Message-ID:
 <SN6PR02MB4157C74E0E83E63175278153D4A22@SN6PR02MB4157.namprd02.prod.outlook.com>
Date: Sun, 30 Mar 2025 15:35:07 +0000
From: Michael Kelley <mhklinux@...look.com>
To: Naman Jain <namjain@...ux.microsoft.com>, "K . Y . Srinivasan"
	<kys@...rosoft.com>, Haiyang Zhang <haiyangz@...rosoft.com>, Wei Liu
	<wei.liu@...nel.org>, Dexuan Cui <decui@...rosoft.com>, Greg Kroah-Hartman
	<gregkh@...uxfoundation.org>, Stephen Hemminger <stephen@...workplumber.org>
CC: "linux-hyperv@...r.kernel.org" <linux-hyperv@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"stable@...nel.org" <stable@...nel.org>, Saurabh Sengar
	<ssengar@...ux.microsoft.com>
Subject: RE: [PATCH v3 1/2] uio_hv_generic: Fix sysfs creation path for ring
 buffer

From: Naman Jain <namjain@...ux.microsoft.com> Sent: Thursday, March 27, 2025 10:28 PM
> 
> On regular bootup, devices get registered to VMBus first, so when
> uio_hv_generic driver for a particular device type is probed,
> the device is already initialized and added, so sysfs creation in
> uio_hv_generic probe works fine. However, when device is removed
> and brought back, the channel rescinds and device again gets
> registered to VMBus. However this time, the uio_hv_generic driver is
> already registered to probe for that device and in this case sysfs
> creation is tried before the device's kobject gets initialized
> completely.
> 
> Fix this by moving the core logic of sysfs creation for ring buffer,
> from uio_hv_generic to HyperV's VMBus driver, where rest of the sysfs
> attributes for the channels are defined. While doing that, make use
> of attribute groups and macros, instead of creating sysfs directly,
> to ensure better error handling and code flow.
> 
> Problem path:
> vmbus_process_offer (new offer comes for the VMBus device)
>   vmbus_add_channel_work
>     vmbus_device_register
>       |-> device_register
>       |     |...
>       |     |-> hv_uio_probe
>       |           |...
>       |           |-> sysfs_create_bin_file (leads to a warning as
>       |                 primary channel's kobject, which is used to
>       |                 create sysfs is not yet initialized)
>       |-> kset_create_and_add
>       |-> vmbus_add_channel_kobj (initialization of primary channel's
>                                   kobject happens later)
> 
> Above code flow is sequential and the warning is always reproducible in
> this path.
> 
> Fixes: 9ab877a6ccf8 ("uio_hv_generic: make ring buffer attribute for primary channel")
> Cc: stable@...nel.org
> Suggested-by: Saurabh Sengar <ssengar@...ux.microsoft.com>
> Suggested-by: Michael Kelley <mhklinux@...look.com>
> Signed-off-by: Naman Jain <namjain@...ux.microsoft.com>
> ---
>  drivers/hv/hyperv_vmbus.h    |   6 ++
>  drivers/hv/vmbus_drv.c       | 110 ++++++++++++++++++++++++++++++++++-
>  drivers/uio/uio_hv_generic.c |  33 +++++------
>  include/linux/hyperv.h       |   6 ++
>  4 files changed, 134 insertions(+), 21 deletions(-)
> 

[snip]

> +/**
> + * hv_create_ring_sysfs() - create "ring" sysfs entry corresponding to ring buffers for a channel.
> + * @channel: Pointer to vmbus_channel structure
> + * @hv_mmap_ring_buffer: function pointer for initializing the function to be called on mmap of
> + *                       channel's "ring" sysfs node, which is for the ring buffer of that channel.
> + *                       Function pointer is of below type:
> + *                       int (*hv_mmap_ring_buffer)(struct vmbus_channel *channel,
> + *                                                  struct vm_area_struct *vma))
> + *                       This has a pointer to the channel and a pointer to vm_area_struct,
> + *                       used for mmap, as arguments.
> + *
> + * Sysfs node for ring buffer of a channel is created along with other fields, however its
> + * visibility is disabled by default. Sysfs creation needs to be controlled when the use-case
> + * is running.
> + * For example, HV_NIC device is used either by uio_hv_generic or hv_netvsc at any given point of
> + * time, and "ring" sysfs is needed only when uio_hv_generic is bound to that device. To avoid
> + * exposing the ring buffer by default, this function is reponsible to enable visibility of
> + * ring for userspace to use.
> + * Note: Race conditions can happen with userspace and it is not encouraged to create new
> + * use-cases for this. This was added to maintain backward compatibility, while solving
> + * one of the race conditions in uio_hv_generic while creating sysfs.
> + *
> + * Returns 0 on success or error code on failure.
> + */
> +int hv_create_ring_sysfs(struct vmbus_channel *channel,
> +			 int (*hv_mmap_ring_buffer)(struct vmbus_channel *channel,
> +						    struct vm_area_struct *vma))
> +{
> +	struct kobject *kobj = &channel->kobj;
> +	struct vmbus_channel *primary_channel = channel->primary_channel ?
> +		channel->primary_channel : channel;
> +
> +	channel->mmap_ring_buffer = hv_mmap_ring_buffer;
> +	channel->ring_sysfs_visible = true;
> +
> +	/*
> +	 * Skip updating the sysfs group if the primary channel is not yet initialized and sysfs
> +	 * group is not yet created. In those cases, the 'ring' will be created later in
> +	 * vmbus_device_register() -> vmbus_add_channel_kobj().
> +	 */
> +	if  (!primary_channel->device_obj->channels_kset)
> +		return 0;

This test doesn't accomplish what you want. It tests if the "channels" directory
has been created, but not if the numbered subdirectory for this channel has been
created. sysfs_update_group() operates on the numbered subdirectory and
could still fail because it hasn't been created yet.

My recommendation is to not try to do a test, and just let sysfs_update_group()
fail in that case (and ignore the error).

Michael

> +
> +	return sysfs_update_group(kobj, &vmbus_chan_group);
> +}
> +EXPORT_SYMBOL_GPL(hv_create_ring_sysfs);

Powered by blists - more mailing lists