Re: [PATCH 1/1] Drivers: hv: vmbus: enable VMBus protocol version 5.0

[Dan Carpenter]

On Mon, May 14, 2018 at 11:17:55AM -0700, Stephen Hemminger wrote:
> On Mon, 14 May 2018 18:14:15 +
> Dexuan Cui  wrote:
> 
> > > From: devel  On Behalf Of
> > > Stephen Hemminger
> > > Sent: Sunday, May 13, 2018 10:24  
> > > > ...
> > > > @@ -372,6 +400,18 @@ int vmbus_post_msg(void *buffer, size_t buflen,  
> > > bool can_sleep)  
> > > > ...
> > > > +   hdr = (struct vmbus_channel_message_header 
> > > > *)buffer;  
> > > 
> > > Hate to pick o the details, but buffer is void * so cast is not necessary 
> > > here.  
> > 
> > Yes, it's unnecessary in C, though it's necessary in C++.
> > 
> > I found the patch went into char-misc 4 hours ago, so it looks we may
> > as well leave it as is. IMHO an explicit cast is not a bad thing. :-)
> > 
> > Thanks,
> > -- Dexuan
> 
> Kernel developers like to be concise. In fact there is a smatch script that 
> perodically
> gets run and more cleanup patches get sent.

It's a Coccinelle script, not Smatch.  Coccinelle generates patches
automatically so it's a better tool for cleanup than Smatch.

I would generate a lot more Smatch information if there was a way to
integrate it easily into a code editor.  For example, we could highlight
unecessary casts or pointer dereferences where Smatch wasn't 100% sure
if it was correct.  Or you could hover over function name to see what
resources it allocates.

regards,
dan carpenter

___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel

RE: [PATCH 1/1] Drivers: hv: vmbus: enable VMBus protocol version 5.0

[Dexuan Cui]

> From: Stephen Hemminger 
> Sent: Monday, May 14, 2018 11:18
> To: Dexuan Cui 
> > > ...
> > > Hate to pick o the details, but buffer is void * so cast is not necessary 
> > > here.
> >
> > Yes, it's unnecessary in C, though it's necessary in C++.
> >
> > I found the patch went into char-misc 4 hours ago, so it looks we may
> > as well leave it as is. IMHO an explicit cast is not a bad thing. :-)
> >
> > Thanks,
> > -- Dexuan
> 
> Kernel developers like to be concise. In fact there is a smatch script that
> perodically  gets run and more cleanup patches get sent.

I checked the "git log" and confimed you're correct: there are a lot of 
patches that removed the cast from "void *". :-)

Then let me post a small patch for this.

-- Dexuan


___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel

Re: [PATCH 1/1] Drivers: hv: vmbus: enable VMBus protocol version 5.0

[Stephen Hemminger]

On Mon, 14 May 2018 18:14:15 +
Dexuan Cui  wrote:

> > From: devel  On Behalf Of
> > Stephen Hemminger
> > Sent: Sunday, May 13, 2018 10:24  
> > > ...
> > > @@ -372,6 +400,18 @@ int vmbus_post_msg(void *buffer, size_t buflen,  
> > bool can_sleep)  
> > > ...
> > > + hdr = (struct vmbus_channel_message_header *)buffer;  
> > 
> > Hate to pick o the details, but buffer is void * so cast is not necessary 
> > here.  
> 
> Yes, it's unnecessary in C, though it's necessary in C++.
> 
> I found the patch went into char-misc 4 hours ago, so it looks we may
> as well leave it as is. IMHO an explicit cast is not a bad thing. :-)
> 
> Thanks,
> -- Dexuan

Kernel developers like to be concise. In fact there is a smatch script that 
perodically
gets run and more cleanup patches get sent.
___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel

RE: [PATCH 1/1] Drivers: hv: vmbus: enable VMBus protocol version 5.0

[Dexuan Cui]

> From: devel  On Behalf Of
> Stephen Hemminger
> Sent: Sunday, May 13, 2018 10:24
> > ...
> > @@ -372,6 +400,18 @@ int vmbus_post_msg(void *buffer, size_t buflen,
> bool can_sleep)
> > ...
> > +   hdr = (struct vmbus_channel_message_header *)buffer;
> 
> Hate to pick o the details, but buffer is void * so cast is not necessary 
> here.

Yes, it's unnecessary in C, though it's necessary in C++.

I found the patch went into char-misc 4 hours ago, so it looks we may
as well leave it as is. IMHO an explicit cast is not a bad thing. :-)

Thanks,
-- Dexuan
___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel

Re: [PATCH 1/1] Drivers: hv: vmbus: enable VMBus protocol version 5.0

[Stephen Hemminger]

On Sat, 12 May 2018 02:30:33 -0700
k...@linuxonhyperv.com wrote:

>  int vmbus_post_msg(void *buffer, size_t buflen, bool can_sleep)
>  {
> + struct vmbus_channel_message_header *hdr;
>   union hv_connection_id conn_id;
>   int ret = 0;
>   int retries = 0;
>   u32 usec = 1;
>  
>   conn_id.asu32 = 0;
> - conn_id.u.id = VMBUS_MESSAGE_CONNECTION_ID;
> + conn_id.u.id = vmbus_connection.msg_conn_id;
>  
>   /*
>* hv_post_message() can have transient failures because of
> @@ -372,6 +400,18 @@ int vmbus_post_msg(void *buffer, size_t buflen, bool 
> can_sleep)
>  
>   switch (ret) {
>   case HV_STATUS_INVALID_CONNECTION_ID:
> + /*
> +  * See vmbus_negotiate_version(): VMBus protocol 5.0
> +  * requires that we must use
> +  * VMBUS_MESSAGE_CONNECTION_ID_4 for the Initiate
> +  * Contact message, but on old hosts that only
> +  * support VMBus protocol 4.0 or lower, here we get
> +  * HV_STATUS_INVALID_CONNECTION_ID and we should
> +  * return an error immediately without retrying.
> +  */
> + hdr = (struct vmbus_channel_message_header *)buffer;

Hate to pick o the details, but buffer is void * so cast is not necessary here.
___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel