From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-yw1-f174.google.com (mail-yw1-f174.google.com [209.85.128.174]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id BA7631F03D9 for ; Sat, 25 Apr 2026 02:17:35 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.174 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777083457; cv=none; b=VxpqRzTgP8t5oPSthF4poKFIkwq45JvXjOxdPu/aeFeND+tSZ2EIdqgpvK041qmyaaT3w/6FHDzH0Seqi+baYsqxlS19/VaSz5DW5YHWCAEdn3rMNLQmXtFSOHMi4tbzbMCfXghBqiMETeVXImNIZhgZaiIfzgh660ina1WTqvo= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777083457; c=relaxed/simple; bh=B7eKxT0CklwU4a+KUGe/ivPvc9NgmkKrcfstup9BHH0=; h=Date:From:To:Cc:Subject:Message-ID:References:MIME-Version: Content-Type:Content-Disposition:In-Reply-To; b=HxuZ01HfXE/mBjnIuYNs0uqy0gtkcuC/5QwH0Vg9FeFZtZDjqq+G1twc7IcOFfejmC2tWKZaCNPGxM+Pbrhbhm4IwFTWjQYXQCa+ZZ9r4tJpJR771AUqA6VGJxvL1cRkTfCAWQbMhnPuxyO+e7OxbbFPk+EWvT5gYEQ81fcUJRU= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=FVKNVxxU; arc=none smtp.client-ip=209.85.128.174 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="FVKNVxxU" Received: by mail-yw1-f174.google.com with SMTP id 00721157ae682-793fdbb8d3aso89368387b3.3 for ; Fri, 24 Apr 2026 19:17:35 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1777083455; x=1777688255; darn=lists.linux.dev; h=in-reply-to:content-transfer-encoding:content-disposition :mime-version:references:message-id:subject:cc:to:from:date:from:to :cc:subject:date:message-id:reply-to; bh=yNwxkb+9lc8LbYT7XYKKUUPdOhjNGqAa21W0MQ/ZC/Q=; b=FVKNVxxU/Hxh5F44EdsgnmowkParegewcR56ktd5TmgUh47VZiMWHq82mLp8yblbaV abPdX7XJ7SrFpG7zaBbRqxz2sEOVuIz2mTw+zduk0bAE1AILIyWEK74fm+/Lqbu5boKr uAGvEaFSuxO8PrjK11qZKjDWCimTsh+AbpGdrElepXoJ0j8QJwiVvRtvXRsU9DHGKoUl s35ftKpYErpFYsLe7OEaSY7GsPg/Jw24BkNn/8rym9IVY26iZdQAv5OiMxAJNh7srPoX gte9Yxr90vXKNZK5ffGuBA8NnAgNRQUFZvA/BXQuPPSqK5f5oyzcM1CM3li7CI4IJ2B+ Uo2A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1777083455; x=1777688255; h=in-reply-to:content-transfer-encoding:content-disposition :mime-version:references:message-id:subject:cc:to:from:date:x-gm-gg :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=yNwxkb+9lc8LbYT7XYKKUUPdOhjNGqAa21W0MQ/ZC/Q=; b=GDI12mrexOmZFXAv08ayrxseJhVTjamBSVjierP1bfDTNsb6OUeeSVnh4bDh3vMtE/ Gx8CDtsclo0gXKu7qnrM7ql0x63NIKh+lkIoXAXVdqM/JwBAslmyqad2UFBoh1QB8bB2 0CgIUKiWVTWBTCXay5ROhhEMbb9zwEK8xHGX1xoeefG/hjegaC4Of4gryQH+ArpvPPr1 9DDL/tZIWhrO4bzmLyN4VylPlpducLYmryzwP6SX3qeRGsVx6SoRlIUrhmbi9zwZ911e ufkYz78sMsix7RhkPOWkGt+KwzeffKWZq7NH2g19DD/OuedyCAeNkCxZSTan6or1/5D5 19Dw== X-Gm-Message-State: AOJu0YyZ5qi0+fzF9h3I2Ue9rHAbloRvBoRPLQ3NXC35WiLnSawP0IB8 fDz3wAn0picjWrp3+VqY1bUx27gUJHlMT7aWe3ZrMYB5SHOffiVbI3NP X-Gm-Gg: AeBDietLOy6udqpbq5PpIZTe1CQO/l3hIcraXF0/yHdU2aPcqdjkVmtD59Y4+fJdg8y Vp1a+L/j4iPm1wgHZtH0b4LLbj4bqb12eI75TMyPidXvYYqUlUYgbo8OC03MyxwgZLQGejVCURT WSCCKi+xc72vmc/eR1N0E+kZ7V07prnkvbh2mTSEXAw0WeP36mGvE22MMOiA+ruTwPMpkKRPA2v QO1GK8/0PxB3Y45Pe7SM94It4H0NBAbFLqS4bPNiZSf9HYhYPOnz3W9KRAaxCBJI1oZuy64VSrV wzi8JH3iLQN/FqBdmQ5GlkBX4qGRRkE1cTL7FV9V4CCSHFiq9L3A8h6Kn0c25Ce6z4QMcwvyr3U MUVzxQtXqhweqEQw4X2bJy6XzFUp0q7ZjG5OJp3AW5DAUTALsb1C7UClnYK11hmKX7iQscxeiVX vEf7vRaz9ZzhR3AbV8VkYm1SWpwIPYyDe8ukXUHfdPOGb7PIme+84acVNGKF/wxQ3WcTi4bdCwJ LXEbg== X-Received: by 2002:a05:690c:a05c:b0:79f:4e02:50cc with SMTP id 00721157ae682-7b9ecfe1297mr273316837b3.42.1777083454774; Fri, 24 Apr 2026 19:17:34 -0700 (PDT) Received: from zenbox ([2600:1700:18fb:6011:cc96:e003:f6cc:4ed]) by smtp.gmail.com with ESMTPSA id 00721157ae682-7b9ee89abd7sm98290457b3.1.2026.04.24.19.17.34 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 24 Apr 2026 19:17:34 -0700 (PDT) Date: Fri, 24 Apr 2026 22:17:32 -0400 From: Justin Suess To: Alexei Starovoitov Cc: sashiko@lists.linux.dev, bpf Subject: Re: [PATCH bpf-next v3 1/2] bpf: Implement dtor for struct file BTF ID Message-ID: References: <20260424192252.695240-2-utilityemal77@gmail.com> <20260424195958.AED4EC19425@smtp.kernel.org> Precedence: bulk X-Mailing-List: sashiko@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: On Fri, Apr 24, 2026 at 06:25:31PM -0700, Alexei Starovoitov wrote: > On Fri, Apr 24, 2026 at 4:20 PM Justin Suess wrote: > > > > > > So either: > > > > 1. allocating memory for a new irq_work on the fly, which is an > > operation that can fail. So if we're under memory pressure the > > element never gets freed. > > > > 2. preallocate an irq work for every element ahead of time. > > obviously that's not what I was referring to. > irq_work is rarely used directly like that. > The common pattern is: > defer_free() > { > if (llist_add(head + s->offset, &df->objects)) > irq_work_queue(&df->work); > That's smart. So you can use it like a doorbell pattern and avoid re-calling it. And avoid locking with llist. > but before going there. > Please enumerate "other dtors broken in NMI". > What exactly is broken? > What is the sequence of events? > All map types with kptrs? or particular ones? Only referenced kptrs. I only tested hashmaps, not any other map types. But I don't see any reason why this wouldn't apply to other map types. Here are the specific dtors that know/suspect are buggy in NMI. 1. bpf_task_struct_release_dtor I confirmed. See below 2. bpf_kfree_skb_dtor can do call_rcu_hurry. 3. bpf_crypto_ctx_release_dtor does call_rcu. 4. bpf_cgroup_release_dtor frees via work queue and takes cgroup_mutex. bpf_cpumask_release_dtor I think is safe. I made a reproducer for one of these (task_struct dtor). I triggered the issue by deleting the last reference to a task_struct kptr within the tp_bpf/nmi_handler prog. `bpf_map_delete_elem()` ` -> htab_map_delete_elem()` ` -> free_htab_elem()` ` -> bpf_obj_free_fields()` ` -> bpf_task_release_dtor()` ` -> put_task_struct_rcu_user()` ` -> call_rcu()` Log: [ 1.358433] ================================ [ 1.358433] WARNING: inconsistent lock state [ 1.358434] 7.0.0-11169-ge4ef174588b8-dirty #16 Tainted: G OE [ 1.358435] -------------------------------- [ 1.358436] inconsistent {INITIAL USE} -> {IN-NMI} usage. [ 1.358436] test_progs/134 [HC1[1]:SC0[0]:HE0:SE1] takes: [ 1.358438] ffff8ae3bbc6f0e8 (&rdp->nocb_lock){....}-{2:2}, at: __call_rcu_common.constprop.0+0x316/0x740 This is done by loading a task_struct referenced kptr into a map and deleting it later from within a tp_btf:nmi_handler BPF_PROG_TYPE_TRACING program. For the task_struct case; I have a bug report: https://lore.kernel.org/bpf/20260421201035.1729473-1-utilityemal77@gmail.com/ With reproducer: https://gist.githubusercontent.com/RazeLighter777/5539336d79ab1854f9e9550c6dcab118/raw/082f1eeb2dd445936e64dd3a33861764690bde82/task_struct_dtor_deadlock.patch I didn't make a reproducer for other cases. Thanks for the tip with llist and irq_work, that blew my mind a bit. Justin