From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) (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 3C7B028F51E for ; Thu, 3 Jul 2025 06:24:16 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.133.124 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751523859; cv=none; b=c/oLAHDhuqzKGv4lkLV1ll2/E9ds5xxNCrx+r3TXYy0cXJzWr55EnOw9fi3TuMIVLVbIaiucRV+ATOSCsswKvx3jq5ts3f+euoxYHWSgA9Qk9wAooxExqKSjl8ZRUWRAeOYJ4lw6jfk/in5FR5ZTCtGIHVnIxWNaEeHXG93t/N0= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751523859; c=relaxed/simple; bh=dao9nTpC5C1q0JJDCBYH/D3fgkfHONt6s1hu70kqblQ=; h=Date:From:To:Cc:Subject:Message-ID:References:MIME-Version: In-Reply-To:Content-Type:Content-Disposition; b=sZcCmxwlalzKBC7KttcpGh9yXtaDdhKM1ECuohulOaga1ouSqCQVsK1aCxgCCUe3kRJ3UTrAVr+ifzg0e6eOBTIHXrfpzHGjUsXIVRVhYJ8s7tWa3qGKiWX/eHZZbrHocamA4skhQwdFg3x6n4zk6pQFLSYARu+v7LoVrHTEQJE= 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=OwccRZ4V; arc=none smtp.client-ip=170.10.133.124 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="OwccRZ4V" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1751523856; 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: in-reply-to:in-reply-to:references:references; bh=oNnWJxGDTYj2oLSlWuvTf5AvF7vq7QQGZ22ViGGEQZQ=; b=OwccRZ4Vxh5DvST4BMUgW1HUdqLBFGQUQZM8IJemps9xeyNn+i1bjIyLLHzIkfgc+knLrx oq2I4EKT2fXGv38kcl7Iz15z3f/WDdfYF+D3jpJYaA3cBq3X//p5ODYZmnqgNQRfZhlqeu GBDw58dC/CblSeyCbj+vA78Wr2TVVDE= 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-82-gOKLg_X1MjKHs_sGVoLOuQ-1; Thu, 03 Jul 2025 02:24:14 -0400 X-MC-Unique: gOKLg_X1MjKHs_sGVoLOuQ-1 X-Mimecast-MFC-AGG-ID: gOKLg_X1MjKHs_sGVoLOuQ_1751523853 Received: by mail-wr1-f69.google.com with SMTP id ffacd0b85a97d-3a6d1394b07so5009221f8f.3 for ; Wed, 02 Jul 2025 23:24:14 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1751523853; x=1752128653; h=in-reply-to:content-disposition:mime-version:references:message-id :subject:cc:to:from:date:x-gm-message-state:from:to:cc:subject:date :message-id:reply-to; bh=oNnWJxGDTYj2oLSlWuvTf5AvF7vq7QQGZ22ViGGEQZQ=; b=PV3TT9rDf00wEcyn5wr8va1Qu2aKnZdQSDOsAXlFVrFAYveLYcJtzdYE40E5a5F8qq vTDtIoL7y59t82405nPG9CNFxkttSF9NcF7CV0re9kUv4o1jEh5kFT4GaOyE5bP/hCie zA7K9DfjI2iFA3M1cDrxURDCTkgKuh1447YSaUa+srbLKK4twAjR9oz6FlD6UYJiIdwF CiM0bC9T4LkNPar/k3Djk5klw/ew+xP22L139GmepR+MD7rtI8s/9NoZOSSoeqsbaiwz Wz/VxkGPpPC2ESVijTaVhCpdvwcCeS4ppBw3akMrQIM3S3ILBMWPl+TXbE5vMW+AOkuq Zm5g== X-Forwarded-Encrypted: i=1; AJvYcCWtOWsYPns9h3GRjUKrxMPLhGzdFGcsfwGjw+jMzrhbNo5yonzUx+GxxBc7h9E2j6Sp8J/GlAq2bFdtGzNWJQ==@lists.linux.dev X-Gm-Message-State: AOJu0YxRNhMPyI+HFlg1zEtASE2YAdHnJGitI4+afyCTloob41oCwggJ 5QKYbpzL13f2ikbW1YULytk6ElMgmgv7YjzbaTxYjcTfXBPsawGqMACrn44UCfNUerd7W8o7YXg jatebqQ9KsKb600OLuARID0hFrfXXe/ze69T29GU/SjYI++I93daoXzT9+tYYBdKUufZB X-Gm-Gg: ASbGncvrTYyLhsgH1gzmeTmN1+0crtRG1C2l0ZLBhzMLHIjbQdTg0tVsKMqh0tvs4Zm SGYn/Y2iHtRl3+XW5I4eEXChu6FVPcy70rzR63YMmiSShVHJK/+PBr5XBSi1N6QwUU8TP/eCGsq /16tGpkvzKjD/QopcGxn1B+ufRXoeY0tTO9gQ1b5zPscttC+24+eNYIsvgAPKdIRj7QsFkoUz4E TbDfQdVMppC4NDs38AKk5A6BD/3rSEwXyW7OIlKqTNxnv50EGdWXOj/k/S61GfPUxwIBRXNJCEW cCqSjhToGn3TmTPk X-Received: by 2002:a05:6000:490a:b0:3b1:8db7:d1fc with SMTP id ffacd0b85a97d-3b1fe5c07d4mr4360122f8f.21.1751523853191; Wed, 02 Jul 2025 23:24:13 -0700 (PDT) X-Google-Smtp-Source: AGHT+IGGxmSX9oSYQVzW9zlIGbTQXs8hI4c+MXz2QtluF728gF5D2N2yEQOE0oBdUYHzNVwqbLlzgg== X-Received: by 2002:a05:6000:490a:b0:3b1:8db7:d1fc with SMTP id ffacd0b85a97d-3b1fe5c07d4mr4360105f8f.21.1751523852675; Wed, 02 Jul 2025 23:24:12 -0700 (PDT) Received: from redhat.com ([2a0d:6fc0:152e:1400:856d:9957:3ec3:1ddc]) by smtp.gmail.com with ESMTPSA id ffacd0b85a97d-3a88c7fb20esm18034642f8f.36.2025.07.02.23.24.11 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 02 Jul 2025 23:24:12 -0700 (PDT) Date: Thu, 3 Jul 2025 02:24:09 -0400 From: "Michael S. Tsirkin" To: Parav Pandit Cc: "linux-kernel@vger.kernel.org" , Bjorn Helgaas , "linux-pci@vger.kernel.org" , "stefanha@redhat.com" , "alok.a.tiwari@oracle.com" , "virtualization@lists.linux.dev" Subject: Re: [PATCH RFC v3] pci: report surprise removal event Message-ID: <20250703022224-mutt-send-email-mst@kernel.org> References: <1eac13450ade12cc98b15c5864e5bcd57f9e9882.1751440755.git.mst@redhat.com> <20250702132314-mutt-send-email-mst@kernel.org> Precedence: bulk X-Mailing-List: virtualization@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 In-Reply-To: X-Mimecast-Spam-Score: 0 X-Mimecast-MFC-PROC-ID: J70U1yKU9idGN06pJY-u0-EcubSp8DLUnQ7HqR0SJIE_1751523853 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Thu, Jul 03, 2025 at 05:02:13AM +0000, Parav Pandit wrote: > > > From: Michael S. Tsirkin > > Sent: 02 July 2025 10:54 PM > > > > On Wed, Jul 02, 2025 at 03:20:52AM -0400, Michael S. Tsirkin wrote: > > > At the moment, in case of a surprise removal, the regular remove > > > callback is invoked, exclusively. This works well, because mostly, > > > the cleanup would be the same. > > > > > > However, there's a race: imagine device removal was initiated by a > > > user action, such as driver unbind, and it in turn initiated some > > > cleanup and is now waiting for an interrupt from the device. If the > > > device is now surprise-removed, that never arrives and the remove > > > callback hangs forever. > > > > > > For example, this was reported for virtio-blk: > > > > > > 1. the graceful removal is ongoing in the remove() callback, where disk > > > deletion del_gendisk() is ongoing, which waits for the requests +to > > > complete, > > > > > > 2. Now few requests are yet to complete, and surprise removal started. > > > > > > At this point, virtio block driver will not get notified by the driver > > > core layer, because it is likely serializing remove() happening by > > > +user/driver unload and PCI hotplug driver-initiated device removal. > > So > > > vblk driver doesn't know that device is removed, block layer is waiting > > > for requests completions to arrive which it never gets. So > > > del_gendisk() gets stuck. > > > > > > Drivers can artificially add timeouts to handle that, but it can be > > > flaky. > > > > > > Instead, let's add a way for the driver to be notified about the > > > disconnect. It can then do any necessary cleanup, knowing that the > > > device is inactive. > > > > > > Since cleanups can take a long time, this takes an approach of a work > > > struct that the driver initiates and enables on probe, and tears down > > > on remove. > > > > > > Signed-off-by: Michael S. Tsirkin > > > --- > > > > > > > Parav what do you think of this patch? > The async notification part without holding the device lock is good part of this patch. > > However, large part of the systems and use cases does not involve pci hot plug removal. > An average system that I came across using has 150+ pci devices, and none of them uses hotplug. > > So increasing pci dev struct for rare hot unplug, that too for the race condition does not look the best option. > > I believe the intent of async notification without device lock can be achieved by adding a non-blocking async notifier callback. > This can go in the pci ops struct. > > Such callback scale far better being part of the ops struct instead of pci_dev struct. Sorry, I don't