0vf3kW9RmCXgOeSDP/rh8/RLMvhQ5Nkm/BOaNgCO3HGBipmoDkGO D0uoIgP7oZ3LLYeeoS5BNij4MWX8NBz0A7PutcUIzK0PkcyNRrmFB4Smn2/MzU3yHpHf AqJg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1776068365; x=1776673165; h=mime-version:user-agent:content-transfer-encoding:autocrypt :references:in-reply-to:date:cc:to:from:subject:message-id:x-gm-gg :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=NcrSdL+6/tRRW05NAZ6inc2WHBwAEQDecnXM8gaacXk=; b=VtLzSprGPFFP/l2UL6T/8zvV/xj/s0JxD+8MEGzJwyzSmqVT+GKs88oCEtYuIIbYY/ J4yDEZiKFPvUmGq5XjzdeSS+xFUgEW3sku7/I6fuFNPkA9ZdF49YPcMW1zNxL9EFGgjU IWTVOlGlTVphwUmk1YyXx9ODjeyVyqyuy3ZGLwwDxEAY5gv7Tmq8Wj8vDz72SAWWIDVb S6vL5kCgTFZCvBxz522KwNzuRD0+akW2XzpwnmKcYWZ3QD0ui3H+R5wjBs0bR6JuFNLe /zR5MjUo+7oCTTwpZwAGYuAtOM5gGMfX73n1BO0ykstTb7pw568V9AcyStPKL9TRzZcA h5AA== X-Forwarded-Encrypted: i=1; AFNElJ8g0oX20b44zUTfB9z9c+GjeZf2NlpwN5TAos3ilEmHyhiQ92jqZqjZirkHrF9La/W6qQskeiE3c6oaYdY=@vger.kernel.org X-Gm-Message-State: AOJu0YwZmqX5Y3TTxdhXyc1qRBnpcjp1tDsCrU1WLf23r6XQPGxbeLET C1UYpG1z6L4jf4/+AH7U7Y0ZG06ue8JJ96cL/bISgREm1Ebxi5cUcCLPzxdl4TJZZdOnLJS+tjn S6KL681NrxW+YvhWfkX9r36WcRCqEL9XEp2ZIV45S2s5u+ZY3hg3kvl6iIHrYEhghPQTnd5IrTd OQ X-Gm-Gg: AeBDietHhESimoNj3URa7l6XK9qimcvxQ7ot1DRPOUPYin9Ed3QUP5GQAs8uuO3GgSv 799yn4MnyuZ09pgzZegt4HYkPH3BSNOYNNSHfuDm5m7PK8Vup5evIzJpb4zI/xyz3EF9R9MQG8G +qXhDKSw5EwPtKLD+u5nKoceuQVtUlu0EdsPga1H5GN6qZOFcFat8KEKs0NaOuuC6UpiUcZERCp YhQOce8B9UL7xF5rNJlY6NEN5csPao0KOTFcla9SWRU8ZJYGnOG896bH5nv/hUt3iaZ6FIg1Ays PXTPlfKlhBHiV0yktKsc46FH5VNYRugfle3PZrsau5545S3kg58N4vZuvVNNMmduDhQVzxNuycT myRuRYxOi8pzaJdUSjN5o1u0dI1qSwYodiijAzhlkkvqWPaiW2vQiKSkwRqVEWijUzCLKDnt3Dm 1Yh/RQNDjQ0P5KGyo= X-Received: by 2002:a5d:64c4:0:b0:43b:498f:dceb with SMTP id ffacd0b85a97d-43d6427bac4mr15843313f8f.9.1776068363070; Mon, 13 Apr 2026 01:19:23 -0700 (PDT) X-Received: by 2002:a5d:64c4:0:b0:43b:498f:dceb with SMTP id ffacd0b85a97d-43d6427bac4mr15843051f8f.9.1776068359558; Mon, 13 Apr 2026 01:19:19 -0700 (PDT) Received: from gmonaco-thinkpadt14gen3.rmtit.csb (212-8-243-115.hosted-by-worldstream.net. [212.8.243.115]) by smtp.gmail.com with ESMTPSA id ffacd0b85a97d-43d762decf6sm15365087f8f.8.2026.04.13.01.19.17 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 13 Apr 2026 01:19:18 -0700 (PDT) Message-ID: <74a624434b59c00f9407909b8696f041536d9418.camel@redhat.com> Subject: Re: [RFC PATCH 2/4] rv/tlob: Add tlob deterministic automaton monitor From: Gabriele Monaco To: wen.yang@linux.dev Cc: Steven Rostedt , Masami Hiramatsu , Mathieu Desnoyers , linux-trace-kernel@vger.kernel.org, linux-kernel@vger.kernel.org Date: Mon, 13 Apr 2026 10:19:17 +0200 In-Reply-To: References: Autocrypt: addr=gmonaco@redhat.com; prefer-encrypt=mutual; keydata=mDMEZuK5YxYJKwYBBAHaRw8BAQdAmJ3dM9Sz6/Hodu33Qrf8QH2bNeNbOikqYtxWFLVm0 1a0JEdhYnJpZWxlIE1vbmFjbyA8Z21vbmFjb0BrZXJuZWwub3JnPoiZBBMWCgBBFiEEysoR+AuB3R Zwp6j270psSVh4TfIFAmjKX2MCGwMFCQWjmoAFCwkIBwICIgIGFQoJCAsCBBYCAwECHgcCF4AACgk Q70psSVh4TfIQuAD+JulczTN6l7oJjyroySU55Fbjdvo52xiYYlMjPG7dCTsBAMFI7dSL5zg98I+8 cXY1J7kyNsY6/dcipqBM4RMaxXsOtCRHYWJyaWVsZSBNb25hY28gPGdtb25hY29AcmVkaGF0LmNvb T6InAQTFgoARAIbAwUJBaOagAULCQgHAgIiAgYVCgkICwIEFgIDAQIeBwIXgBYhBMrKEfgLgd0WcK eo9u9KbElYeE3yBQJoymCyAhkBAAoJEO9KbElYeE3yjX4BAJ/ETNnlHn8OjZPT77xGmal9kbT1bC1 7DfrYVISWV2Y1AP9HdAMhWNAvtCtN2S1beYjNybuK6IzWYcFfeOV+OBWRDQ== Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.58.3 (3.58.3-1.fc43) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 On Mon, 2026-04-13 at 03:27 +0800, wen.yang@linux.dev wrote: > From: Wen Yang >=20 > Add the tlob (task latency over budget) RV monitor. tlob tracks the > monotonic elapsed time (CLOCK_MONOTONIC) of a marked per-task code > path, including time off-CPU, and fires a per-task hrtimer when the > elapsed time exceeds a configurable budget. >=20 > Three-state DA (unmonitored/on_cpu/off_cpu) driven by trace_start, > switch_in/out, and budget_expired events. Per-task state lives in a > fixed-size hash table (TLOB_MAX_MONITORED slots) with RCU-deferred > free. >=20 > Two userspace interfaces: > =C2=A0- tracefs: uprobe pair registration via the monitor file using the > =C2=A0=C2=A0 format "pid:threshold_us:offset_start:offset_stop:binary_pat= h" > =C2=A0- /dev/rv ioctls (CONFIG_RV_CHARDEV): TLOB_IOCTL_TRACE_START / > =C2=A0=C2=A0 TRACE_STOP; TRACE_STOP returns -EOVERFLOW on violation >=20 > Each /dev/rv fd has a per-fd mmap ring buffer (physically contiguous > pages). A control page (struct tlob_mmap_page) at offset 0 exposes > head/tail/dropped for lockless userspace reads; struct tlob_event > records follow at data_offset. Drop-new policy on overflow. >=20 > UAPI: include/uapi/linux/rv.h (tlob_start_args, tlob_event, > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 tlob_mmap_page, ioctl numbers), monitor_tl= ob.rst, > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ioctl-number.rst (RV_IOC_MAGIC=3D0xB9). >=20 I'm not fully grasping all the requirements for the monitors yet, but I see= you are reimplementing a lot of functionality in the monitor itself rather than within RV, let's see if we can consolidate some of them: * you're using timer expirations, can we do it with timed automata? [1] * RV automata usually don't have an /unmonitored/ state, your trace_start = event would be the start condition (da_event_start) and the monitor will get non= - running at each violation (it calls da_monitor_reset() automatically), all setup/cleanup logic should be handled implicitly within RV. I believe that = would also save you that ugly trace_event_tlob() redefinition. * you're maintaining a local hash table for each task_struct, that could u= se the per-object monitors [2] where your "object" is in fact your struct, allocated when you start the monitor with all appropriate fields and indexe= d by pid * you are handling violations manually, considering timed automata trigger= a full fledged violation on timeouts, can you use the RV-way (error tracepoin= ts or reactors only)? Do you need the additional reporting within the tracepoint/ioctl? Cannot the userspace consumer desume all those from other events and let RV do just the monitoring? * I like the uprobe thing, we could probably move all that to a common hel= per once we figure out how to make it generic. Note: [1] and [2] didn't reach upstream yet, but should reach linux-next so= on. Thanks, Gabriele [1] - https://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace.git/commi= t/?h=3Drv/for-next&id=3Df5587d1b6ec938afb2f74fe399a68020d66923e4 [2] - https://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace.git/commi= t/?h=3Drv/for-next&id=3Dda282bf7fadb095ee0a40c32ff0126429c769b45 > Signed-off-by: Wen Yang > --- > =C2=A0Documentation/trace/rv/index.rst=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 1 + > =C2=A0Documentation/trace/rv/monitor_tlob.rst=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 | 381 +++++++ > =C2=A0.../userspace-api/ioctl/ioctl-number.rst=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0 |=C2=A0=C2=A0 1 + > =C2=A0include/uapi/linux/rv.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 | 181 ++++ > =C2=A0kernel/trace/rv/Kconfig=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 |=C2=A0 17 + > =C2=A0kernel/trace/rv/Makefile=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 |=C2=A0=C2=A0 2 + > =C2=A0kernel/trace/rv/monitors/tlob/Kconfig=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 |=C2=A0 51 + > =C2=A0kernel/trace/rv/monitors/tlob/tlob.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 | 986 ++++++++++++++++++ > =C2=A0kernel/trace/rv/monitors/tlob/tlob.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 | 145 +++ > =C2=A0kernel/trace/rv/monitors/tlob/tlob_trace.h=C2=A0=C2=A0=C2=A0 |=C2= =A0 42 + > =C2=A0kernel/trace/rv/rv.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 4 + > =C2=A0kernel/trace/rv/rv_dev.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 | 602 +++++++++++ > =C2=A0kernel/trace/rv/rv_trace.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= |=C2=A0 50 + > =C2=A013 files changed, 2463 insertions(+) > =C2=A0create mode 100644 Documentation/trace/rv/monitor_tlob.rst > =C2=A0create mode 100644 include/uapi/linux/rv.h > =C2=A0create mode 100644 kernel/trace/rv/monitors/tlob/Kconfig > =C2=A0create mode 100644 kernel/trace/rv/monitors/tlob/tlob.c > =C2=A0create mode 100644 kernel/trace/rv/monitors/tlob/tlob.h > =C2=A0create mode 100644 kernel/trace/rv/monitors/tlob/tlob_trace.h > =C2=A0create mode 100644 kernel/trace/rv/rv_dev.c >=20 > diff --git a/Documentation/trace/rv/index.rst > b/Documentation/trace/rv/index.rst > index a2812ac5c..4f2bfaf38 100644 > --- a/Documentation/trace/rv/index.rst > +++ b/Documentation/trace/rv/index.rst > @@ -15,3 +15,4 @@ Runtime Verification > =C2=A0=C2=A0=C2=A0 monitor_wwnr.rst > =C2=A0=C2=A0=C2=A0 monitor_sched.rst > =C2=A0=C2=A0=C2=A0 monitor_rtapp.rst > +=C2=A0=C2=A0 monitor_tlob.rst > diff --git a/Documentation/trace/rv/monitor_tlob.rst > b/Documentation/trace/rv/monitor_tlob.rst > new file mode 100644 > index 000000000..d498e9894 > --- /dev/null > +++ b/Documentation/trace/rv/monitor_tlob.rst > @@ -0,0 +1,381 @@ > +.. SPDX-License-Identifier: GPL-2.0 > + > +Monitor tlob > +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > + > +- Name: tlob - task latency over budget > +- Type: per-task deterministic automaton > +- Author: Wen Yang > + > +Description > +----------- > + > +The tlob monitor tracks per-task elapsed time (CLOCK_MONOTONIC, includin= g > +both on-CPU and off-CPU time) and reports a violation when the monitored > +task exceeds a configurable latency budget threshold. > + > +The monitor implements a three-state deterministic automaton:: > + > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 | (initial) > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 v > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +--------------+ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +-------> | unmon= itored=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +--------------+ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 trace_start > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= v > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +--------------+ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0=C2=A0=C2= =A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +--------------+ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 switch_ou= t|=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 | trace_stop / budget_ex= pired > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 v=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 v > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 +--------= ------+=C2=A0 (unmonitored) > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 |=C2=A0= =C2=A0 off_cpu=C2=A0=C2=A0=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 +--------= ------+ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0 |=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0 | switch_in| trace_stop / budget_expired > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0=C2= =A0=C2=A0 v=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 v > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 (on_cpu)= =C2=A0 (unmonitored) > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 | > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 +-- trace_stop (f= rom on_cpu or off_cpu) > + > +=C2=A0 Key transitions: > +=C2=A0=C2=A0=C2=A0 unmonitored=C2=A0=C2=A0 --(trace_start)-->=C2=A0=C2= =A0 on_cpu > +=C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(s= witch_out)-->=C2=A0=C2=A0=C2=A0 off_cpu > +=C2=A0=C2=A0=C2=A0 off_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(switch= _in)-->=C2=A0=C2=A0=C2=A0=C2=A0 on_cpu > +=C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(t= race_stop)-->=C2=A0=C2=A0=C2=A0 unmonitored > +=C2=A0=C2=A0=C2=A0 off_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(trace_= stop)-->=C2=A0=C2=A0=C2=A0 unmonitored > +=C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(b= udget_expired)-> unmonitored=C2=A0=C2=A0 [violation] > +=C2=A0=C2=A0=C2=A0 off_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 --(budget= _expired)-> unmonitored=C2=A0=C2=A0 [violation] > + > +=C2=A0 sched_wakeup self-loops in on_cpu and unmonitored; switch_out and > +=C2=A0 sched_wakeup self-loop in off_cpu.=C2=A0 budget_expired is fired = by the one-shot > hrtimer; it always > +=C2=A0 transitions to unmonitored regardless of whether the task is on-C= PU > +=C2=A0 or off-CPU when the timer fires. > + > +State Descriptions > +------------------ > + > +- **unmonitored**: Task is not being traced.=C2=A0 Scheduling events > +=C2=A0 (``switch_in``, ``switch_out``, ``sched_wakeup``) are silently > +=C2=A0 ignored (self-loop).=C2=A0 The monitor waits for a ``trace_start`= ` event > +=C2=A0 to begin a new observation window. > + > +- **on_cpu**: Task is running on the CPU with the deadline timer armed. > +=C2=A0 A one-shot hrtimer was set for ``threshold_us`` microseconds at > +=C2=A0 ``trace_start`` time.