From: "K. Y. Srinivasan" <kys@microsoft.com> Date: 2011-08-31 21:56:31
Greg,
The TODO file for Hyper-V drivers has not been updated in a while and does
not reflect the current state of these drivers:
1) There are no more checkpatch warnings/errors in this code. One of the TODO
work items is "fix remaining checkpatch warnings and errors".
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
3) Like all other virtualization platforms, the protocol to communicate with
the host is very host specific. The ringbuffer control structures are shared
between the host and the guest and so, the guest is compelled to follow the
mandates of the host. Thus, it is not possible for us to merge the vmbus with
other virtual buses in the system. One of the TODO work items is "see if the
vmbus can be merged with the other virtual busses in the kernel".
4) A couple of months ago, we had posted the network driver for community review.
We have submitted patches (and these have been applied) to address the review
comments. One of the TODO work items is "audit the network driver"
5) Recently, we have merged the IDE and scsi drivers into a single driver that
can handle both IDE and SCSI devices. These patches have been already checked in.
As part of this effort, we have addressed all of the community comments on the
combined storage driver. The TODO file currently has two work items on this
issue: "audit the block driver" and "audit the scsi driver".
Greg, as I have been pestering you for some time now, we are very interested in
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
Regards,
K. Y
From: Peter Foley <hidden> Date: 2011-08-31 23:15:13
On Wed, 31 Aug 2011, K. Y. Srinivasan wrote:
Greg, as I have been pestering you for some time now, we are very interested in
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
An issue I've come across in the hyper-v drivers is the forced modular
build. You might want to see if the "depends on m" is still needed.
Thanks,
Peter Foley
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
Greg,
The TODO file for Hyper-V drivers has not been updated in a while and does
not reflect the current state of these drivers:
1) There are no more checkpatch warnings/errors in this code. One of the TODO
work items is "fix remaining checkpatch warnings and errors".
Great, you can delete it.
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
3) Like all other virtualization platforms, the protocol to communicate with
the host is very host specific. The ringbuffer control structures are shared
between the host and the guest and so, the guest is compelled to follow the
mandates of the host. Thus, it is not possible for us to merge the vmbus with
other virtual buses in the system. One of the TODO work items is "see if the
vmbus can be merged with the other virtual busses in the kernel".
Sure, you can delete it.
4) A couple of months ago, we had posted the network driver for community review.
We have submitted patches (and these have been applied) to address the review
comments. One of the TODO work items is "audit the network driver"
Leave that until the network subsystem authors agree that this is all
taken care of by merging the driver into their subsystem.
5) Recently, we have merged the IDE and scsi drivers into a single driver that
can handle both IDE and SCSI devices. These patches have been already checked in.
As part of this effort, we have addressed all of the community comments on the
combined storage driver. The TODO file currently has two work items on this
issue: "audit the block driver" and "audit the scsi driver".
Same as above for the network driver, but you can leave just one entry
now.
I can't apply patches right now due to the kernel.org infrastructure
reworking, so you will have to wait a few days for me to apply your
previous ones, but feel free to send a patch cleaning up the TODO file
now.
thanks,
greg k-h
On Wed, Aug 31, 2011 at 07:11:39PM -0400, Peter Foley wrote:
On Wed, 31 Aug 2011, K. Y. Srinivasan wrote:
quoted
Greg, as I have been pestering you for some time now, we are very interested in
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
An issue I've come across in the hyper-v drivers is the forced modular
build. You might want to see if the "depends on m" is still needed.
Ah, good point, KY, can you try this out and see if that can be changed?
If not, something is wrong and that needs to get fixed.
thanks,
greg k-h
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
- rename the vmbus_child_* functions to vmbus_* as there's no need to
think of "children" here.
- vmbus_onoffer comment is incorrect. You handle offers from lots of
other types. Or if not, am I reading the code incorrectly?
- the static hv_cb_utils array needs to go away. In the hv_utils.c
util_probe() call, properly register the channel callback, and the
same goes for the util_remove() call, unregister things there.
Note, you can use the driver_data field to determine exactly which
callback needs to be registered to make things easy. Something like
the following (pseudo code only):
static const struct hv_vmbus_device_id id_table[] = {
/* Shutdown guid */
{ VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49,
0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB),
.driver_data = &shutdown_onchannelcallback },
....
};
util_probe(struct hv_device *dev, const struct hv_vmbus_device_id *id)
[ Yes, you will have to change the probe callback signature, but that's fine. ]
{
void *fn(void *context);
u8 *buffer;
fn = id->driver_data;
buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
/* Hook up callback and buffer with a call to the proper vmbus
* function
*/
...
}
util_remove()
{
/* undo what you did in util_probe(), unhooking the callback and
* freeing the data */
}
Does that make any sense?
thanks,
greg k-h
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-01 20:57:41
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:28 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
Greg,
The TODO file for Hyper-V drivers has not been updated in a while and does
not reflect the current state of these drivers:
1) There are no more checkpatch warnings/errors in this code. One of the
TODO
quoted
work items is "fix remaining checkpatch warnings and errors".
Great, you can delete it.
Will do.
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
quoted
3) Like all other virtualization platforms, the protocol to communicate with
the host is very host specific. The ringbuffer control structures are shared
between the host and the guest and so, the guest is compelled to follow the
mandates of the host. Thus, it is not possible for us to merge the vmbus with
other virtual buses in the system. One of the TODO work items is "see if the
vmbus can be merged with the other virtual busses in the kernel".
Sure, you can delete it.
Will do.
quoted
4) A couple of months ago, we had posted the network driver for community
review.
quoted
We have submitted patches (and these have been applied) to address the
review
quoted
comments. One of the TODO work items is "audit the network driver"
Leave that until the network subsystem authors agree that this is all
taken care of by merging the driver into their subsystem.
Ok; will do.
quoted
5) Recently, we have merged the IDE and scsi drivers into a single driver that
can handle both IDE and SCSI devices. These patches have been already
checked in.
quoted
As part of this effort, we have addressed all of the community comments on
the
quoted
combined storage driver. The TODO file currently has two work items on this
issue: "audit the block driver" and "audit the scsi driver".
Same as above for the network driver, but you can leave just one entry
now.
Ok; will do.
I can't apply patches right now due to the kernel.org infrastructure
reworking, so you will have to wait a few days for me to apply your
previous ones, but feel free to send a patch cleaning up the TODO file
now.
Thanks Greg. I will get the patches out soon.
Regards,
K. Y
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-01 21:08:46
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
quoted
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
- rename the vmbus_child_* functions to vmbus_* as there's no need to
think of "children" here.
Ok; will do.
- vmbus_onoffer comment is incorrect. You handle offers from lots of
other types. Or if not, am I reading the code incorrectly?
You are right; I will clean this up.
- the static hv_cb_utils array needs to go away. In the hv_utils.c
util_probe() call, properly register the channel callback, and the
same goes for the util_remove() call, unregister things there.
Note, you can use the driver_data field to determine exactly which
callback needs to be registered to make things easy. Something like
the following (pseudo code only):
static const struct hv_vmbus_device_id id_table[] = {
/* Shutdown guid */
{ VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49,
0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB),
.driver_data = &shutdown_onchannelcallback },
....
};
util_probe(struct hv_device *dev, const struct hv_vmbus_device_id *id)
[ Yes, you will have to change the probe callback signature, but that's fine. ]
{
void *fn(void *context);
u8 *buffer;
fn = id->driver_data;
buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
/* Hook up callback and buffer with a call to the proper vmbus
* function
*/
...
}
util_remove()
{
/* undo what you did in util_probe(), unhooking the callback and
* freeing the data */
}
Does that make any sense?
Yes; I will get these patches to you soon. I see how the data ptr can be used;
I will buy you the beer the next time we meet.
Regards,
K. Y
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-02 14:09:59
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: Peter Foley
Cc: KY Srinivasan; gregkh@suse.de; Linux Kernel Mailing List;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Wed, Aug 31, 2011 at 07:11:39PM -0400, Peter Foley wrote:
quoted
On Wed, 31 Aug 2011, K. Y. Srinivasan wrote:
quoted
Greg, as I have been pestering you for some time now, we are very
interested in
quoted
quoted
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
An issue I've come across in the hyper-v drivers is the forced modular
build. You might want to see if the "depends on m" is still needed.
Ah, good point, KY, can you try this out and see if that can be changed?
If not, something is wrong and that needs to get fixed.
I will check it and let you know. If there are issues, I fill fix them
Regards,
K. Y
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-02 15:17:21
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: Peter Foley
Cc: KY Srinivasan; gregkh@suse.de; Linux Kernel Mailing List;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Wed, Aug 31, 2011 at 07:11:39PM -0400, Peter Foley wrote:
quoted
On Wed, 31 Aug 2011, K. Y. Srinivasan wrote:
quoted
Greg, as I have been pestering you for some time now, we are very
interested in
quoted
quoted
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
An issue I've come across in the hyper-v drivers is the forced modular
build. You might want to see if the "depends on m" is still needed.
Ah, good point, KY, can you try this out and see if that can be changed?
If not, something is wrong and that needs to get fixed.
Just verified that we can build into the kernel (not as modules) all hyper-V
drivers except hv_storvsc. hv_storvsc depends upon functionality that in my build
is being built as modules (SCSI etc.) and this is the reason why I could not build
hv_storvsc as part of the kernel.
Greg, do you want me to submit a patch to Kconfig, getting rid of the module
dependency (for hyperv-v drivers).
Regards,
K. Y
On Fri, Sep 02, 2011 at 03:17:19PM +0000, KY Srinivasan wrote:
quoted
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: Peter Foley
Cc: KY Srinivasan; gregkh@suse.de; Linux Kernel Mailing List;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Wed, Aug 31, 2011 at 07:11:39PM -0400, Peter Foley wrote:
quoted
On Wed, 31 Aug 2011, K. Y. Srinivasan wrote:
quoted
Greg, as I have been pestering you for some time now, we are very
interested in
quoted
quoted
exiting staging and we are willing to dedicate the necessary development
resources to address whatever issues that may still be pending. So, your help
in identifying what needs to be done will be greatly appreciated. To that end,
I think it will be useful to update the TODO file to reflect the current state of
the drivers. Let us know how we should proceed here.
An issue I've come across in the hyper-v drivers is the forced modular
build. You might want to see if the "depends on m" is still needed.
Ah, good point, KY, can you try this out and see if that can be changed?
If not, something is wrong and that needs to get fixed.
Just verified that we can build into the kernel (not as modules) all hyper-V
drivers except hv_storvsc. hv_storvsc depends upon functionality that in my build
is being built as modules (SCSI etc.) and this is the reason why I could not build
hv_storvsc as part of the kernel.
So 'make allyesconfig' with the module dependancy turned off doesn't
break the build? If so, great.
Greg, do you want me to submit a patch to Kconfig, getting rid of the module
dependency (for hyperv-v drivers).
On Thu, Sep 01, 2011 at 01:30:31PM -0700, Greg KH wrote:
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
quoted
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
Oh, here's another issue I forgot about. You can get rid of the 'ext'
field of the hv_device structure and use the proper function in the
driver core for this type of thing, just wrap it up in your own function
to make it "sane" looking, like USB and PCI does with their
usb_get_intfdata(), usb_set_intfdata(), pci_get_drvdata(), and
pci_set_drvdata() functions.
thanks,
greg k-h
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-02 19:47:16
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
quoted
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
- rename the vmbus_child_* functions to vmbus_* as there's no need to
think of "children" here.
- vmbus_onoffer comment is incorrect. You handle offers from lots of
other types. Or if not, am I reading the code incorrectly?
- the static hv_cb_utils array needs to go away. In the hv_utils.c
util_probe() call, properly register the channel callback, and the
same goes for the util_remove() call, unregister things there.
Note, you can use the driver_data field to determine exactly which
callback needs to be registered to make things easy. Something like
the following (pseudo code only):
static const struct hv_vmbus_device_id id_table[] = {
/* Shutdown guid */
{ VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49,
0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB),
.driver_data = &shutdown_onchannelcallback },
....
};
util_probe(struct hv_device *dev, const struct hv_vmbus_device_id *id)
[ Yes, you will have to change the probe callback signature, but that's fine. ]
Greg, I think if I understand you correctly, the id parameter to the probe function
would be pointing to relevant entry in the hv_vmbus_device_id array. Since the driver
can handle multiple ids (guids), we need to either in the driver specific probe function or
in the bus specific probe function, figure out which of the many devices the driver
supports is being probed. I have couple of implementation options and would appreciate
your preference:
1) Do the guid parsing at the vmbus level parsing function. If I go this route, the driver specific probe
function would get an extra parameter as you have described in pseudo code.
2) Do the guid parsing in the util probe function. In this case, we don't need to change the signature for
the probe function as all the id information is available in the (util) driver.
I personally would prefer the second approach since it does not affect other drivers (no need to change
the signature for the probe function). Let me know how you want me to proceed here.
Regards,
K. Y
On Fri, Sep 02, 2011 at 07:47:12PM +0000, KY Srinivasan wrote:
quoted
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
quoted
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work items is
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
- rename the vmbus_child_* functions to vmbus_* as there's no need to
think of "children" here.
- vmbus_onoffer comment is incorrect. You handle offers from lots of
other types. Or if not, am I reading the code incorrectly?
- the static hv_cb_utils array needs to go away. In the hv_utils.c
util_probe() call, properly register the channel callback, and the
same goes for the util_remove() call, unregister things there.
Note, you can use the driver_data field to determine exactly which
callback needs to be registered to make things easy. Something like
the following (pseudo code only):
static const struct hv_vmbus_device_id id_table[] = {
/* Shutdown guid */
{ VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49,
0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB),
.driver_data = &shutdown_onchannelcallback },
....
};
util_probe(struct hv_device *dev, const struct hv_vmbus_device_id *id)
[ Yes, you will have to change the probe callback signature, but that's fine. ]
Greg, I think if I understand you correctly, the id parameter to the probe function
would be pointing to relevant entry in the hv_vmbus_device_id array.
Yes, just like it does for USB, PCI, etc.
Since the driver can handle multiple ids (guids), we need to either in
the driver specific probe function or in the bus specific probe
function, figure out which of the many devices the driver supports is
being probed. I have couple of implementation options and would
appreciate your preference:
1) Do the guid parsing at the vmbus level parsing function. If I go
this route, the driver specific probe function would get an extra
parameter as you have described in pseudo code.
Yes, that's the proper way to do this, as your match function already
found the proper id structure, so you have the pointer, just pass it to
the probe function callback.
2) Do the guid parsing in the util probe function. In this case, we
don't need to change the signature for the probe function as all the
id information is available in the (util) driver.
Yes, but you end up doing the matching all over again in the util
driver, which isn't nice and ripe for duplication and bugs.
I personally would prefer the second approach since it does not affect
other drivers (no need to change the signature for the probe
function). Let me know how you want me to proceed here.
As you only have 3 probe functions, it's not a big deal to change them
all :)
thanks,
greg k-h
From: KY Srinivasan <kys@microsoft.com> Date: 2011-09-02 20:54:39
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Friday, September 02, 2011 3:58 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Fri, Sep 02, 2011 at 07:47:12PM +0000, KY Srinivasan wrote:
quoted
quoted
-----Original Message-----
From: Greg KH [mailto:greg@kroah.com]
Sent: Thursday, September 01, 2011 4:31 PM
To: KY Srinivasan
Cc: gregkh@suse.de; linux-kernel@vger.kernel.org;
devel@linuxdriverproject.org; virtualization@lists.osdl.org
Subject: Re: Hyper-V TODO file
On Thu, Sep 01, 2011 at 01:27:33PM -0700, Greg KH wrote:
quoted
On Wed, Aug 31, 2011 at 03:16:51PM -0700, K. Y. Srinivasan wrote:
quoted
2) With your help, we have fixed all of the issues related to these drivers
conforming to the Linux Device Driver model. One of the TODO work
items is
quoted
quoted
quoted
quoted
"audit the vmbus to verify it is working properly with the driver model".
I have a few comments about this, I'll respond in another email.
Ok, it looks a _lot_ better, but I have a few minor nits, and one larger
one:
- rename the vmbus_child_* functions to vmbus_* as there's no need to
think of "children" here.
- vmbus_onoffer comment is incorrect. You handle offers from lots of
other types. Or if not, am I reading the code incorrectly?
- the static hv_cb_utils array needs to go away. In the hv_utils.c
util_probe() call, properly register the channel callback, and the
same goes for the util_remove() call, unregister things there.
Note, you can use the driver_data field to determine exactly which
callback needs to be registered to make things easy. Something like
the following (pseudo code only):
static const struct hv_vmbus_device_id id_table[] = {
/* Shutdown guid */
{ VMBUS_DEVICE(0x31, 0x60, 0x0B, 0X0E, 0x13, 0x52, 0x34, 0x49,
0x81, 0x8B, 0x38, 0XD9, 0x0C, 0xED, 0x39, 0xDB),
.driver_data = &shutdown_onchannelcallback },
....
};
util_probe(struct hv_device *dev, const struct hv_vmbus_device_id *id)
[ Yes, you will have to change the probe callback signature, but that's fine. ]
Greg, I think if I understand you correctly, the id parameter to the probe
function
quoted
would be pointing to relevant entry in the hv_vmbus_device_id array.
Yes, just like it does for USB, PCI, etc.
quoted
Since the driver can handle multiple ids (guids), we need to either in
the driver specific probe function or in the bus specific probe
function, figure out which of the many devices the driver supports is
being probed. I have couple of implementation options and would
appreciate your preference:
1) Do the guid parsing at the vmbus level parsing function. If I go
this route, the driver specific probe function would get an extra
parameter as you have described in pseudo code.
Yes, that's the proper way to do this, as your match function already
found the proper id structure, so you have the pointer, just pass it to
the probe function callback.
quoted
2) Do the guid parsing in the util probe function. In this case, we
don't need to change the signature for the probe function as all the
id information is available in the (util) driver.
Yes, but you end up doing the matching all over again in the util
driver, which isn't nice and ripe for duplication and bugs.
quoted
I personally would prefer the second approach since it does not affect
other drivers (no need to change the signature for the probe
function). Let me know how you want me to proceed here.
As you only have 3 probe functions, it's not a big deal to change them
all :)
Ok; will do. I am glad I checked with you!
Regards,
K. Y