Re: [RFC PATCH 33/33] Add Xen virtual block device driver.

[Jeremy Fitzhardinge <jeremy@goop.org>]

Dave Boutcher wrote:
> On Tue, 18 Jul 2006 00:00:33 -0700, Chris Wright <chrisw@sous-sol.org> said:
>   
>> The block device frontend driver allows the kernel to access block
>> devices exported exported by a virtual machine containing a physical
>> block device driver.
>>     
>
> First, I think this belongs in drivers/block (and the network driver
> belongs in drivers/net).  If we're going to bring xen to the party,
> lets not leave it hiding out in a corner.
>
>   
>> +static void connect(struct blkfront_info *);
>> +static void blkfront_closing(struct xenbus_device *);
>> +static int blkfront_remove(struct xenbus_device *);
>> +static int talk_to_backend(struct xenbus_device *, struct blkfront_info *);
>> +static int setup_blkring(struct xenbus_device *, struct blkfront_info *);
>> +
>> +static void kick_pending_request_queues(struct blkfront_info *);
>> +
>> +static irqreturn_t blkif_int(int irq, void *dev_id, struct pt_regs *ptregs);
>> +static void blkif_restart_queue(void *arg);
>> +static void blkif_recover(struct blkfront_info *);
>> +static void blkif_completion(struct blk_shadow *);
>> +static void blkif_free(struct blkfront_info *, int);
>>     
>
> I'm pretty sure you can rearrange the code to get rid of the forward
> references. 
>
>   
>> +/**
>> + * We are reconnecting to the backend, due to a suspend/resume, or a backend
>> + * driver restart.  We tear down our blkif structure and recreate it, but
>> + * leave the device-layer structures intact so that this is transparent to the
>> + * rest of the kernel.
>> + */
>> +static int blkfront_resume(struct xenbus_device *dev)
>> +{
>> +	struct blkfront_info *info = dev->dev.driver_data;
>> +	int err;
>> +
>> +	DPRINTK("blkfront_resume: %s\n", dev->nodename);
>> +
>> +	blkif_free(info, 1);
>> +
>> +	err = talk_to_backend(dev, info);
>> +	if (!err)
>> +		blkif_recover(info);
>> +
>> +	return err;
>> +}
>>     
> Should blkfront_resume grab blkif_io_lock?
>   

There should be no concurrent activity until info->connected has been 
set to BLKIF_STATE_CONNECTED, which doesn't happen until blkif_recover 
has completed successfully.  blkif_queue_request and blkif_int both test 
the connection state before doing anything.  (Not sure if a concurrent 
XenBus event can happen though.)

>   
>> +static inline int GET_ID_FROM_FREELIST(
>> +	struct blkfront_info *info)
>> +{
>> +	unsigned long free = info->shadow_free;
>> +	BUG_ON(free > BLK_RING_SIZE);
>> +	info->shadow_free = info->shadow[free].req.id;
>> +	info->shadow[free].req.id = 0x0fffffee; /* debug */
>> +	return free;
>> +}
>> +
>> +static inline void ADD_ID_TO_FREELIST(
>> +	struct blkfront_info *info, unsigned long id)
>> +{
>> +	info->shadow[id].req.id  = info->shadow_free;
>> +	info->shadow[id].request = 0;
>> +	info->shadow_free = id;
>> +}
>>     
>
> A real nit..but why are these routines SHOUTING?
>
>   
>> +int blkif_release(struct inode *inode, struct file *filep)
>> +{
>> +	struct blkfront_info *info = inode->i_bdev->bd_disk->private_data;
>> +	info->users--;
>> +	if (info->users == 0) {
>>     
>
> Hrm...this strikes me as racey.  Don't you need at least a memory
> barrier here to handle SMP?
>   
Hm.  Doesn't look good to me.

>> +static struct xlbd_major_info xvd_major_info = {
>> +	.major = 201,
>> +	.type = &xvd_type_info
>> +};
>>     
>
> I've forgotten what the current policy is around new major numbers. 
>   
This is wrong.  201 is allocated to Veritas, but 202 has been allocated 
for the Xen VBD.

    J