=C2=A0 A ``switch_out`` event transitions to > +=C2=A0 ``off_cpu``; the hrtimer keeps running (off-CPU time counts towar= d > +=C2=A0 the budget).=C2=A0 A ``trace_stop`` cancels the timer and returns= to > +=C2=A0 ``unmonitored`` (normal completion).=C2=A0 If the hrtimer fires > +=C2=A0 (``budget_expired``) the violation is recorded and the automaton > +=C2=A0 transitions to ``unmonitored``. > + > +- **off_cpu**: Task was preempted or blocked.=C2=A0 The one-shot hrtimer > +=C2=A0 continues to run.=C2=A0 A ``switch_in`` event returns to ``on_cpu= ``. > +=C2=A0 A ``trace_stop`` cancels the timer and returns to ``unmonitored``= . > +=C2=A0 If the hrtimer fires (``budget_expired``) while the task is off-C= PU, > +=C2=A0 the violation is recorded and the automaton transitions to > +=C2=A0 ``unmonitored``. > + > +Rationale > +--------- > + > +The per-task latency budget threshold allows operators to express timing > +requirements in microseconds and receive an immediate ftrace event when = a > +task exceeds its budget.=C2=A0 This is useful for real-time tasks > +(``SCHED_FIFO`` / ``SCHED_DEADLINE``) where total elapsed time must > +remain within a known bound. > + > +Each task has an independent threshold, so up to ``TLOB_MAX_MONITORED`` > +(64) tasks with different timing requirements can be monitored > +simultaneously. > + > +On threshold violation the automaton records a ``tlob_budget_exceeded`` > +ftrace event carrying the final on-CPU / off-CPU time breakdown, but doe= s > +not kill or throttle the task.=C2=A0 Monitoring can be restarted by issu= ing a > +new ``trace_start`` event (or a new ``TLOB_IOCTL_TRACE_START`` ioctl). > + > +A per-task one-shot hrtimer is armed at ``trace_start`` for exactly > +``threshold_us`` microseconds.=C2=A0 It fires at most once per monitorin= g > +window, performs an O(1) hash lookup, records the violation, and injects > +the ``budget_expired`` event into the DA.=C2=A0 When ``CONFIG_RV_MON_TLO= B`` > +is not set there is zero runtime cost. > + > +Usage > +----- > + > +tracefs interface (uprobe-based external monitoring) > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > + > +The ``monitor`` tracefs file allows any privileged user to instrument an > +unmodified binary via uprobes, without changing its source code.=C2=A0 W= rite a > +four-field record to attach two plain entry uprobes: one at > +``offset_start`` fires ``tlob_start_task()`` and one at ``offset_stop`` > +fires ``tlob_stop_task()``, so the latency budget covers exactly the cod= e > +region between the two offsets:: > + > +=C2=A0 threshold_us:offset_start:offset_stop:binary_path > + > +``binary_path`` comes last so it may freely contain ``:`` (e.g. paths > +inside a container namespace). > + > +The uprobes fire for every task that executes the probed instruction in > +the binary, consistent with the native uprobe semantics.=C2=A0 All tasks= that > +execute the code region get independent per-task monitoring slots. > + > +Using two plain entry uprobes (rather than a uretprobe for the stop) mea= ns > +that a mistyped offset can never corrupt the call stack; the worst outco= me > +of a bad ``offset_stop`` is a missed stop that causes the hrtimer to fir= e > +and report a budget violation. > + > +Example=C2=A0 --=C2=A0 monitor a code region in ``/usr/bin/myapp`` with = a 5 ms > +budget, where the region starts at offset 0x12a0 and ends at 0x12f0:: > + > +=C2=A0 echo 1 > /sys/kernel/tracing/rv/monitors/tlob/enable > + > +=C2=A0 # Bind uprobes: start probe starts the clock, stop probe stops it > +=C2=A0 echo "5000:0x12a0:0x12f0:/usr/bin/myapp" \ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > /sys/kernel/tracing/rv/monitors/tlob/mo= nitor > + > +=C2=A0 # Remove the uprobe binding for this code region > +=C2=A0 echo "-0x12a0:/usr/bin/myapp" > > /sys/kernel/tracing/rv/monitors/tlob/monitor > + > +=C2=A0 # List registered uprobe bindings (mirrors the write format) > +=C2=A0 cat /sys/kernel/tracing/rv/monitors/tlob/monitor > +=C2=A0 # -> 5000:0x12a0:0x12f0:/usr/bin/myapp > + > +=C2=A0 # Read violations from the trace buffer > +=C2=A0 cat /sys/kernel/tracing/trace > + > +Up to ``TLOB_MAX_MONITORED`` tasks may be monitored simultaneously. > + > +The offsets can be obtained with ``nm`` or ``readelf``:: > + > +=C2=A0 nm -n /usr/bin/myapp | grep my_function > +=C2=A0 # -> 0000000000012a0 T my_function > + > +=C2=A0 readelf -s /usr/bin/myapp | grep my_function > +=C2=A0 # -> 42: 0000000000012a0=C2=A0 336 FUNC GLOBAL DEFAULT=C2=A0 13 m= y_function > + > +=C2=A0 # offset_start =3D 0x12a0 (function entry) > +=C2=A0 # offset_stop=C2=A0 =3D 0x12a0 + 0x50 =3D 0x12f0 (or any instruct= ion before return) > + > +Notes: > + > +- The uprobes fire for every task that executes the probed instruction, > +=C2=A0 so concurrent calls from different threads each get independent > +=C2=A0 monitoring slots. > +- ``offset_stop`` need not be a function return; it can be any instructi= on > +=C2=A0 within the region.=C2=A0 If the stop probe is never reached (e.g.= early exit > +=C2=A0 path bypasses it), the hrtimer fires and a budget violation is re= ported. > +- Each ``(binary_path, offset_start)`` pair may only be registered once. > +=C2=A0 A second write with the same ``offset_start`` for the same binary= is > +=C2=A0 rejected with ``-EEXIST``.=C2=A0 Two entry uprobes at the same ad= dress would > +=C2=A0 both fire for every task, causing ``tlob_start_task()`` to be cal= led > +=C2=A0 twice; the second call would silently fail with ``-EEXIST`` and t= he > +=C2=A0 second binding's threshold would never take effect.=C2=A0 Differe= nt code > +=C2=A0 regions that share the same ``offset_stop`` (common exit point) a= re > +=C2=A0 explicitly allowed. > +- The uprobe binding is removed when ``-offset_start:binary_path`` is > +=C2=A0 written to ``monitor``, or when the monitor is disabled. > +- The ``tag`` field in every ``tlob_budget_exceeded`` event is > +=C2=A0 automatically set to ``offset_start`` for the tracefs path, so > +=C2=A0 violation events for different code regions are immediately > +=C2=A0 distinguishable even when ``threshold_us`` values are identical. > + > +ftrace ring buffer (budget violation events) > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > + > +When a monitored task exceeds its latency budget the hrtimer fires, > +records the violation, and emits a single ``tlob_budget_exceeded`` event > +into the ftrace ring buffer.=C2=A0 **Nothing is written to the ftrace ri= ng > +buffer while the task is within budget.** > + > +The event carries the on-CPU / off-CPU time breakdown so that root-cause > +analysis (CPU-bound vs. scheduling / I/O overrun) is immediate:: > + > +=C2=A0 cat /sys/kernel/tracing/trace > + > +Example output:: > + > +=C2=A0 myapp-1234 [003] .... 12345.678: tlob_budget_exceeded: \ > +=C2=A0=C2=A0=C2=A0 myapp[1234]: budget exceeded threshold=3D5000 \ > +=C2=A0=C2=A0=C2=A0 on_cpu=3D820 off_cpu=3D4500 switches=3D3 state=3Doff_= cpu tag=3D0x00000000000012a0 > + > +Field descriptions: > + > +``threshold`` > +=C2=A0 Configured latency budget in microseconds. > + > +``on_cpu`` > +=C2=A0 Cumulative on-CPU time since ``trace_start``, in microseconds. > + > +``off_cpu`` > +=C2=A0 Cumulative off-CPU (scheduling + I/O wait) time since ``trace_sta= rt``, > +=C2=A0 in microseconds. > + > +``switches`` > +=C2=A0 Number of times the task was scheduled out during this window. > + > +``state`` > +=C2=A0 DA state when the hrtimer fired: ``on_cpu`` means the task was ex= ecuting > +=C2=A0 when the budget expired (CPU-bound overrun); ``off_cpu`` means th= e task > +=C2=A0 was preempted or blocked (scheduling / I/O overrun). > + > +``tag`` > +=C2=A0 Opaque 64-bit cookie supplied by the caller via ``tlob_start_args= .tag`` > +=C2=A0 (ioctl path) or automatically set to ``offset_start`` (tracefs up= robe > +=C2=A0 path).=C2=A0 Use it to distinguish violations from different code= regions > +=C2=A0 monitored by the same thread.=C2=A0 Zero when not set. > + > +To capture violations in a file:: > + > +=C2=A0 trace-cmd record -e tlob_budget_exceeded & > +=C2=A0 # ... run workload ... > +=C2=A0 trace-cmd report > + > +/dev/rv ioctl interface (self-instrumentation) > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > + > +Tasks can self-instrument their own code paths via the ``/dev/rv`` misc > +device (requires ``CONFIG_RV_CHARDEV``).=C2=A0 The kernel key is > +``task_struct``; multiple threads sharing a single fd each get their own > +independent monitoring slot. > + > +**Synchronous mode**=C2=A0 --=C2=A0 the calling thread checks its own re= sult:: > + > +=C2=A0 int fd =3D open("/dev/rv", O_RDWR); > + > +=C2=A0 struct tlob_start_args args =3D { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .threshold_us =3D 50000,=C2=A0=C2=A0 /* 5= 0 ms */ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .tag=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 =3D 0,=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 /* optional; = 0 =3D don't care */ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .notify_fd=C2=A0=C2=A0=C2=A0 =3D -1,=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 /* no fd notification */ > +=C2=A0 }; > +=C2=A0 ioctl(fd, TLOB_IOCTL_TRACE_START, &args); > + > +=C2=A0 /* ... code path under observation ... */ > + > +=C2=A0 int ret =3D ioctl(fd, TLOB_IOCTL_TRACE_STOP, NULL); > +=C2=A0 /* ret =3D=3D 0:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0 within budget=C2=A0 */ > +=C2=A0 /* ret =3D=3D -EOVERFLOW: budget exceeded */ > + > +=C2=A0 close(fd); > + > +**Asynchronous mode**=C2=A0 --=C2=A0 a dedicated monitor thread receives= violation > +records via ``read()`` on a shared fd, decoupling the observation from > +the critical path:: > + > +=C2=A0 /* Monitor thread: open a dedicated fd. */ > +=C2=A0 int monitor_fd =3D open("/dev/rv", O_RDWR); > + > +=C2=A0 /* Worker thread: set notify_fd =3D monitor_fd in TRACE_START arg= s. */ > +=C2=A0 int work_fd =3D open("/dev/rv", O_RDWR); > +=C2=A0 struct tlob_start_args args =3D { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .threshold_us =3D 10000,=C2=A0=C2=A0 /* 1= 0 ms */ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .tag=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 =3D REGION_A, > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .notify_fd=C2=A0=C2=A0=C2=A0 =3D monitor_= fd, > +=C2=A0 }; > +=C2=A0 ioctl(work_fd, TLOB_IOCTL_TRACE_START, &args); > +=C2=A0 /* ... critical section ... */ > +=C2=A0 ioctl(work_fd, TLOB_IOCTL_TRACE_STOP, NULL); > + > +=C2=A0 /* Monitor thread: blocking read() returns one or more tlob_event= records. > */ > +=C2=A0 struct tlob_event ntfs[8]; > +=C2=A0 ssize_t n =3D read(monitor_fd, ntfs, sizeof(ntfs)); > +=C2=A0 for (int i =3D 0; i < n / sizeof(struct tlob_event); i++) { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 struct tlob_event *ntf =3D &ntfs[i]; > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 printf("tid=3D%u tag=3D0x%llx exceeded bu= dget=3D%llu us " > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= "(on_cpu=3D%llu off_cpu=3D%llu switches=3D%u state=3D%s)\n", > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= ntf->tid, ntf->tag, ntf->threshold_us, > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= ntf->on_cpu_us, ntf->off_cpu_us, ntf->switches, > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= ntf->state ? "on_cpu" : "off_cpu"); > +=C2=A0 } > + > +**mmap ring buffer**=C2=A0 --=C2=A0 zero-copy consumption of violation e= vents:: > + > +=C2=A0 int fd =3D open("/dev/rv", O_RDWR); > +=C2=A0 struct tlob_start_args args =3D { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .threshold_us =3D 1000,=C2=A0=C2=A0 /* 1 = ms */ > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 .notify_fd=C2=A0=C2=A0=C2=A0 =3D fd,=C2= =A0=C2=A0=C2=A0=C2=A0 /* push violations to own ring buffer */ > +=C2=A0 }; > +=C2=A0 ioctl(fd, TLOB_IOCTL_TRACE_START, &args); > + > +=C2=A0 /* Map the ring: one control page + capacity data records. */ > +=C2=A0 size_t pagesize =3D sysconf(_SC_PAGESIZE); > +=C2=A0 size_t cap =3D 64;=C2=A0=C2=A0 /* read from page->capacity after = mmap */ > +=C2=A0 size_t len =3D pagesize + cap * sizeof(struct tlob_event); > +=C2=A0 void *map =3D mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_SHARED,= fd, 0); > + > +=C2=A0 struct tlob_mmap_page *page =3D map; > +=C2=A0 struct tlob_event *data =3D > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 (struct tlob_event *)((char *)map + page-= >data_offset); > + > +=C2=A0 /* Consumer loop: poll for events, read without copying. */ > +=C2=A0 while (1) { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 poll(&(struct pollfd){fd, POLLIN, 0}, 1, = -1); > + > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 uint32_t head =3D __atomic_load_n(&page->= data_head, __ATOMIC_ACQUIRE); > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 uint32_t tail =3D page->data_tail; > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 while (tail !