see a way to achieve that, as the driver can go away while hotunplug happens. You would be welcome to try but you mentioned you have no plans to do so. > > This you can try using this in virtio blk to > > address the hang you reported? > > > The hang I reported was not the race condition between remove() and hotunplug during remove. > It was the simple remove() as hot-unplug issue due to commit 43bb40c5b926. > > The race condition hang is hard to reproduce as_is. > I can try to reproduce by adding extra sleep() etc code in remove() with v4 of this version with ops callback. > > However, that requires lot more code to be developed on top of current proposed fix [1]. > > [1] https://lore.kernel.org/linux-block/20250624185622.GB5519@fedora/ > > I need to re-arrange the hardware with hotplug resources. Will try to arrange on v4. > > > > Compile tested only. > > > > > > Note: this minimizes core code. I considered a more elaborate API that > > > would be easier to use, but decided to be conservative until there are > > > multiple users. > > > > > > changes from v2 > > > v2 was corrupted, fat fingers :( > > > > > > changes from v1: > > > switched to a WQ, with APIs to enable/disable > > > added motivation > > > > > > > > > drivers/pci/pci.h | 6 ++++++ > > > include/linux/pci.h | 27 +++++++++++++++++++++++++++ > > > 2 files changed, 33 insertions(+) > > > > > > diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h index > > > b81e99cd4b62..208b4cab534b 100644 > > > --- a/drivers/pci/pci.h > > > +++ b/drivers/pci/pci.h > > > @@ -549,6 +549,12 @@ static inline int pci_dev_set_disconnected(struct > > pci_dev *dev, void *unused) > > > pci_dev_set_io_state(dev, pci_channel_io_perm_failure); > > > pci_doe_disconnected(dev); > > > > > > + if (READ_ONCE(dev->disconnect_work_enable)) { > > > + /* Make sure work is up to date. */ > > > + smp_rmb(); > > > + schedule_work(&dev->disconnect_work); > > > + } > > > + > > > return 0; > > > } > > > > > > diff --git a/include/linux/pci.h b/include/linux/pci.h index > > > 51e2bd6405cd..b2168c5d0679 100644 > > > --- a/include/linux/pci.h > > > +++ b/include/linux/pci.h > > > @@ -550,6 +550,10 @@ struct pci_dev { > > > /* These methods index pci_reset_fn_methods[] */ > > > u8 reset_methods[PCI_NUM_RESET_METHODS]; /* In priority order */ > > > > > > + /* Report disconnect events */ > > > + u8 disconnect_work_enable; > > > + struct work_struct disconnect_work; > > > + > > > > #ifdef CONFIG_PCIE_TPH > > > u16 tph_cap; /* TPH capability offset */ > > > u8 tph_mode; /* TPH mode */ > > > @@ -2657,6 +2661,29 @@ static inline bool pci_is_dev_assigned(struct > > pci_dev *pdev) > > > return (pdev->dev_flags & PCI_DEV_FLAGS_ASSIGNED) == > > > PCI_DEV_FLAGS_ASSIGNED; } > > > > > > +/* > > > + * Caller must initialize @pdev->disconnect_work before invoking this. > > > + * Caller also must check pci_device_is_present afterwards, since > > > + * if device is already gone when this is called, work will not run. > > > + */ > > > +static inline void pci_set_disconnect_work(struct pci_dev *pdev) { > > > + /* Make sure WQ has been initialized already */ > > > + smp_wmb(); > > > + > > > + WRITE_ONCE(pdev->disconnect_work_enable, 0x1); } > > > + > > > +static inline void pci_clear_disconnect_work(struct pci_dev *pdev) { > > > + WRITE_ONCE(pdev->disconnect_work_enable, 0x0); > > > + > > > + /* Make sure to stop using work from now on. */ > > > + smp_wmb(); > > > + > > > + cancel_work_sync(&pdev->disconnect_work); > > > +} > > > + > > > /** > > > * pci_ari_enabled - query ARI forwarding status > > > * @bus: the PCI bus > > > -- > > > MST