=3D head) { > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 handle(&data[tail= & (page->capacity - 1)]); > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 tail++; > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 } > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 __atomic_store_n(&page->data_tail, tail, = __ATOMIC_RELEASE); > +=C2=A0 } > + > +Note: ``read()`` and ``mmap()`` share the same ring and ``data_tail`` > +cursor.=C2=A0 Do not use both simultaneously on the same fd. > + > +``tlob_event`` fields: > + > +``tid`` > +=C2=A0 Thread ID (``task_pid_vnr``) of the violating task. > + > +``threshold_us`` > +=C2=A0 Budget that was exceeded, in microseconds. > + > +``on_cpu_us`` > +=C2=A0 Cumulative on-CPU time at violation time, in microseconds. > + > +``off_cpu_us`` > +=C2=A0 Cumulative off-CPU time at violation time, in microseconds. > + > +``switches`` > +=C2=A0 Number of context switches since ``TRACE_START``. > + > +``state`` > +=C2=A0 1 =3D timer fired while task was on-CPU; 0 =3D timer fired while = off-CPU. > + > +``tag`` > +=C2=A0 Cookie from ``tlob_start_args.tag``; for the tracefs uprobe path = this > +=C2=A0 equals ``offset_start``.=C2=A0 Zero when not set. > + > +tracefs files > +------------- > + > +The following files are created under > +``/sys/kernel/tracing/rv/monitors/tlob/``: > + > +``enable`` (rw) > +=C2=A0 Write ``1`` to enable the monitor; write ``0`` to disable it and > +=C2=A0 stop all currently monitored tasks. > + > +``desc`` (ro) > +=C2=A0 Human-readable description of the monitor. > + > +``monitor`` (rw) > +=C2=A0 Write ``threshold_us:offset_start:offset_stop:binary_path`` to bi= nd two > +=C2=A0 plain entry uprobes in *binary_path*.=C2=A0 The uprobe at *offset= _start* fires > +=C2=A0 ``tlob_start_task()``; the uprobe at *offset_stop* fires > +=C2=A0 ``tlob_stop_task()``.=C2=A0 Returns ``-EEXIST`` if a binding with= the same > +=C2=A0 *offset_start* already exists for *binary_path*.=C2=A0 Write > +=C2=A0 ``-offset_start:binary_path`` to remove the binding.=C2=A0 Read t= o list > +=C2=A0 registered bindings, one > +=C2=A0 ``threshold_us:0xoffset_start:0xoffset_stop:binary_path`` entry p= er line. > + > +Specification > +------------- > + > +Graphviz DOT file in tools/verification/models/tlob.dot > diff --git a/Documentation/userspace-api/ioctl/ioctl-number.rst > b/Documentation/userspace-api/ioctl/ioctl-number.rst > index 331223761..8d3af68db 100644 > --- a/Documentation/userspace-api/ioctl/ioctl-number.rst > +++ b/Documentation/userspace-api/ioctl/ioctl-number.rst > @@ -385,6 +385,7 @@ Code=C2=A0 Seq#=C2=A0=C2=A0=C2=A0 Include > File=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Comments > =C2=A00xB8=C2=A0 01-02=C2=A0 uapi/misc/mrvl_cn10k_dpi.h=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 > Marvell CN10K DPI driver > =C2=A00xB8=C2=A0 all=C2=A0=C2=A0=C2=A0 uapi/linux/mshv.h=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > Microsoft Hyper-V /dev/mshv driver > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > > +0xB9=C2=A0 00-3F=C2=A0 linux/rv.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 > Runtime Verification (RV) monitors > =C2=A00xBA=C2=A0 00-0F=C2=A0 uapi/linux/liveupdate.h=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Pasha > Tatashin > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > > =C2=A00xC0=C2=A0 00-0F=C2=A0 linux/usb/iowarrior.h > diff --git a/include/uapi/linux/rv.h b/include/uapi/linux/rv.h > new file mode 100644 > index 000000000..d1b96d8cd > --- /dev/null > +++ b/include/uapi/linux/rv.h > @@ -0,0 +1,181 @@ > +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ > +/* > + * UAPI definitions for Runtime Verification (RV) monitors. > + * > + * All RV monitors that expose an ioctl self-instrumentation interface > + * share the magic byte RV_IOC_MAGIC (0xB9), registered in > + * Documentation/userspace-api/ioctl/ioctl-number.rst. > + * > + * A single /dev/rv misc device serves as the entry point.=C2=A0 ioctl n= umbers > + * encode both the monitor identity and the operation: > + * > + *=C2=A0=C2=A0 0x01 - 0x1F=C2=A0 tlob (task latency over budget) > + *=C2=A0=C2=A0 0x20 - 0x3F=C2=A0 reserved for future RV monitors > + * > + * Usage examples and design rationale are in: > + *=C2=A0=C2=A0 Documentation/trace/rv/monitor_tlob.rst > + */ > + > +#ifndef _UAPI_LINUX_RV_H > +#define _UAPI_LINUX_RV_H > + > +#include > +#include > + > +/* Magic byte shared by all RV monitor ioctls. */ > +#define RV_IOC_MAGIC 0xB9 > + > +/* ---------------------------------------------------------------------= -- > + * tlob: task latency over budget monitor=C2=A0 (nr 0x01 - 0x1F) > + * ---------------------------------------------------------------------= -- > + */ > + > +/** > + * struct tlob_start_args - arguments for TLOB_IOCTL_TRACE_START > + * @threshold_us: Latency budget for this critical section, in microseco= nds. > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 Must be greater than zero. > + * @tag:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Opaque 64-bit c= ookie supplied by the caller.=C2=A0 Echoed back > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 verbatim in the tlob_budget_exceeded ftrace event and in an= y > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 tlob_event record delivered via @notify_fd.=C2=A0 Use it to > identify > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 which code region triggered a violation when the same threa= d > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 monitors multiple regions sequentially.=C2=A0 Set to 0 if n= ot > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 needed. > + * @notify_fd:=C2=A0=C2=A0 File descriptor that will receive a tlob_even= t record on > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 violation.=C2=A0 Must refer to an open /dev/rv fd.=C2=A0 Ma= y equal > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 the calling fd (self-notification, useful for retrieving th= e > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 on_cpu_us / off_cpu_us breakdown after TRACE_STOP returns > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 -EOVERFLOW).=C2=A0 Set to -1 to disable fd notification; in= that > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 case violations are only signalled via the TRACE_STOP retur= n > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 value and the tlob_budget_exceeded ftrace event. > + * @flags:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Must be 0.=C2=A0 Reserved= for future extensions. > + */ > +struct tlob_start_args { > + __u64 threshold_us; > + __u64 tag; > + __s32 notify_fd; > + __u32 flags; > +}; > + > +/** > + * struct tlob_event - one budget-exceeded event > + * > + * Consumed by read() on the notify_fd registered at TLOB_IOCTL_TRACE_ST= ART. > + * Each record describes a single budget exceedance for one task. > + * > + * @tid:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Thread ID= (task_pid_vnr) of the violating task. > + * @threshold_us: Budget that was exceeded, in microseconds. > + * @on_cpu_us:=C2=A0=C2=A0=C2=A0 Cumulative on-CPU time at violation tim= e, in microseconds. > + * @off_cpu_us:=C2=A0=C2=A0 Cumulative off-CPU (scheduling + I/O wait) t= ime at > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 violation time, in microseconds. > + * @switches:=C2=A0=C2=A0=C2=A0=C2=A0 Number of context switches since T= RACE_START. > + * @state:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 DA state at violati= on: 1 =3D on_cpu, 0 =3D off_cpu. > + * @tag:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Cookie fr= om tlob_start_args.tag; for the tracefs uprobe > path > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 this is the offset_start value.=C2=A0 Zero when not set. > + */ > +struct tlob_event { > + __u32 tid; > + __u32 pad; > + __u64 threshold_us; > + __u64 on_cpu_us; > + __u64 off_cpu_us; > + __u32 switches; > + __u32 state;=C2=A0=C2=A0 /* 1 =3D on_cpu, 0 =3D off_cpu */ > + __u64 tag; > +}; > + > +/** > + * struct tlob_mmap_page - control page for the mmap'd violation ring bu= ffer > + * > + * Mapped at offset 0 of the mmap region returned by mmap(2) on a /dev/r= v fd. > + * The data array of struct tlob_event records begins at offset @data_of= fset > + * (always one page from the mmap base; use this field rather than hard- > coding > + * PAGE_SIZE so the code remains correct across architectures). > + * > + * Ring layout: > + * > + *=C2=A0=C2=A0 mmap base + 0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 : struct tlob_mmap_page=C2=A0 (one page) > + *=C2=A0=C2=A0 mmap base + data_offset=C2=A0=C2=A0 : struct tlob_event[c= apacity] > + * > + * The mmap length determines the ring capacity.=C2=A0 Compute it as: > + * > + *=C2=A0=C2=A0 raw=C2=A0=C2=A0=C2=A0 =3D sysconf(_SC_PAGESIZE) + capacit= y * sizeof(struct tlob_event) > + *=C2=A0=C2=A0 length =3D (raw + sysconf(_SC_PAGESIZE) - 1) & ~(sysconf(= _SC_PAGESIZE) - > 1) > + * > + * i.e. round the raw byte count up to the next page boundary before > + * passing it to mmap(2).=C2=A0 The kernel requires a page-aligned lengt= h. > + * capacity must be a power of 2.=C2=A0 Read @capacity after a successfu= l > + * mmap(2) for the actual value. > + * > + * Producer/consumer ordering contract: > + * > + *=C2=A0=C2=A0 Kernel (producer): > + *=C2=A0=C2=A0=C2=A0=C2=A0 data[data_head & (capacity - 1)] =3D event; > + *=C2=A0=C2=A0=C2=A0=C2=A0 // pairs with load-acquire in userspace: > + *=C2=A0=C2=A0=C2=A0=C2=A0 smp_store_release(&page->data_head, data_head= + 1); > + * > + *=C2=A0=C2=A0 Userspace (consumer): > + *=C2=A0=C2=A0=C2=A0=C2=A0 // pairs with store-release in kernel: > + *=C2=A0=C2=A0=C2=A0=C2=A0 head =3D __atomic_load_n(&page->data_head, __= ATOMIC_ACQUIRE); > + *=C2=A0=C2=A0=C2=A0=C2=A0 for (tail =3D page->data_tail; tail !=3D head= ; tail++) > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 handle(&data[tail & (= capacity - 1)]); > + *=C2=A0=C2=A0=C2=A0=C2=A0 __atomic_store_n(&page->data_tail, tail, __AT= OMIC_RELEASE); > + * > + * @data_head and @data_tail are monotonically increasing __u32 counters > + * in units of records.=C2=A0 Unsigned 32-bit wrap-around is handled cor= rectly > + * by modular arithmetic; the ring is full when > + * (data_head - data_tail) =3D=3D capacity. > + * > + * When the ring is full the kernel drops the incoming record and increm= ents > + * @dropped.=C2=A0 The consumer should check @dropped periodically to de= tect loss. > + * > + * read() and mmap() share the same ring buffer.=C2=A0 Do not use both > + * simultaneously on the same fd. > + * > + * @data_head:=C2=A0=C2=A0 Next write slot index.=C2=A0 Updated by the k= ernel with > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 store-release ordering.=C2=A0 Read by userspace with load- > acquire. > + * @data_tail:=C2=A0=C2=A0 Next read slot index.=C2=A0 Updated by usersp= ace.=C2=A0 Read by the > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 kernel to detect overflow. > + * @capacity:=C2=A0=C2=A0=C2=A0 Actual ring capacity in records (power o= f 2).=C2=A0 Written once > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 by the kernel at mmap time; read-only for userspace > thereafter. > + * @version:=C2=A0=C2=A0=C2=A0=C2=A0 Ring buffer ABI version; currently = 1. > + * @data_offset: Byte offset from the mmap base to the data array. > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 Always equal to sysconf(_SC_PAGESIZE) on the running kernel= . > + * @record_size: sizeof(struct tlob_event) as seen by the kernel.=C2=A0 = Verify > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 this matches userspace's sizeof before indexing the array. > + * @dropped:=C2=A0=C2=A0=C2=A0=C2=A0 Number of events dropped because th= e ring was full. > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 Monotonically increasing; read with __ATOMIC_RELAXED. > + */ > +struct tlob_mmap_page { > + __u32=C2=A0 data_head; > + __u32=C2=A0 data_tail; > + __u32=C2=A0 capacity; > + __u32=C2=A0 version; > + __u32=C2=A0 data_offset; > + __u32=C2=A0 record_size; > + __u64=C2=A0 dropped; > +}; > + > +/* > + * TLOB_IOCTL_TRACE_START - begin monitoring the calling task. > + * > + * Arms a per-task hrtimer for threshold_us microseconds.=C2=A0 If args.= notify_fd > + * is >=3D 0, a tlob_event record is pushed into that fd's ring buffer o= n > + * violation in addition to the tlob_budget_exceeded ftrace event. > + * args.notify_fd =3D=3D -1 disables fd notification. > + * > + * Violation records are consumed by read() on the notify_fd (blocking o= r > + * non-blocking depending on O_NONBLOCK).=C2=A0 On violation, > TLOB_IOCTL_TRACE_STOP > + * also returns -EOVERFLOW regardless of whether notify_fd is set. > + * > + * args.flags must be 0. > + */ > +#define TLOB_IOCTL_TRACE_START _IOW(RV_IOC_MAGIC, 0x01, struct > tlob_start_args) > + > +/* > + * TLOB_IOCTL_TRACE_STOP - end monitoring the calling task. > + * > + * Returns 0 if within budget, -EOVERFLOW if the budget was exceeded. > + */ > +#define TLOB_IOCTL_TRACE_STOP _IO(RV_IOC_MAGIC,=C2=A0 0x02) > + > +#endif /* _UAPI_LINUX_RV_H */ > diff --git a/kernel/trace/rv/Kconfig b/kernel/trace/rv/Kconfig > index 5b4be87ba..227573cda 100644 > --- a/kernel/trace/rv/Kconfig > +++ b/kernel/trace/rv/Kconfig > @@ -65,6 +65,7 @@ source "kernel/trace/rv/monitors/pagefault/Kconfig" > =C2=A0source "kernel/trace/rv/monitors/sleep/Kconfig" > =C2=A0# Add new rtapp monitors here > =C2=A0 > +source "kernel/trace/rv/monitors/tlob/Kconfig" > =C2=A0# Add new monitors here > =C2=A0 > =C2=A0config RV_REACTORS > @@ -93,3 +94,19 @@ config RV_REACT_PANIC > =C2=A0 help > =C2=A0 =C2=A0 Enables the panic reactor. The panic reactor emits a printk= () > =C2=A0 =C2=A0 message if an exception is found and panic()s the system. > + > +config RV_CHARDEV > + bool "RV ioctl interface via /dev/rv" > + depends on RV > + default n > + help > + =C2=A0 Register a /dev/rv misc device that exposes an ioctl interface > + =C2=A0 for RV monitor self-instrumentation.=C2=A0 All RV monitors share= the > + =C2=A0 single device node; ioctl numbers encode the monitor identity. > + > + =C2=A0 When enabled, user-space programs can open /dev/rv and use > + =C2=A0 monitor-specific ioctl commands to bracket code regions they > + =C2=A0 want the kernel RV subsystem to observe. > + > + =C2=A0 Say Y here if you want to use the tlob self-instrumentation > + =C2=A0 ioctl interface; otherwise say N. > diff --git a/kernel/trace/rv/Makefile b/kernel/trace/rv/Makefile > index 750e4ad6f..cc3781a3b 100644 > --- a/kernel/trace/rv/Makefile > +++ b/kernel/trace/rv/Makefile > @@ -3,6 +3,7 @@ > =C2=A0ccflags-y +=3D -I $(src) # needed for trace events > =C2=A0 > =C2=A0obj-$(CONFIG_RV) +=3D rv.o > +obj-$(CONFIG_RV_CHARDEV) +=3D rv_dev.o > =C2=A0obj-$(CONFIG_RV_MON_WIP) +=3D monitors/wip/wip.o > =C2=A0obj-$(CONFIG_RV_MON_WWNR) +=3D monitors/wwnr/wwnr.o > =C2=A0obj-$(CONFIG_RV_MON_SCHED) +=3D monitors/sched/sched.o > @@ -17,6 +18,7 @@ obj-$(CONFIG_RV_MON_STS) +=3D monitors/sts/sts.o > =C2=A0obj-$(CONFIG_RV_MON_NRP) +=3D monitors/nrp/nrp.o > =C2=A0obj-$(CONFIG_RV_MON_SSSW) +=3D monitors/sssw/sssw.o > =C2=A0obj-$(CONFIG_RV_MON_OPID) +=3D monitors/opid/opid.o > +obj-$(CONFIG_RV_MON_TLOB) +=3D monitors/tlob/tlob.o > =C2=A0# Add new monitors here > =C2=A0obj-$(CONFIG_RV_REACTORS) +=3D rv_reactors.o > =C2=A0obj-$(CONFIG_RV_REACT_PRINTK) +=3D reactor_printk.o > diff --git a/kernel/trace/rv/monitors/tlob/Kconfig > b/kernel/trace/rv/monitors/tlob/Kconfig > new file mode 100644 > index 000000000..010237480 > --- /dev/null > +++ b/kernel/trace/rv/monitors/tlob/Kconfig > @@ -0,0 +1,51 @@ > +# SPDX-License-Identifier: GPL-2.0-only > +# > +config RV_MON_TLOB > + depends on RV > + depends on UPROBES > + select DA_MON_EVENTS_ID > + bool "tlob monitor" > + help > + =C2=A0 Enable the tlob (task latency over budget) monitor. This monitor > + =C2=A0 tracks the elapsed time (CLOCK_MONOTONIC) of a marked code path > within a > + =C2=A0 task (including both on-CPU and off-CPU time) and reports a > + =C2=A0 violation when the elapsed time exceeds a configurable budget > + =C2=A0 threshold. > + > + =C2=A0 The monitor implements a three-state deterministic automaton. > + =C2=A0 States: unmonitored, on_cpu, off_cpu. > + =C2=A0 Key transitions: > + =C2=A0=C2=A0=C2=A0 unmonitored=C2=A0=C2=A0=C2=A0 --(trace_start)-->=C2= =A0=C2=A0=C2=A0 on_cpu > + =C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0 --(switch_out)-->=C2=A0=C2=A0=C2= =A0=C2=A0 off_cpu > + =C2=A0=C2=A0=C2=A0 off_cpu=C2=A0 --(switch_in)-->=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 on_cpu > + =C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0 --(trace_stop)-->=C2=A0=C2=A0=C2= =A0 unmonitored > + =C2=A0=C2=A0=C2=A0 off_cpu=C2=A0 --(trace_stop)-->=C2=A0=C2=A0=C2=A0 un= monitored > + =C2=A0=C2=A0=C2=A0 on_cpu=C2=A0=C2=A0 --(budget_expired)--> unmonitored > + =C2=A0=C2=A0=C2=A0 off_cpu=C2=A0 --(budget_expired)--> unmonitored > + > + =C2=A0 External configuration is done via the tracefs "monitor" file: > + =C2=A0=C2=A0=C2=A0 echo pid:threshold_us:binary:offset_start:offset_sto= p > > .../rv/monitors/tlob/monitor > + =C2=A0=C2=A0=C2=A0 echo -pid=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > .../rv/monitors/tlob/monitor=C2=A0 (remove > task) > + =C2=A0=C2=A0=C2=A0 cat=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ..= ./rv/monitors/tlob/monitor=C2=A0 (list > tasks) > + > + =C2=A0 The uprobe binding places two plain entry uprobes at offset_star= t > and > + =C2=A0 offset_stop in the binary; these trigger tlob_start_task() and > + =C2=A0 tlob_stop_task() respectively.=C2=A0 Using two entry uprobes (ra= ther > than a > + =C2=A0 uretprobe) means that a mistyped offset can never corrupt the ca= ll > + =C2=A0 stack; the worst outcome is a missed stop, which causes the hrti= mer > to > + =C2=A0 fire and report a budget violation. > + > + =C2=A0 Violation events are delivered via a lock-free mmap ring buffer = on > + =C2=A0 /dev/rv (enabled by CONFIG_RV_CHARDEV).=C2=A0 The consumer mmap(= )s the > + =C2=A0 device, reads records from the data array using the head/tail > indices > + =C2=A0 in the control page, and advances data_tail when done. > + > + =C2=A0 For self-instrumentation, use TLOB_IOCTL_TRACE_START / > + =C2=A0 TLOB_IOCTL_TRACE_STOP via the /dev/rv misc device (enabled by > + =C2=A0 CONFIG_RV_CHARDEV). > + > + =C2=A0 Up to TLOB_MAX_MONITORED tasks may be monitored simultaneously. > + > + =C2=A0 For further information, see: > + =C2=A0=C2=A0=C2=A0 Documentation/trace/rv/monitor_tlob.rst > + > diff --git a/kernel/trace/rv/monitors/tlob/tlob.c > b/kernel/trace/rv/monitors/tlob/tlob.c > new file mode 100644 > index 000000000..a6e474025 > --- /dev/null > +++ b/kernel/trace/rv/monitors/tlob/tlob.c > @@ -0,0 +1,986 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* > + * tlob: task latency over budget monitor > + * > + * Track the elapsed wall-clock time of a marked code path and detect wh= en > + * a monitored task exceeds its per-task latency budget.=C2=A0 CLOCK_MON= OTONIC > + * is used so both on-CPU and off-CPU time count toward the budget. > + * > + * Per-task state is maintained in a spinlock-protected hash table.=C2= =A0 A > + * one-shot hrtimer fires at the deadline; if the task has not called > + * trace_stop by then, a violation is recorded. > + * > + * Up to TLOB_MAX_MONITORED tasks may be tracked simultaneously. > + * > + * Copyright (C) 2026 Wen Yang > + */ > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > + > +/* rv_interface_lock is defined in kernel/trace/rv/rv.c */ > +extern struct mutex rv_interface_lock; > + > +#define MODULE_NAME "tlob" > + > +#include > +#include > + > +#define RV_MON_TYPE RV_MON_PER_TASK > +#include "tlob.h" > +#include > + > +/* Hash table size; must be a power of two. */ > +#define TLOB_HTABLE_BITS 6 > +#define TLOB_HTABLE_SIZE (1 << TLOB_HTABLE_BITS) > + > +/* Maximum binary path length for uprobe binding. */ > +#define TLOB_MAX_PATH 256 > + > +/* Per-task latency monitoring state. */ > +struct tlob_task_state { > + struct hlist_node hlist; > + struct task_struct *task; > + u64 threshold_us; > + u64 tag; > + struct hrtimer deadline_timer; > + int canceled; /* protected by entry_lock */ > + struct file *notify_file; /* NULL or held reference */ > + > + /* > + * entry_lock serialises the mutable accounting fields below. > + * Lock order: tlob_table_lock -> entry_lock (never reverse). > + */ > + raw_spinlock_t entry_lock; > + u64 on_cpu_us; > + u64 off_cpu_us; > + ktime_t last_ts; > + u32 switches; > + u8 da_state; > + > + struct rcu_head rcu; /* for call_rcu() teardown */ > +}; > + > +/* Per-uprobe-binding state: a start + stop probe pair for one binary re= gion. > */ > +struct tlob_uprobe_binding { > + struct list_head list; > + u64 threshold_us; > + struct path path; > + char binpath[TLOB_MAX_PATH]; /* canonical > path for read/remove */ > + loff_t offset_start; > + loff_t offset_stop; > + struct uprobe_consumer entry_uc; > + struct uprobe_consumer stop_uc; > + struct uprobe *entry_uprobe; > + struct uprobe *stop_uprobe; > +}; > + > +/* Object pool for tlob_task_state. */ > +static struct kmem_cache *tlob_state_cache; > + > +/* Hash table and lock protecting table structure (insert/delete/cancele= d). > */ > +static struct hlist_head tlob_htable[TLOB_HTABLE_SIZE]; > +static DEFINE_RAW_SPINLOCK(tlob_table_lock); > +static atomic_t tlob_num_monitored =3D ATOMIC_INIT(0); > + > +/* Uprobe binding list; protected by tlob_uprobe_mutex. */ > +static LIST_HEAD(tlob_uprobe_list); > +static DEFINE_MUTEX(tlob_uprobe_mutex); > + > +/* Forward declaration */ > +static enum hrtimer_restart tlob_deadline_timer_fn(struct hrtimer *timer= ); > + > +/* Hash table helpers */ > + > +static unsigned int tlob_hash_task(const struct task_struct *task) > +{ > + return hash_ptr((void *)task, TLOB_HTABLE_BITS); > +} > + > +/* > + * tlob_find_rcu - look up per-task state. > + * Must be called under rcu_read_lock() or with tlob_table_lock held. > + */ > +static struct tlob_task_state *tlob_find_rcu(struct task_struct *task) > +{ > + struct tlob_task_state *ws; > + unsigned int h =3D tlob_hash_task(task); > + > + hlist_for_each_entry_rcu(ws, &tlob_htable[h], hlist, > + lockdep_is_held(&tlob_table_lock)) > + if (ws->task =3D=3D task) > + return ws; > + return NULL; > +} > + > +/* Allocate and initialise a new per-task state entry. */ > +static struct tlob_task_state *tlob_alloc(struct task_struct *task, > + =C2=A0 u64 threshold_us, u64 tag) > +{ > + struct tlob_task_state *ws; > + > + ws =3D kmem_cache_zalloc(tlob_state_cache, GFP_ATOMIC); > + if (!ws) > + return NULL; > + > + ws->task =3D task; > + get_task_struct(task); > + ws->threshold_us =3D threshold_us; > + ws->tag =3D tag; > + ws->last_ts =3D ktime_get(); > + ws->da_state =3D on_cpu_tlob; > + raw_spin_lock_init(&ws->entry_lock); > + hrtimer_setup(&ws->deadline_timer, tlob_deadline_timer_fn, > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 CLOCK_MONOTONIC, HRTIMER_MODE_REL); > + return ws; > +} > + > +/* RCU callback: free the slab once no readers remain. */ > +static void tlob_free_rcu_slab(struct rcu_head *head) > +{ > + struct tlob_task_state *ws =3D > + container_of(head, struct tlob_task_state, rcu); > + kmem_cache_free(tlob_state_cache, ws); > +} > + > +/* Arm the one-shot deadline timer for threshold_us microseconds. */ > +static void tlob_arm_deadline(struct tlob_task_state *ws) > +{ > + hrtimer_start(&ws->deadline_timer, > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ns_to_ktime(ws->threshold_us * NSEC_PER= _USEC), > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 HRTIMER_MODE_REL); > +} > + > +/* > + * Push a violation record into a monitor fd's ring buffer (softirq cont= ext). > + * Drop-new policy: discard incoming record when full.=C2=A0 smp_store_r= elease on > + * data_head pairs with smp_load_acquire in the consumer. > + */ > +static void tlob_event_push(struct rv_file_priv *priv, > + =C2=A0=C2=A0=C2=A0 const struct tlob_event *info) > +{ > + struct tlob_ring *ring =3D &priv->ring; > + unsigned long flags; > + u32 head, tail; > + > + spin_lock_irqsave(&ring->lock, flags); > + > + head =3D ring->page->data_head; > + tail =3D READ_ONCE(ring->page->data_tail); > + > + if (head - tail > ring->mask) { > + /* Ring full: drop incoming record. */ > + ring->page->dropped++; > + spin_unlock_irqrestore(&ring->lock, flags); > + return; > + } > + > + ring->data[head & ring->mask] =3D *info; > + /* pairs with smp_load_acquire() in the consumer */ > + smp_store_release(&ring->page->data_head, head + 1); > + > + spin_unlock_irqrestore(&ring->lock, flags); > + > + wake_up_interruptible_poll(&priv->waitq, EPOLLIN | EPOLLRDNORM); > +} > + > +#if IS_ENABLED(CONFIG_KUNIT) > +void tlob_event_push_kunit(struct rv_file_priv *priv, > + =C2=A0 const struct tlob_event *info) > +{ > + tlob_event_push(priv, info); > +} > +EXPORT_SYMBOL_IF_KUNIT(tlob_event_push_kunit); > +#endif /* CONFIG_KUNIT */ > + > +/* > + * Budget exceeded: remove the entry, record the violation, and inject > + * budget_expired into the DA. > + * > + * Lock order: tlob_table_lock -> entry_lock.=C2=A0 tlob_stop_task() set= s > + * ws->canceled under both locks; if we see it here the stop path owns > cleanup. > + * fput/put_task_struct are done before call_rcu(); the RCU callback onl= y > + * reclaims the slab. > + */ > +static enum hrtimer_restart tlob_deadline_timer_fn(struct hrtimer *timer= ) > +{ > + struct tlob_task_state *ws =3D > + container_of(timer, struct tlob_task_state, deadline_timer); > + struct tlob_event info =3D {}; > + struct file *notify_file; > + struct task_struct *task; > + unsigned long flags; > + /* snapshots taken under entry_lock */ > + u64 on_cpu_us, off_cpu_us, threshold_us, tag; > + u32 switches; > + bool on_cpu; > + bool push_event =3D false; > + > + raw_spin_lock_irqsave(&tlob_table_lock, flags); > + /* stop path sets canceled under both locks; if set it owns cleanup > */ > + if (ws->canceled) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + return HRTIMER_NORESTART; > + } > + > + /* Finalize accounting and snapshot all fields under entry_lock. */ > + raw_spin_lock(&ws->entry_lock); > + > + { > + ktime_t now =3D ktime_get(); > + u64 delta_us =3D ktime_to_us(ktime_sub(now, ws->last_ts)); > + > + if (ws->da_state =3D=3D on_cpu_tlob) > + ws->on_cpu_us +=3D delta_us; > + else > + ws->off_cpu_us +=3D delta_us; > + } > + > + ws->canceled=C2=A0 =3D 1; > + on_cpu_us=C2=A0=C2=A0=C2=A0=C2=A0 =3D ws->on_cpu_us; > + off_cpu_us=C2=A0=C2=A0=C2=A0 =3D ws->off_cpu_us; > + threshold_us=C2=A0 =3D ws->threshold_us; > + tag=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D ws-= >tag; > + switches=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D ws->switches; > + on_cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D (ws->da_state =3D= =3D on_cpu_tlob); > + notify_file=C2=A0=C2=A0 =3D ws->notify_file; > + if (notify_file) { > + info.tid=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D tas= k_pid_vnr(ws->task); > + info.threshold_us =3D threshold_us; > + info.on_cpu_us=C2=A0=C2=A0=C2=A0 =3D on_cpu_us; > + info.off_cpu_us=C2=A0=C2=A0 =3D off_cpu_us; > + info.switches=C2=A0=C2=A0=C2=A0=C2=A0 =3D switches; > + info.state=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D on_cpu ? 1 : = 0; > + info.tag=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D tag= ; > + push_event=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 =3D true; > + } > + > + raw_spin_unlock(&ws->entry_lock); > + > + hlist_del_rcu(&ws->hlist); > + atomic_dec(&tlob_num_monitored); > + /* > + * Hold a reference so task remains valid across da_handle_event() > + * after we drop tlob_table_lock. > + */ > + task =3D ws->task; > + get_task_struct(task); > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + > + /* > + * Both locks are now released; ws is exclusively owned (removed from > + * the hash table with canceled=3D1).=C2=A0 Emit the tracepoint and pus= h the > + * violation record. > + */ > + trace_tlob_budget_exceeded(ws->task, threshold_us, on_cpu_us, > + =C2=A0=C2=A0 off_cpu_us, switches, on_cpu, tag); > + > + if (push_event) { > + struct rv_file_priv *priv =3D notify_file->private_data; > + > + if (priv) > + tlob_event_push(priv, &info); > + } > + > + da_handle_event(task, budget_expired_tlob); > + > + if (notify_file) > + fput(notify_file); /* ref from fget() at > TRACE_START */ > + put_task_struct(ws->task); /* ref from tlob_alloc() */ > + put_task_struct(task); /* extra ref from > get_task_struct() above */ > + call_rcu(&ws->rcu, tlob_free_rcu_slab); > + return HRTIMER_NORESTART; > +} > + > +/* Tracepoint handlers */ > + > +/* > + * handle_sched_switch - advance the DA and accumulate on/off-CPU time. > + * > + * RCU read-side for lock-free lookup; entry_lock for per-task accountin= g. > + * da_handle_event() is called after rcu_read_unlock() to avoid holding = the > + * read-side critical section across the RV framework. > + */ > +static void handle_sched_switch(void *data, bool preempt, > + struct task_struct *prev, > + struct task_struct *next, > + unsigned int prev_state) > +{ > + struct tlob_task_state *ws; > + unsigned long flags; > + bool do_prev =3D false, do_next =3D false; > + ktime_t now; > + > + rcu_read_lock(); > + > + ws =3D tlob_find_rcu(prev); > + if (ws) { > + raw_spin_lock_irqsave(&ws->entry_lock, flags); > + if (!ws->canceled) { > + now =3D ktime_get(); > + ws->on_cpu_us +=3D ktime_to_us(ktime_sub(now, ws- > >last_ts)); > + ws->last_ts =3D now; > + ws->switches++; > + ws->da_state =3D off_cpu_tlob; > + do_prev =3D true; > + } > + raw_spin_unlock_irqrestore(&ws->entry_lock, flags); > + } > + > + ws =3D tlob_find_rcu(next); > + if (ws) { > + raw_spin_lock_irqsave(&ws->entry_lock, flags); > + if (!ws->canceled) { > + now =3D ktime_get(); > + ws->off_cpu_us +=3D ktime_to_us(ktime_sub(now, ws- > >last_ts)); > + ws->last_ts =3D now; > + ws->da_state =3D on_cpu_tlob; > + do_next =3D true; > + } > + raw_spin_unlock_irqrestore(&ws->entry_lock, flags); > + } > + > + rcu_read_unlock(); > + > + if (do_prev) > + da_handle_event(prev, switch_out_tlob); > + if (do_next) > + da_handle_event(next, switch_in_tlob); > +} > + > +static void handle_sched_wakeup(void *data, struct task_struct *p) > +{ > + struct tlob_task_state *ws; > + unsigned long flags; > + bool found =3D false; > + > + rcu_read_lock(); > + ws =3D tlob_find_rcu(p); > + if (ws) { > + raw_spin_lock_irqsave(&ws->entry_lock, flags); > + found =3D !ws->canceled; > + raw_spin_unlock_irqrestore(&ws->entry_lock, flags); > + } > + rcu_read_unlock(); > + > + if (found) > + da_handle_event(p, sched_wakeup_tlob); > +} > + > +/* ---------------------------------------------------------------------= -- > + * Core start/stop helpers (also called from rv_dev.c) > + * ---------------------------------------------------------------------= -- > + */ > + > +/* > + * __tlob_insert - insert @ws into the hash table and arm its deadline t= imer. > + * > + * Re-checks for duplicates and capacity under tlob_table_lock; the call= er > + * may have done a lock-free pre-check before allocating @ws.=C2=A0 On f= ailure @ws > + * is freed directly (never in table, so no call_rcu needed). > + */ > +static int __tlob_insert(struct task_struct *task, struct tlob_task_stat= e > *ws) > +{ > + unsigned int h; > + unsigned long flags; > + > + raw_spin_lock_irqsave(&tlob_table_lock, flags); > + if (tlob_find_rcu(task)) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + if (ws->notify_file) > + fput(ws->notify_file); > + put_task_struct(ws->task); > + kmem_cache_free(tlob_state_cache, ws); > + return -EEXIST; > + } > + if (atomic_read(&tlob_num_monitored) >=3D TLOB_MAX_MONITORED) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + if (ws->notify_file) > + fput(ws->notify_file); > + put_task_struct(ws->task); > + kmem_cache_free(tlob_state_cache, ws); > + return -ENOSPC; > + } > + h =3D tlob_hash_task(task); > + hlist_add_head_rcu(&ws->hlist, &tlob_htable[h]); > + atomic_inc(&tlob_num_monitored); > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + > + da_handle_start_run_event(task, trace_start_tlob); > + tlob_arm_deadline(ws); > + return 0; > +} > + > +/** > + * tlob_start_task - begin monitoring @task with latency budget > @threshold_us. > + * > + * @notify_file: /dev/rv fd whose ring buffer receives a tlob_event on > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 violation; caller transfers the fget() reference to tlob.c. > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 Pass NULL for synchronous mode (violations only via > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 TRACE_STOP return value and the tlob_budget_exceeded event)= . > + * > + * Returns 0, -ENODEV, -EEXIST, -ENOSPC, or -ENOMEM.=C2=A0 On failure th= e caller > + * retains responsibility for any @notify_file reference. > + */ > +int tlob_start_task(struct task_struct *task, u64 threshold_us, > + =C2=A0=C2=A0=C2=A0 struct file *notify_file, u64 tag) > +{ > + struct tlob_task_state *ws; > + unsigned long flags; > + > + if (!tlob_state_cache) > + return -ENODEV; > + > + if (threshold_us > (u64)KTIME_MAX / NSEC_PER_USEC) > + return -ERANGE; > + > + /* Quick pre-check before allocation. */ > + raw_spin_lock_irqsave(&tlob_table_lock, flags); > + if (tlob_find_rcu(task)) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + return -EEXIST; > + } > + if (atomic_read(&tlob_num_monitored) >=3D TLOB_MAX_MONITORED) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + return -ENOSPC; > + } > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + > + ws =3D tlob_alloc(task, threshold_us, tag); > + if (!ws) > + return -ENOMEM; > + > + ws->notify_file =3D notify_file; > + return __tlob_insert(task, ws); > +} > +EXPORT_SYMBOL_GPL(tlob_start_task); > + > +/** > + * tlob_stop_task - stop monitoring @task before the deadline fires. > + * > + * Sets canceled under entry_lock (inside tlob_table_lock) before callin= g > + * hrtimer_cancel(), racing safely with the timer callback. > + * > + * Returns 0 if within budget, -ESRCH if the entry is gone (deadline alr= eady > + * fired, or TRACE_START was never called). > + */ > +int tlob_stop_task(struct task_struct *task) > +{ > + struct tlob_task_state *ws; > + struct file *notify_file; > + unsigned long flags; > + > + raw_spin_lock_irqsave(&tlob_table_lock, flags); > + ws =3D tlob_find_rcu(task); > + if (!ws) { > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + return -ESRCH; > + } > + > + /* Prevent handle_sched_switch from updating accounting after > removal. */ > + raw_spin_lock(&ws->entry_lock); > + ws->canceled =3D 1; > + raw_spin_unlock(&ws->entry_lock); > + > + hlist_del_rcu(&ws->hlist); > + atomic_dec(&tlob_num_monitored); > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + > + hrtimer_cancel(&ws->deadline_timer); > + > + da_handle_event(task, trace_stop_tlob); > + > + notify_file =3D ws->notify_file; > + if (notify_file) > + fput(notify_file); > + put_task_struct(ws->task); > + call_rcu(&ws->rcu, tlob_free_rcu_slab); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(tlob_stop_task); > + > +/* Stop monitoring all tracked tasks; called on monitor disable. */ > +static void tlob_stop_all(void) > +{ > + struct tlob_task_state *batch[TLOB_MAX_MONITORED]; > + struct tlob_task_state *ws; > + struct hlist_node *tmp; > + unsigned long flags; > + int n =3D 0, i; > + > + raw_spin_lock_irqsave(&tlob_table_lock, flags); > + for (i =3D 0; i < TLOB_HTABLE_SIZE; i++) { > + hlist_for_each_entry_safe(ws, tmp, &tlob_htable[i], hlist) { > + raw_spin_lock(&ws->entry_lock); > + ws->canceled =3D 1; > + raw_spin_unlock(&ws->entry_lock); > + hlist_del_rcu(&ws->hlist); > + atomic_dec(&tlob_num_monitored); > + if (n < TLOB_MAX_MONITORED) > + batch[n++] =3D ws; > + } > + } > + raw_spin_unlock_irqrestore(&tlob_table_lock, flags); > + > + for (i =3D 0; i < n; i++) { > + ws =3D batch[i]; > + hrtimer_cancel(&ws->deadline_timer); > + da_handle_event(ws->task, trace_stop_tlob); > + if (ws->notify_file) > + fput(ws->notify_file); > + put_task_struct(ws->task); > + call_rcu(&ws->rcu, tlob_free_rcu_slab); > + } > +} > + > +/* uprobe binding helpers */ > + > +static int tlob_uprobe_entry_handler(struct uprobe_consumer *uc, > + =C2=A0=C2=A0=C2=A0=C2=A0 struct pt_regs *regs, __u64 *data) > +{ > + struct tlob_uprobe_binding *b =3D > + container_of(uc, struct tlob_uprobe_binding, entry_uc); > + > + tlob_start_task(current, b->threshold_us, NULL, (u64)b- > >offset_start); > + return 0; > +} > + > +static int tlob_uprobe_stop_handler(struct uprobe_consumer *uc, > + =C2=A0=C2=A0=C2=A0 struct pt_regs *regs, __u64 *data) > +{ > + tlob_stop_task(current); > + return 0; > +} > + > +/* > + * Register start + stop entry uprobes for a binding. > + * Both are plain entry uprobes (no uretprobe), so a wrong offset never > + * corrupts the call stack; the worst outcome is a missed stop (hrtimer > + * fires and reports a budget violation). > + * Called with tlob_uprobe_mutex held. > + */ > +static int tlob_add_uprobe(u64 threshold_us, const char *binpath, > + =C2=A0=C2=A0 loff_t offset_start, loff_t offset_stop) > +{ > + struct tlob_uprobe_binding *b, *tmp_b; > + char pathbuf[TLOB_MAX_PATH]; > + struct inode *inode; > + char *canon; > + int ret; > + > + b =3D kzalloc(sizeof(*b), GFP_KERNEL); > + if (!b) > + return -ENOMEM; > + > + if (binpath[0] !=3D '/') { > + kfree(b); > + return -EINVAL; > + } > + > + b->threshold_us =3D threshold_us; > + b->offset_start =3D offset_start; > + b->offset_stop=C2=A0 =3D offset_stop; > + > + ret =3D kern_path(binpath, LOOKUP_FOLLOW, &b->path); > + if (ret) > + goto err_free; > + > + if (!d_is_reg(b->path.dentry)) { > + ret =3D -EINVAL; > + goto err_path; > + } > + > + /* Reject duplicate start offset for the same binary. */ > + list_for_each_entry(tmp_b, &tlob_uprobe_list, list) { > + if (tmp_b->offset_start =3D=3D offset_start && > + =C2=A0=C2=A0=C2=A0 tmp_b->path.dentry =3D=3D b->path.dentry) { > + ret =3D -EEXIST; > + goto err_path; > + } > + } > + > + /* Store canonical path for read-back and removal matching. */ > + canon =3D d_path(&b->path, pathbuf, sizeof(pathbuf)); > + if (IS_ERR(canon)) { > + ret =3D PTR_ERR(canon); > + goto err_path; > + } > + strscpy(b->binpath, canon, sizeof(b->binpath)); > + > + b->entry_uc.handler =3D tlob_uprobe_entry_handler; > + b->stop_uc.handler=C2=A0 =3D tlob_uprobe_stop_handler; > + > + inode =3D d_real_inode(b->path.dentry); > + > + b->entry_uprobe =3D uprobe_register(inode, offset_start, 0, &b- > >entry_uc); > + if (IS_ERR(b->entry_uprobe)) { > + ret =3D PTR_ERR(b->entry_uprobe); > + b->entry_uprobe =3D NULL; > + goto err_path; > + } > + > + b->stop_uprobe =3D uprobe_register(inode, offset_stop, 0, &b->stop_uc); > + if (IS_ERR(b->stop_uprobe)) { > + ret =3D PTR_ERR(b->stop_uprobe); > + b->stop_uprobe =3D NULL; > + goto err_entry; > + } > + > + list_add_tail(&b->list, &tlob_uprobe_list); > + return 0; > + > +err_entry: > + uprobe_unregister_nosync(b->entry_uprobe, &b->entry_uc); > + uprobe_unregister_sync(); > +err_path: > + path_put(&b->path); > +err_free: > + kfree(b); > + return ret; > +} > + > +/* > + * Remove the uprobe binding for (offset_start, binpath). > + * binpath is resolved to a dentry for comparison so symlinks are handle= d > + * correctly.=C2=A0 Called with tlob_uprobe_mutex held. > + */ > +static void tlob_remove_uprobe_by_key(loff_t offset_start, const char > *binpath) > +{ > + struct tlob_uprobe_binding *b, *tmp; > + struct path remove_path; > + > + if (kern_path(binpath, LOOKUP_FOLLOW, &remove_path)) > + return; > + > + list_for_each_entry_safe(b, tmp, &tlob_uprobe_list, list) { > + if (b->offset_start !=3D offset_start) > + continue; > + if (b->path.dentry !=3D remove_path.dentry) > + continue; > + uprobe_unregister_nosync(b->entry_uprobe, &b->entry_uc); > + uprobe_unregister_nosync(b->stop_uprobe,=C2=A0 &b->stop_uc); > + list_del(&b->list); > + uprobe_unregister_sync(); > + path_put(&b->path); > + kfree(b); > + break; > + } > + > + path_put(&remove_path); > +} > + > +/* Unregister all uprobe bindings; called from disable_tlob(). */ > +static void tlob_remove_all_uprobes(void) > +{ > + struct tlob_uprobe_binding *b, *tmp; > + > + mutex_lock(&tlob_uprobe_mutex); > + list_for_each_entry_safe(b, tmp, &tlob_uprobe_list, list) { > + uprobe_unregister_nosync(b->entry_uprobe, &b->entry_uc); > + uprobe_unregister_nosync(b->stop_uprobe,=C2=A0 &b->stop_uc); > + list_del(&b->list); > + path_put(&b->path); > + kfree(b); > + } > + mutex_unlock(&tlob_uprobe_mutex); > + uprobe_unregister_sync(); > +} > + > +/* > + * tracefs "monitor" file > + * > + * Read:=C2=A0 one "threshold_us:0xoffset_start:0xoffset_stop:binary_pat= h\n" > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 line per registered uprobe = binding. > + * Write: "threshold_us:offset_start:offset_stop:binary_path" - add upro= be > binding > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 "-offset_start:binary_path"= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 - rem= ove uprobe > binding > + */ > + > +static ssize_t tlob_monitor_read(struct file *file, > + char __user *ubuf, > + size_t count, loff_t *ppos) > +{ > + /* pid(10) + threshold(20) + 2 offsets(2*18) + path(256) + delimiters > */ > + const int line_sz =3D TLOB_MAX_PATH + 72; > + struct tlob_uprobe_binding *b; > + char *buf, *p; > + int n =3D 0, buf_sz, pos =3D 0; > + ssize_t ret; > + > + mutex_lock(&tlob_uprobe_mutex); > + list_for_each_entry(b, &tlob_uprobe_list, list) > + n++; > + mutex_unlock(&tlob_uprobe_mutex); > + > + buf_sz =3D (n ? n : 1) * line_sz + 1; > + buf =3D kmalloc(buf_sz, GFP_KERNEL); > + if (!buf) > + return -ENOMEM; > + > + mutex_lock(&tlob_uprobe_mutex); > + list_for_each_entry(b, &tlob_uprobe_list, list) { > + p =3D b->binpath; > + pos +=3D scnprintf(buf + pos, buf_sz - pos, > + "%llu:0x%llx:0x%llx:%s\n", > + b->threshold_us, > + (unsigned long long)b->offset_start, > + (unsigned long long)b->offset_stop, > + p); > + } > + mutex_unlock(&tlob_uprobe_mutex); > + > + ret =3D simple_read_from_buffer(ubuf, count, ppos, buf, pos); > + kfree(buf); > + return ret; > +} > + > +/* > + * Parse "threshold_us:offset_start:offset_stop:binary_path". > + * binary_path comes last so it may freely contain ':'. > + * Returns 0 on success. > + */ > +VISIBLE_IF_KUNIT int tlob_parse_uprobe_line(char *buf, u64 *thr_out, > + =C2=A0=C2=A0=C2=A0 char **path_out, > + =C2=A0=C2=A0=C2=A0 loff_t *start_out, loff_t > *stop_out) > +{ > + unsigned long long thr; > + long long start, stop; > + int n =3D 0; > + > + /* > + * %llu : decimal-only (microseconds) > + * %lli : auto-base, accepts 0x-prefixed hex for offsets > + * %n=C2=A0=C2=A0 : records the byte offset of the first path character > + */ > + if (sscanf(buf, "%llu:%lli:%lli:%n", &thr, &start, &stop, &n) !=3D 3) > + return -EINVAL; > + if (thr =3D=3D 0 || n =3D=3D 0 || buf[n] =3D=3D '\0') > + return -EINVAL; > + if (start < 0 || stop < 0) > + return -EINVAL; > + > + *thr_out=C2=A0=C2=A0 =3D thr; > + *start_out =3D start; > + *stop_out=C2=A0 =3D stop; > + *path_out=C2=A0 =3D buf + n; > + return 0; > +} > + > +static ssize_t tlob_monitor_write(struct file *file, > + =C2=A0 const char __user *ubuf, > + =C2=A0 size_t count, loff_t *ppos) > +{ > + char buf[TLOB_MAX_PATH + 64]; > + loff_t offset_start, offset_stop; > + u64 threshold_us; > + char *binpath; > + int ret; > + > + if (count >=3D sizeof(buf)) > + return -EINVAL; > + if (copy_from_user(buf, ubuf, count)) > + return -EFAULT; > + buf[count] =3D '\0'; > + > + if (count > 0 && buf[count - 1] =3D=3D '\n') > + buf[count - 1] =3D '\0'; > + > + /* Remove request: "-offset_start:binary_path" */ > + if (buf[0] =3D=3D '-') { > + long long off; > + int n =3D 0; > + > + if (sscanf(buf + 1, "%lli:%n", &off, &n) !=3D 1 || n =3D=3D 0) > + return -EINVAL; > + binpath =3D buf + 1 + n; > + if (binpath[0] !=3D '/') > + return -EINVAL; > + > + mutex_lock(&tlob_uprobe_mutex); > + tlob_remove_uprobe_by_key((loff_t)off, binpath); > + mutex_unlock(&tlob_uprobe_mutex); > + > + return (ssize_t)count; > + } > + > + /* > + * Uprobe binding: > "threshold_us:offset_start:offset_stop:binary_path" > + * binpath points into buf at the start of the path field. > + */ > + ret =3D tlob_parse_uprobe_line(buf, &threshold_us, > + =C2=A0=C2=A0=C2=A0=C2=A0 &binpath, &offset_start, &offset_stop); > + if (ret) > + return ret; > + > + mutex_lock(&tlob_uprobe_mutex); > + ret =3D tlob_add_uprobe(threshold_us, binpath, offset_start, > offset_stop); > + mutex_unlock(&tlob_uprobe_mutex); > + return ret ? ret : (ssize_t)count; > +} > + > +static const struct file_operations tlob_monitor_fops =3D { > + .open =3D simple_open, > + .read =3D tlob_monitor_read, > + .write =3D tlob_monitor_write, > + .llseek =3D noop_llseek, > +}; > + > +/* > + * __tlob_init_monitor / __tlob_destroy_monitor - called with > rv_interface_lock > + * held (required by da_monitor_init/destroy via > rv_get/put_task_monitor_slot). > + */ > +static int __tlob_init_monitor(void) > +{ > + int i, retval; > + > + tlob_state_cache =3D kmem_cache_create("tlob_task_state", > + =C2=A0=C2=A0=C2=A0=C2=A0 sizeof(struct tlob_task_state), > + =C2=A0=C2=A0=C2=A0=C2=A0 0, 0, NULL); > + if (!tlob_state_cache) > + return -ENOMEM; > + > + for (i =3D 0; i < TLOB_HTABLE_SIZE; i++) > + INIT_HLIST_HEAD(&tlob_htable[i]); > + atomic_set(&tlob_num_monitored, 0); > + > + retval =3D da_monitor_init(); > + if (retval) { > + kmem_cache_destroy(tlob_state_cache); > + tlob_state_cache =3D NULL; > + return retval; > + } > + > + rv_this.enabled =3D 1; > + return 0; > +} > + > +static void __tlob_destroy_monitor(void) > +{ > + rv_this.enabled =3D 0; > + tlob_stop_all(); > + tlob_remove_all_uprobes(); > + /* > + * Drain pending call_rcu() callbacks from tlob_stop_all() before > + * destroying the kmem_cache. > + */ > + synchronize_rcu(); > + da_monitor_destroy(); > + kmem_cache_destroy(tlob_state_cache); > + tlob_state_cache =3D NULL; > +} > + > +/* > + * tlob_init_monitor / tlob_destroy_monitor - KUnit wrappers that acquir= e > + * rv_interface_lock, satisfying the lockdep_assert_held() inside > + * rv_get/put_task_monitor_slot(). > + */ > +VISIBLE_IF_KUNIT int tlob_init_monitor(void) > +{ > + int ret; > + > + mutex_lock(&rv_interface_lock); > + ret =3D __tlob_init_monitor(); > + mutex_unlock(&rv_interface_lock); > + return ret; > +} > +EXPORT_SYMBOL_IF_KUNIT(tlob_init_monitor); > + > +VISIBLE_IF_KUNIT void tlob_destroy_monitor(void) > +{ > + mutex_lock(&rv_interface_lock); > + __tlob_destroy_monitor(); > + mutex_unlock(&rv_interface_lock); > +} > +EXPORT_SYMBOL_IF_KUNIT(tlob_destroy_monitor); > + > +VISIBLE_IF_KUNIT int tlob_enable_hooks(void) > +{ > + rv_attach_trace_probe("tlob", sched_switch, handle_sched_switch); > + rv_attach_trace_probe("tlob", sched_wakeup, handle_sched_wakeup); > + return 0; > +} > +EXPORT_SYMBOL_IF_KUNIT(tlob_enable_hooks); > + > +VISIBLE_IF_KUNIT void tlob_disable_hooks(void) > +{ > + rv_detach_trace_probe("tlob", sched_switch, handle_sched_switch); > + rv_detach_trace_probe("tlob", sched_wakeup, handle_sched_wakeup); > +} > +EXPORT_SYMBOL_IF_KUNIT(tlob_disable_hooks); > + > +/* > + * enable_tlob / disable_tlob - called by rv_enable/disable_monitor() wh= ich > + * already holds rv_interface_lock; call the __ variants directly. > + */ > +static int enable_tlob(void) > +{ > + int retval; > + > + retval =3D __tlob_init_monitor(); > + if (retval) > + return retval; > + > + return tlob_enable_hooks(); > +} > + > +static void disable_tlob(void) > +{ > + tlob_disable_hooks(); > + __tlob_destroy_monitor(); > +} > + > +static struct rv_monitor rv_this =3D { > + .name =3D "tlob", > + .description =3D "Per-task latency-over-budget monitor.", > + .enable =3D enable_tlob, > + .disable =3D disable_tlob, > + .reset =3D da_monitor_reset_all, > + .enabled =3D 0, > +}; > + > +static int __init register_tlob(void) > +{ > + int ret; > + > + ret =3D rv_register_monitor(&rv_this, NULL); > + if (ret) > + return ret; > + > + if (rv_this.root_d) { > + tracefs_create_file("monitor", 0644, rv_this.root_d, NULL, > + =C2=A0=C2=A0=C2=A0 &tlob_monitor_fops); > + } > + > + return 0; > +} > + > +static void __exit unregister_tlob(void) > +{ > + rv_unregister_monitor(&rv_this); > +} > + > +module_init(register_tlob); > +module_exit(unregister_tlob); > + > +MODULE_LICENSE("GPL"); > +MODULE_AUTHOR("Wen Yang "); > +MODULE_DESCRIPTION("tlob: task latency over budget per-task monitor."); > diff --git a/kernel/trace/rv/monitors/tlob/tlob.h > b/kernel/trace/rv/monitors/tlob/tlob.h > new file mode 100644 > index 000000000..3438a6175 > --- /dev/null > +++ b/kernel/trace/rv/monitors/tlob/tlob.h > @@ -0,0 +1,145 @@ > +/* SPDX-License-Identifier: GPL-2.0 */ > +#ifndef _RV_TLOB_H > +#define _RV_TLOB_H > + > +/* > + * C representation of the tlob automaton, generated from tlob.dot via r= vgen > + * and extended with tlob_start_task()/tlob_stop_task() declarations. > + * For the format description see > Documentation/trace/rv/deterministic_automata.rst > + */ > + > +#include > +#include > + > +#define MONITOR_NAME tlob > + > +enum states_tlob { > + unmonitored_tlob, > + on_cpu_tlob, > + off_cpu_tlob, > + state_max_tlob, > +}; > + > +#define INVALID_STATE state_max_tlob > + > +enum events_tlob { > + trace_start_tlob, > + switch_in_tlob, > + switch_out_tlob, > + sched_wakeup_tlob, > + trace_stop_tlob, > + budget_expired_tlob, > + event_max_tlob, > +}; > + > +struct automaton_tlob { > + char *state_names[state_max_tlob]; > + char *event_names[event_max_tlob]; > + unsigned char function[state_max_tlob][event_max_tlob]; > + unsigned char initial_state; > + bool final_states[state_max_tlob]; > +}; > + > +static const struct automaton_tlob automaton_tlob =3D { > + .state_names =3D { > + "unmonitored", > + "on_cpu", > + "off_cpu", > + }, > + .event_names =3D { > + "trace_start", > + "switch_in", > + "switch_out", > + "sched_wakeup", > + "trace_stop", > + "budget_expired", > + }, > + .function =3D { > + /* unmonitored */ > + { > + on_cpu_tlob, /* trace_start=C2=A0=C2=A0=C2=A0 */ > + unmonitored_tlob, /* switch_in=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 */ > + unmonitored_tlob, /* switch_out=C2=A0=C2=A0=C2=A0=C2=A0 */ > + unmonitored_tlob, /* sched_wakeup=C2=A0=C2=A0 */ > + INVALID_STATE, /* trace_stop=C2=A0=C2=A0=C2=A0=C2=A0 */ > + INVALID_STATE, /* budget_expired */ > + }, > + /* on_cpu */ > + { > + INVALID_STATE, /* trace_start=C2=A0=C2=A0=C2=A0 */ > + INVALID_STATE, /* switch_in=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 */ > + off_cpu_tlob, /* switch_out=C2=A0=C2=A0=C2=A0=C2=A0 */ > + on_cpu_tlob, /* sched_wakeup=C2=A0=C2=A0 */ > + unmonitored_tlob, /* trace_stop=C2=A0=C2=A0=C2=A0=C2=A0 */ > + unmonitored_tlob, /* budget_expired */ > + }, > + /* off_cpu */ > + { > + INVALID_STATE, /* trace_start=C2=A0=C2=A0=C2=A0 */ > + on_cpu_tlob, /* switch_in=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 */ > + off_cpu_tlob, /* switch_out=C2=A0=C2=A0=C2=A0=C2=A0 */ > + off_cpu_tlob, /* sched_wakeup=C2=A0=C2=A0 */ > + unmonitored_tlob, /* trace_stop=C2=A0=C2=A0=C2=A0=C2=A0 */ > + unmonitored_tlob, /* budget_expired */ > + }, > + }, > + /* > + * final_states: unmonitored is the sole accepting state. > + * Violations are recorded via ntf_push and tlob_budget_exceeded. > + */ > + .initial_state =3D unmonitored_tlob, > + .final_states =3D { 1, 0, 0 }, > +}; > + > +/* Exported for use by the RV ioctl layer (rv_dev.c) */ > +int tlob_start_task(struct task_struct *task, u64 threshold_us, > + =C2=A0=C2=A0=C2=A0 struct file *notify_file, u64 tag); > +int tlob_stop_task(struct task_struct *task); > + > +/* Maximum number of concurrently monitored tasks (also used by KUnit). = */ > +#define TLOB_MAX_MONITORED 64U > + > +/* > + * Ring buffer constants (also published in UAPI for mmap size calculati= on). > + */ > +#define TLOB_RING_DEFAULT_CAP 64U /* records allocated at open()=C2=A0 *= / > +#define TLOB_RING_MIN_CAP 8U /* minimum accepted by mmap()=C2=A0=C2=A0 = */ > +#define TLOB_RING_MAX_CAP 4096U /* maximum accepted by mmap()=C2=A0=C2= =A0 */ > + > +/** > + * struct tlob_ring - per-fd mmap-capable violation ring buffer. > + * > + * Allocated as a contiguous page range at rv_open() time: > + *=C2=A0=C2=A0 page 0:=C2=A0=C2=A0=C2=A0 struct tlob_mmap_page=C2=A0 (sh= ared with userspace) > + *=C2=A0=C2=A0 pages 1-N: struct tlob_event[capacity] > + */ > +struct tlob_ring { > + struct tlob_mmap_page *page; > + struct tlob_event *data; > + u32 mask; > + spinlock_t lock; > + unsigned long base; > + unsigned int order; > +}; > + > +/** > + * struct rv_file_priv - per-fd private data for /dev/rv. > + */ > +struct rv_file_priv { > + struct tlob_ring ring; > + wait_queue_head_t waitq; > +}; > + > +#if IS_ENABLED(CONFIG_KUNIT) > +int tlob_init_monitor(void); > +void tlob_destroy_monitor(void); > +int tlob_enable_hooks(void); > +void tlob_disable_hooks(void); > +void tlob_event_push_kunit(struct rv_file_priv *priv, > + =C2=A0 const struct tlob_event *info); > +int tlob_parse_uprobe_line(char *buf, u64 *thr_out, > + =C2=A0=C2=A0 char **path_out, > + =C2=A0=C2=A0 loff_t *start_out, loff_t *stop_out); > +#endif /* CONFIG_KUNIT */ > + > +#endif /* _RV_TLOB_H */ > diff --git a/kernel/trace/rv/monitors/tlob/tlob_trace.h > b/kernel/trace/rv/monitors/tlob/tlob_trace.h > new file mode 100644 > index 000000000..b08d67776 > --- /dev/null > +++ b/kernel/trace/rv/monitors/tlob/tlob_trace.h > @@ -0,0 +1,42 @@ > +/* SPDX-License-Identifier: GPL-2.0 */ > + > +/* > + * Snippet to be included in rv_trace.h > + */ > + > +#ifdef CONFIG_RV_MON_TLOB > +/* > + * tlob uses the generic event_da_monitor_id and error_da_monitor_id eve= nt > + * classes so that both event classes are instantiated.=C2=A0 This avoid= s a > + * -Werror=3Dunused-variable warning that the compiler emits when a > + * DECLARE_EVENT_CLASS has no corresponding DEFINE_EVENT instance. > + * > + * The event_tlob tracepoint is defined here but the call-site in > + * da_handle_event() is overridden with a no-op macro below so that no > + * trace record is emitted on every scheduler context switch.=C2=A0 Budg= et > + * violations are reported via the dedicated tlob_budget_exceeded event. > + * > + * error_tlob IS kept active so that invalid DA transitions (programming > + * errors) are still visible in the ftrace ring buffer for debugging. > + */ > +DEFINE_EVENT(event_da_monitor_id, event_tlob, > + =C2=A0=C2=A0=C2=A0=C2=A0 TP_PROTO(int id, char *state, char *event, cha= r *next_state, > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 bool final_state), > + =C2=A0=C2=A0=C2=A0=C2=A0 TP_ARGS(id, state, event, next_state, final_st= ate)); > + > +DEFINE_EVENT(error_da_monitor_id, error_tlob, > + =C2=A0=C2=A0=C2=A0=C2=A0 TP_PROTO(int id, char *state, char *event), > + =C2=A0=C2=A0=C2=A0=C2=A0 TP_ARGS(id, state, event)); > + > +/* > + * Override the trace_event_tlob() call-site with a no-op after the > + * DEFINE_EVENT above has satisfied the event class instantiation > + * requirement.=C2=A0 The tracepoint symbol itself exists (and can be en= abled > + * via tracefs) but the automatic call from da_handle_event() is silence= d > + * to avoid per-context-switch ftrace noise during normal operation. > + */ > +#undef trace_event_tlob > +#define trace_event_tlob(id, state, event, next_state, final_state) \ > + do { (void)(id); (void)(state); (void)(event); \ > + =C2=A0=C2=A0=C2=A0=C2=A0 (void)(next_state); (void)(final_state); } whi= le (0) > +#endif /* CONFIG_RV_MON_TLOB */ > diff --git a/kernel/trace/rv/rv.c b/kernel/trace/rv/rv.c > index ee4e68102..e754e76d5 100644 > --- a/kernel/trace/rv/rv.c > +++ b/kernel/trace/rv/rv.c > @@ -148,6 +148,10 @@ > =C2=A0#include > =C2=A0#endif > =C2=A0 > +#ifdef CONFIG_RV_MON_TLOB > +EXPORT_TRACEPOINT_SYMBOL_GPL(tlob_budget_exceeded); > +#endif > + > =C2=A0#include "rv.h" > =C2=A0 > =C2=A0DEFINE_MUTEX(rv_interface_lock); > diff --git a/kernel/trace/rv/rv_dev.c b/kernel/trace/rv/rv_dev.c > new file mode 100644 > index 000000000..a052f3203 > --- /dev/null > +++ b/kernel/trace/rv/rv_dev.c > @@ -0,0 +1,602 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* > + * rv_dev.c - /dev/rv misc device for RV monitor self-instrumentation > + * > + * A single misc device (MISC_DYNAMIC_MINOR) serves all RV monitors. > + * ioctl numbers encode the monitor identity: > + * > + *=C2=A0=C2=A0 0x01 - 0x1F=C2=A0 tlob (task latency over budget) > + *=C2=A0=C2=A0 0x20 - 0x3F=C2=A0 reserved > + * > + * Each monitor exports tlob_start_task() / tlob_stop_task() which are > + * called here.=C2=A0 The calling task is identified by current. > + * > + * Magic: RV_IOC_MAGIC (0xB9), defined in include/uapi/linux/rv.h > + * > + * Per-fd private data (rv_file_priv) > + * ------------------------------------ > + * Every open() of /dev/rv allocates an rv_file_priv (defined in tlob.h)= . > + * When TLOB_IOCTL_TRACE_START is called with args.notify_fd >=3D 0, vio= lations > + * are pushed as tlob_event records into that fd's per-fd ring buffer > (tlob_ring) > + * and its poll/epoll waitqueue is woken. > + * > + * Consumers drain records with read() on the notify_fd; read() blocks u= ntil > + * at least one record is available (unless O_NONBLOCK is set). > + * > + * Per-thread "started" tracking (tlob_task_handle) > + * ------------------------------------------------- > + * tlob_stop_task() returns -ESRCH in two distinct situations: > + * > + *=C2=A0=C2=A0 (a) The deadline timer already fired and removed the tlob= hash-table > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 entry before TRACE_STOP arrived -= > budget was exceeded -> -EOVERFLOW > + * > + *=C2=A0=C2=A0 (b) TRACE_START was never called for this thread -> progr= amming error > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 -> -ESRCH > + * > + * To distinguish them, rv_dev.c maintains a lightweight hash table > + * (tlob_handles) that records a tlob_task_handle for every task_struct = * > + * for which a successful TLOB_IOCTL_TRACE_START has been > + * issued but the corresponding TLOB_IOCTL_TRACE_STOP has not yet arrive= d. > + * > + * tlob_task_handle is a thin "session ticket"=C2=A0 --=C2=A0 it carries= only the > + * task pointer and the owning file descriptor.=C2=A0 The heavy per-task= state > + * (hrtimer, DA state, threshold) lives in tlob_task_state inside tlob.c= . > + * > + * The table is keyed on task_struct * (same key as tlob.c), protected > + * by tlob_handles_lock (spinlock, irq-safe).=C2=A0 No get_task_struct() > + * refcount is needed here because tlob.c already holds a reference for > + * each live entry. > + * > + * Multiple threads may share the same fd.=C2=A0 Each thread has its own > + * tlob_task_handle in the table, so concurrent TRACE_START / TRACE_STOP > + * calls from different threads do not interfere. > + * > + * The fd release path (rv_release) calls tlob_stop_task() for every > + * handle in tlob_handles that belongs to the closing fd, ensuring clean= up > + * even if the user forgets to call TRACE_STOP. > + */ > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > + > +#ifdef CONFIG_RV_MON_TLOB > +#include "monitors/tlob/tlob.h" > +#endif > + > +/* ---------------------------------------------------------------------= -- > + * tlob_task_handle - per-thread session ticket for the ioctl interface > + * > + * One handle is allocated by TLOB_IOCTL_TRACE_START and freed by > + * TLOB_IOCTL_TRACE_STOP (or by rv_release if the fd is closed). > + * > + * @hlist:=C2=A0 Hash-table linkage in tlob_handles (keyed on task point= er). > + * @task:=C2=A0=C2=A0 The monitored thread.=C2=A0 Plain pointer; no refc= ount held here > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 because tlob.c = holds one for the lifetime of the monitoring > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 window, which e= ncompasses the lifetime of this handle. > + * @file:=C2=A0=C2=A0 The /dev/rv file descriptor that issued TRACE_STAR= T. > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Used by rv_rele= ase() to sweep orphaned handles on close(). > + * ---------------------------------------------------------------------= -- > + */ > +#define TLOB_HANDLES_BITS 5 > +#define TLOB_HANDLES_SIZE (1 << TLOB_HANDLES_BITS) > + > +struct tlob_task_handle { > + struct hlist_node hlist; > + struct task_struct *task; > + struct file *file; > +}; > + > +static struct hlist_head tlob_handles[TLOB_HANDLES_SIZE]; > +static DEFINE_SPINLOCK(tlob_handles_lock); > + > +static unsigned int tlob_handle_hash(const struct task_struct *task) > +{ > + return hash_ptr((void *)task, TLOB_HANDLES_BITS); > +} > + > +/* Must be called with tlob_handles_lock held. */ > +static struct tlob_task_handle * > +tlob_handle_find_locked(struct task_struct *task) > +{ > + struct tlob_task_handle *h; > + unsigned int slot =3D tlob_handle_hash(task); > + > + hlist_for_each_entry(h, &tlob_handles[slot], hlist) { > + if (h->task =3D=3D task) > + return h; > + } > + return NULL; > +} > + > +/* > + * tlob_handle_alloc - record that @task has an active monitoring sessio= n > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 opened via @file. > + * > + * Returns 0 on success, -EEXIST if @task already has a handle (double > + * TRACE_START without TRACE_STOP), -ENOMEM on allocation failure. > + */ > +static int tlob_handle_alloc(struct task_struct *task, struct file *file= ) > +{ > + struct tlob_task_handle *h; > + unsigned long flags; > + unsigned int slot; > + > + h =3D kmalloc(sizeof(*h), GFP_KERNEL); > + if (!h) > + return -ENOMEM; > + h->task =3D task; > + h->file =3D file; > + > + spin_lock_irqsave(&tlob_handles_lock, flags); > + if (tlob_handle_find_locked(task)) { > + spin_unlock_irqrestore(&tlob_handles_lock, flags); > + kfree(h); > + return -EEXIST; > + } > + slot =3D tlob_handle_hash(task); > + hlist_add_head(&h->hlist, &tlob_handles[slot]); > + spin_unlock_irqrestore(&tlob_handles_lock, flags); > + return 0; > +} > + > +/* > + * tlob_handle_free - remove the handle for @task and free it. > + * > + * Returns 1 if a handle existed (TRACE_START was called), 0 if not foun= d > + * (TRACE_START was never called for this thread). > + */ > +static int tlob_handle_free(struct task_struct *task) > +{ > + struct tlob_task_handle *h; > + unsigned long flags; > + > + spin_lock_irqsave(&tlob_handles_lock, flags); > + h =3D tlob_handle_find_locked(task); > + if (h) { > + hlist_del_init(&h->hlist); > + spin_unlock_irqrestore(&tlob_handles_lock, flags); > + kfree(h); > + return 1; > + } > + spin_unlock_irqrestore(&tlob_handles_lock, flags); > + return 0; > +} > + > +/* > + * tlob_handle_sweep_file - release all handles owned by @file. > + * > + * Called from rv_release() when the fd is closed without TRACE_STOP. > + * Calls tlob_stop_task() for each orphaned handle to drain the tlob > + * monitoring entries and prevent resource leaks in tlob.c. > + * > + * Handles are collected under the lock (short critical section), then > + * processed outside it (tlob_stop_task() may sleep/spin internally). > + */ > +#ifdef CONFIG_RV_MON_TLOB > +static void tlob_handle_sweep_file(struct file *file) > +{ > + struct tlob_task_handle *batch[TLOB_HANDLES_SIZE]; > + struct tlob_task_handle *h; > + struct hlist_node *tmp; > + unsigned long flags; > + int i, n =3D 0; > + > + spin_lock_irqsave(&tlob_handles_lock, flags); > + for (i =3D 0; i < TLOB_HANDLES_SIZE; i++) { > + hlist_for_each_entry_safe(h, tmp, &tlob_handles[i], hlist) { > + if (h->file =3D=3D file) { > + hlist_del_init(&h->hlist); > + batch[n++] =3D h; > + } > + } > + } > + spin_unlock_irqrestore(&tlob_handles_lock, flags); > + > + for (i =3D 0; i < n; i++) { > + /* > + * Ignore -ESRCH: the deadline timer may have already fired > + * and cleaned up the tlob entry. > + */ > + tlob_stop_task(batch[i]->task); > + kfree(batch[i]); > + } > +} > +#else > +static inline void tlob_handle_sweep_file(struct file *file) {} > +#endif /* CONFIG_RV_MON_TLOB */ > + > +/* ---------------------------------------------------------------------= -- > + * Ring buffer lifecycle > + * ---------------------------------------------------------------------= -- > + */ > + > +/* > + * tlob_ring_alloc - allocate a ring of @cap records (must be a power of= 2). > + * > + * Allocates a physically contiguous block of pages: > + *=C2=A0=C2=A0 page 0=C2=A0=C2=A0=C2=A0=C2=A0 : struct tlob_mmap_page=C2= =A0 (control page, shared with > userspace) > + *=C2=A0=C2=A0 pages 1..N : struct tlob_event[cap] (data pages) > + * > + * Each page is marked reserved so it can be mapped to userspace via mma= p(). > + */ > +static int tlob_ring_alloc(struct tlob_ring *ring, u32 cap) > +{ > + unsigned int total =3D PAGE_SIZE + cap * sizeof(struct tlob_event); > + unsigned int order =3D get_order(total); > + unsigned long base; > + unsigned int i; > + > + base =3D __get_free_pages(GFP_KERNEL | __GFP_ZERO, order); > + if (!base) > + return -ENOMEM; > + > + for (i =3D 0; i < (1u << order); i++) > + SetPageReserved(virt_to_page((void *)(base + i * > PAGE_SIZE))); > + > + ring->base=C2=A0 =3D base; > + ring->order =3D order; > + ring->page=C2=A0 =3D (struct tlob_mmap_page *)base; > + ring->data=C2=A0 =3D (struct tlob_event *)(base + PAGE_SIZE); > + ring->mask=C2=A0 =3D cap - 1; > + spin_lock_init(&ring->lock); > + > + ring->page->capacity=C2=A0=C2=A0=C2=A0 =3D cap; > + ring->page->version=C2=A0=C2=A0=C2=A0=C2=A0 =3D 1; > + ring->page->data_offset =3D PAGE_SIZE; > + ring->page->record_size =3D sizeof(struct tlob_event); > + return 0; > +} > + > +static void tlob_ring_free(struct tlob_ring *ring) > +{ > + unsigned int i; > + > + if (!ring->base) > + return; > + > + for (i =3D 0; i < (1u << ring->order); i++) > + ClearPageReserved(virt_to_page((void *)(ring->base + i * > PAGE_SIZE))); > + > + free_pages(ring->base, ring->order); > + ring->base =3D 0; > + ring->page =3D NULL; > + ring->data =3D NULL; > +} > + > +/* ---------------------------------------------------------------------= -- > + * File operations > + * ---------------------------------------------------------------------= -- > + */ > + > +static int rv_open(struct inode *inode, struct file *file) > +{ > + struct rv_file_priv *priv; > + int ret; > + > + priv =3D kzalloc(sizeof(*priv), GFP_KERNEL); > + if (!priv) > + return -ENOMEM; > + > + ret =3D tlob_ring_alloc(&priv->ring, TLOB_RING_DEFAULT_CAP); > + if (ret) { > + kfree(priv); > + return ret; > + } > + > + init_waitqueue_head(&priv->waitq); > + file->private_data =3D priv; > + return 0; > +} > + > +static int rv_release(struct inode *inode, struct file *file) > +{ > + struct rv_file_priv *priv =3D file->private_data; > + > + tlob_handle_sweep_file(file); > + tlob_ring_free(&priv->ring); > + kfree(priv); > + file->private_data =3D NULL; > + return 0; > +} > + > +static __poll_t rv_poll(struct file *file, poll_table *wait) > +{ > + struct rv_file_priv *priv =3D file->private_data; > + > + if (!priv) > + return EPOLLERR; > + > + poll_wait(file, &priv->waitq, wait); > + > + /* > + * Pairs with smp_store_release(&ring->page->data_head, ...) in > + * tlob_event_push().=C2=A0 No lock needed: head is written by the kern= el > + * producer and read here; tail is written by the consumer and we > only > + * need an approximate check for the poll fast path. > + */ > + if (smp_load_acquire(&priv->ring.page->data_head) !=3D > + =C2=A0=C2=A0=C2=A0 READ_ONCE(priv->ring.page->data_tail)) > + return EPOLLIN | EPOLLRDNORM; > + > + return 0; > +} > + > +/* > + * rv_read - consume tlob_event violation records from this fd's ring bu= ffer. > + * > + * Each read() returns a whole number of struct tlob_event records.=C2= =A0 @count > must > + * be at least sizeof(struct tlob_event); partial-record sizes are rejec= ted > with > + * -EINVAL. > + * > + * Blocking behaviour follows O_NONBLOCK on the fd: > + *=C2=A0=C2=A0 O_NONBLOCK clear: blocks until at least one record is ava= ilable. > + *=C2=A0=C2=A0 O_NONBLOCK set:=C2=A0=C2=A0 returns -EAGAIN immediately i= f the ring is empty. > + * > + * Returns the number of bytes copied (always a multiple of sizeof > tlob_event), > + * -EAGAIN if non-blocking and empty, or a negative error code. > + * > + * read() and mmap() share the same ring and data_tail cursor; do not us= e > + * both simultaneously on the same fd. > + */ > +static ssize_t rv_read(struct file *file, char __user *buf, size_t count= , > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 loff_t *ppos) > +{ > + struct rv_file_priv *priv =3D file->private_data; > + struct tlob_ring *ring; > + size_t rec =3D sizeof(struct tlob_event); > + unsigned long irqflags; > + ssize_t done =3D 0; > + int ret; > + > + if (!priv) > + return -ENODEV; > + > + ring =3D &priv->ring; > + > + if (count < rec) > + return -EINVAL; > + > + /* Blocking path: sleep until the producer advances data_head. */ > + if (!(file->f_flags & O_NONBLOCK)) { > + ret =3D wait_event_interruptible(priv->waitq, > + /* pairs with smp_store_release() in the producer */ > + smp_load_acquire(&ring->page->data_head) !=3D > + READ_ONCE(ring->page->data_tail)); > + if (ret) > + return ret; > + } > + > + /* > + * Drain records into the caller's buffer.=C2=A0 ring->lock serialises > + * concurrent read() callers and the softirq producer. > + */ > + while (done + rec <=3D count) { > + struct tlob_event record; > + u32 head, tail; > + > + spin_lock_irqsave(&ring->lock, irqflags); > + /* pairs with smp_store_release() in the producer */ > + head =3D smp_load_acquire(&ring->page->data_head); > + tail =3D ring->page->data_tail; > + if (head =3D=3D tail) { > + spin_unlock_irqrestore(&ring->lock, irqflags); > + break; > + } > + record =3D ring->data[tail & ring->mask]; > + WRITE_ONCE(ring->page->data_tail, tail + 1); > + spin_unlock_irqrestore(&ring->lock, irqflags); > + > + if (copy_to_user(buf + done, &record, rec)) > + return done ? done : -EFAULT; > + done +=3D rec; > + } > + > + return done ? done : -EAGAIN; > +} > + > +/* > + * rv_mmap - map the per-fd violation ring buffer into userspace. > + * > + * The mmap region covers the full ring allocation: > + * > + *=C2=A0=C2=A0 offset 0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0 : struct tlob_mmap_page=C2=A0 (control page) > + *=C2=A0=C2=A0 offset PAGE_SIZE=C2=A0 : struct tlob_event[capacity]=C2= =A0 (data pages) > + * > + * The caller must map exactly PAGE_SIZE + capacity * sizeof(struct > tlob_event) > + * bytes starting at offset 0 (vm_pgoff must be 0).=C2=A0 The actual cap= acity is > + * read from tlob_mmap_page.capacity after a successful mmap(2). > + * > + * Private mappings (MAP_PRIVATE) are rejected: the shared data_tail fie= ld > + * written by userspace must be visible to the kernel producer. > + */ > +static int rv_mmap(struct file *file, struct vm_area_struct *vma) > +{ > + struct rv_file_priv *priv =3D file->private_data; > + struct tlob_ring=C2=A0=C2=A0=C2=A0 *ring; > + unsigned long=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 size =3D vma->v= m_end - vma->vm_start; > + unsigned long=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ring_size; > + > + if (!priv) > + return -ENODEV; > + > + ring =3D &priv->ring; > + > + if (vma->vm_pgoff !=3D 0) > + return -EINVAL; > + > + ring_size =3D PAGE_ALIGN(PAGE_SIZE + ((unsigned long)(ring->mask + 1) * > + =C2=A0=C2=A0=C2=A0 sizeof(struct tlob_event))); > + if (size !=3D ring_size) > + return -EINVAL; > + > + if (!(vma->vm_flags & VM_SHARED)) > + return -EINVAL; > + > + return remap_pfn_range(vma, vma->vm_start, > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 page_to_pfn(virt_to_page((void *= )ring->base)), > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ring_size, vma->vm_page_prot); > +} > + > +/* ---------------------------------------------------------------------= -- > + * ioctl dispatcher > + * ---------------------------------------------------------------------= -- > + */ > + > +static long rv_ioctl(struct file *file, unsigned int cmd, unsigned long = arg) > +{ > + unsigned int nr =3D _IOC_NR(cmd); > + > + /* > + * Verify the magic byte so we don't accidentally handle ioctls > + * intended for a different device. > + */ > + if (_IOC_TYPE(cmd) !=3D RV_IOC_MAGIC) > + return -ENOTTY; > + > +#ifdef CONFIG_RV_MON_TLOB > + /* tlob: ioctl numbers 0x01 - 0x1F */ > + switch (cmd) { > + case TLOB_IOCTL_TRACE_START: { > + struct tlob_start_args args; > + struct file *notify_file =3D NULL; > + int ret, hret; > + > + if (copy_from_user(&args, > + =C2=A0=C2=A0 (struct tlob_start_args __user *)arg, > + =C2=A0=C2=A0 sizeof(args))) > + return -EFAULT; > + if (args.threshold_us =3D=3D 0) > + return -EINVAL; > + if (args.flags !=3D 0) > + return -EINVAL; > + > + /* > + * If notify_fd >=3D 0, resolve it to a file pointer. > + * fget() bumps the reference count; tlob.c drops it > + * via fput() when the monitoring window ends. > + * Reject non-/dev/rv fds to prevent type confusion. > + */ > + if (args.notify_fd >=3D 0) { > + notify_file =3D fget(args.notify_fd); > + if (!notify_file) > + return -EBADF; > + if (notify_file->f_op !=3D file->f_op) { > + fput(notify_file); > + return -EINVAL; > + } > + } > + > + ret =3D tlob_start_task(current, args.threshold_us, > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 notify_file, args.tag); > + if (ret !=3D 0) { > + /* tlob.c did not take ownership; drop ref. */ > + if (notify_file) > + fput(notify_file); > + return ret; > + } > + > + /* > + * Record session handle.=C2=A0 Free any stale handle left by > + * a previous window whose deadline timer fired (timer > + * removes tlob_task_state but cannot touch tlob_handles). > + */ > + tlob_handle_free(current); > + hret =3D tlob_handle_alloc(current, file); > + if (hret < 0) { > + tlob_stop_task(current); > + return hret; > + } > + return 0; > + } > + case TLOB_IOCTL_TRACE_STOP: { > + int had_handle; > + int ret; > + > + /* > + * Atomically remove the session handle for current. > + * > + *=C2=A0=C2=A0 had_handle =3D=3D 0: TRACE_START was never called for > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 this thread -> caller bug = -> -ESRCH > + * > + *=C2=A0=C2=A0 had_handle =3D=3D 1: TRACE_START was called.=C2=A0 If > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 tlob_stop_task() now retur= ns > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 -ESRCH, the deadline timer= already > + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 fired -> budget exceeded -= > -EOVERFLOW > + */ > + had_handle =3D tlob_handle_free(current); > + if (!had_handle) > + return -ESRCH; > + > + ret =3D tlob_stop_task(current); > + return (ret =3D=3D -ESRCH) ? -EOVERFLOW : ret; > + } > + default: > + break; > + } > +#endif /* CONFIG_RV_MON_TLOB */ > + > + return -ENOTTY; > +} > + > +/* ---------------------------------------------------------------------= -- > + * Module init / exit > + * ---------------------------------------------------------------------= -- > + */ > + > +static const struct file_operations rv_fops =3D { > + .owner =3D THIS_MODULE, > + .open =3D rv_open, > + .release =3D rv_release, > + .read =3D rv_read, > + .poll =3D rv_poll, > + .mmap =3D rv_mmap, > + .unlocked_ioctl =3D rv_ioctl, > +#ifdef CONFIG_COMPAT > + .compat_ioctl =3D rv_ioctl, > +#endif > + .llseek =3D noop_llseek, > +}; > + > +/* > + * 0666: /dev/rv is a self-instrumentation device.=C2=A0 All ioctls oper= ate > + * exclusively on the calling task (current); no task can monitor anothe= r > + * via this interface.=C2=A0 Opening the device does not grant any privi= lege > + * beyond observing one's own latency, so world-read/write is appropriat= e. > + */ > +static struct miscdevice rv_miscdev =3D { > + .minor =3D MISC_DYNAMIC_MINOR, > + .name =3D "rv", > + .fops =3D &rv_fops, > + .mode =3D 0666, > +}; > + > +static int __init rv_ioctl_init(void) > +{ > + int i; > + > + for (i =3D 0; i < TLOB_HANDLES_SIZE; i++) > + INIT_HLIST_HEAD(&tlob_handles[i]); > + > + return misc_register(&rv_miscdev); > +} > + > +static void __exit rv_ioctl_exit(void) > +{ > + misc_deregister(&rv_miscdev); > +} > + > +module_init(rv_ioctl_init); > +module_exit(rv_ioctl_exit); > + > +MODULE_LICENSE("GPL"); > +MODULE_DESCRIPTION("RV ioctl interface via /dev/rv"); > diff --git a/kernel/trace/rv/rv_trace.h b/kernel/trace/rv/rv_trace.h > index 4a6faddac..65d6c6485 100644 > --- a/kernel/trace/rv/rv_trace.h > +++ b/kernel/trace/rv/rv_trace.h > @@ -126,6 +126,7 @@ DECLARE_EVENT_CLASS(error_da_monitor_id, > =C2=A0#include > =C2=A0#include > =C2=A0#include > +#include > =C2=A0// Add new monitors based on CONFIG_DA_MON_EVENTS_ID here > =C2=A0 > =C2=A0#endif /* CONFIG_DA_MON_EVENTS_ID */ > @@ -202,6 +203,55 @@ TRACE_EVENT(rv_retries_error, > =C2=A0 __get_str(event), __get_str(name)) > =C2=A0); > =C2=A0#endif /* CONFIG_RV_MON_MAINTENANCE_EVENTS */ > + > +#ifdef CONFIG_RV_MON_TLOB > +/* > + * tlob_budget_exceeded - emitted when a monitored task exceeds its late= ncy > + * budget.=C2=A0 Carries the on-CPU / off-CPU time breakdown so that the= cause > + * of the overrun (CPU-bound vs. scheduling/I/O latency) is immediately > + * visible in the ftrace ring buffer without post-processing. > + */ > +TRACE_EVENT(tlob_budget_exceeded, > + > + TP_PROTO(struct task_struct *task, u64 threshold_us, > + u64 on_cpu_us, u64 off_cpu_us, u32 switches, > + bool state_is_on_cpu, u64 tag), > + > + TP_ARGS(task, threshold_us, on_cpu_us, off_cpu_us, switches, > + state_is_on_cpu, tag), > + > + TP_STRUCT__entry( > + __string(comm, task->comm) > + __field(pid_t, pid) > + __field(u64, threshold_us) > + __field(u64, on_cpu_us) > + __field(u64, off_cpu_us) > + __field(u32, switches) > + __field(bool, state_is_on_cpu) > + __field(u64, tag) > + ), > + > + TP_fast_assign( > + __assign_str(comm); > + __entry->pid =3D task->pid; > + __entry->threshold_us =3D threshold_us; > + __entry->on_cpu_us =3D on_cpu_us; > + __entry->off_cpu_us =3D off_cpu_us; > + __entry->switches =3D switches; > + __entry->state_is_on_cpu =3D state_is_on_cpu; > + __entry->tag =3D tag; > + ), > + > + TP_printk("%s[%d]: budget exceeded threshold=3D%llu on_cpu=3D%llu > off_cpu=3D%llu switches=3D%u state=3D%s tag=3D0x%016llx", > + __get_str(comm), __entry->pid, > + __entry->threshold_us, > + __entry->on_cpu_us, __entry->off_cpu_us, > + __entry->switches, > + __entry->state_is_on_cpu ? "on_cpu" : "off_cpu", > + __entry->tag) > +); > +#endif /* CONFIG_RV_MON_TLOB */ > + > =C2=A0#endif /* _TRACE_RV_H */ > =C2=A0 > =C2=A0/* This part must be outside protection */