[PATCH] Driver core: Reduce the level of request_firmware() messages

[PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rjw@sisk.pl>

The messages from _request_firmware() informing that firmware is
being requested or built-in firmware is going to be used are printed
at KERN_INFO, which produces lots of noise on systems with huge
numbers of AMD CPUs.  Reduce the level of these messages to
KERN_DEBUG to get rid of that noise.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 drivers/base/firmware_class.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

Index: linux-2.6.32-SLE11-SP1/drivers/base/firmware_class.c
===================================================================
--- linux-2.6.32-SLE11-SP1.orig/drivers/base/firmware_class.c
+++ linux-2.6.32-SLE11-SP1/drivers/base/firmware_class.c
@@ -487,15 +487,14 @@ _request_firmware(const struct firmware 
 	     builtin++) {
 		if (strcmp(name, builtin->name))
 			continue;
-		dev_info(device, "firmware: using built-in firmware %s\n",
-			 name);
+		dev_dbg(device, "firmware: using built-in firmware %s\n", name);
 		firmware->size = builtin->size;
 		firmware->data = builtin->data;
 		return 0;
 	}
 
 	if (uevent)
-		dev_info(device, "firmware: requesting %s\n", name);
+		dev_dbg(device, "firmware: requesting %s\n", name);
 
 	retval = fw_setup_device(firmware, &f_dev, name, device, uevent);
 	if (retval)

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Greg KH]

On Sat, Feb 27, 2010 at 09:43:22PM +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rjw@sisk.pl>
> 
> The messages from _request_firmware() informing that firmware is
> being requested or built-in firmware is going to be used are printed
> at KERN_INFO, which produces lots of noise on systems with huge
> numbers of AMD CPUs.  Reduce the level of these messages to
> KERN_DEBUG to get rid of that noise.
> 
> Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>

Looks good, I'll queue it up on Tuesday.

thanks,

greg k-h

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Jaswinder Singh Rajput]

Hello Rafael,

On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> From: Rafael J. Wysocki <rjw@sisk.pl>
>
> The messages from _request_firmware() informing that firmware is
> being requested or built-in firmware is going to be used are printed
> at KERN_INFO, which produces lots of noise on systems with huge
> numbers of AMD CPUs.  Reduce the level of these messages to
> KERN_DEBUG to get rid of that noise.
>

Which firmware we are using is very useful information. Because of
huge numbers of CPUs it seems noise then better provide the
information for first cpu and for the rest of the CPUs you can show by
KERN_DEBUG.

Thanks,
--
Jaswinder Singh.

> Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
> ---
>  drivers/base/firmware_class.c |    5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
>
> Index: linux-2.6.32-SLE11-SP1/drivers/base/firmware_class.c
> ===================================================================
> --- linux-2.6.32-SLE11-SP1.orig/drivers/base/firmware_class.c
> +++ linux-2.6.32-SLE11-SP1/drivers/base/firmware_class.c
> @@ -487,15 +487,14 @@ _request_firmware(const struct firmware
>             builtin++) {
>                if (strcmp(name, builtin->name))
>                        continue;
> -               dev_info(device, "firmware: using built-in firmware %s\n",
> -                        name);
> +               dev_dbg(device, "firmware: using built-in firmware %s\n", name);
>                firmware->size = builtin->size;
>                firmware->data = builtin->data;
>                return 0;
>        }
>
>        if (uevent)
> -               dev_info(device, "firmware: requesting %s\n", name);
> +               dev_dbg(device, "firmware: requesting %s\n", name);
>
>        retval = fw_setup_device(firmware, &f_dev, name, device, uevent);
>        if (retval)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> Hello Rafael,
> 
> On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > From: Rafael J. Wysocki <rjw@sisk.pl>
> >
> > The messages from _request_firmware() informing that firmware is
> > being requested or built-in firmware is going to be used are printed
> > at KERN_INFO, which produces lots of noise on systems with huge
> > numbers of AMD CPUs.  Reduce the level of these messages to
> > KERN_DEBUG to get rid of that noise.
> >
> 
> Which firmware we are using is very useful information. Because of
> huge numbers of CPUs it seems noise then better provide the
> information for first cpu and for the rest of the CPUs you can show by
> KERN_DEBUG.

That would have been better indeed, but the problem is _request_firmware()
doesn't allow us to change the level of its messages on demand.

Rafael

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Jaswinder Singh Rajput]

Hello Rafael,

On Sun, 2010-02-28 at 13:13 +0100, Rafael J. Wysocki wrote:
> On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> > On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > >
> > > The messages from _request_firmware() informing that firmware is
> > > being requested or built-in firmware is going to be used are printed
> > > at KERN_INFO, which produces lots of noise on systems with huge
> > > numbers of AMD CPUs.  Reduce the level of these messages to
> > > KERN_DEBUG to get rid of that noise.
> > >
> > 
> > Which firmware we are using is very useful information. Because of
> > huge numbers of CPUs it seems noise then better provide the
> > information for first cpu and for the rest of the CPUs you can show by
> > KERN_DEBUG.
> 
> That would have been better indeed, but the problem is _request_firmware()
> doesn't allow us to change the level of its messages on demand.

Can we try this :

	if (smp_processor_id())
		dev_dbg(..);
	else
		dev_info(..);

Thanks,
--
Jaswinder Singh.

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> Hello Rafael,
> 
> On Sun, 2010-02-28 at 13:13 +0100, Rafael J. Wysocki wrote:
> > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> > > On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > > >
> > > > The messages from _request_firmware() informing that firmware is
> > > > being requested or built-in firmware is going to be used are printed
> > > > at KERN_INFO, which produces lots of noise on systems with huge
> > > > numbers of AMD CPUs.  Reduce the level of these messages to
> > > > KERN_DEBUG to get rid of that noise.
> > > >
> > > 
> > > Which firmware we are using is very useful information. Because of
> > > huge numbers of CPUs it seems noise then better provide the
> > > information for first cpu and for the rest of the CPUs you can show by
> > > KERN_DEBUG.
> > 
> > That would have been better indeed, but the problem is _request_firmware()
> > doesn't allow us to change the level of its messages on demand.
> 
> Can we try this :
> 
> 	if (smp_processor_id())
> 		dev_dbg(..);
> 	else
> 		dev_info(..);

Well, it doesn't look particularly nice, does it?

Besides, say we're requesting firmware for a non-CPU device which happens
to run on CPU1.  Then, dev_dbg() will be used, which most likely is not what we
want.

Rafael

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Jaswinder Singh Rajput]

Hello Rafael,

On Mon, Mar 1, 2010 at 1:24 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
>>
>> On Sun, 2010-02-28 at 13:13 +0100, Rafael J. Wysocki wrote:
>> > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
>> > > On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>> > > > From: Rafael J. Wysocki <rjw@sisk.pl>
>> > > >
>> > > > The messages from _request_firmware() informing that firmware is
>> > > > being requested or built-in firmware is going to be used are printed
>> > > > at KERN_INFO, which produces lots of noise on systems with huge
>> > > > numbers of AMD CPUs.  Reduce the level of these messages to
>> > > > KERN_DEBUG to get rid of that noise.
>> > > >
>> > >
>> > > Which firmware we are using is very useful information. Because of
>> > > huge numbers of CPUs it seems noise then better provide the
>> > > information for first cpu and for the rest of the CPUs you can show by
>> > > KERN_DEBUG.
>> >
>> > That would have been better indeed, but the problem is _request_firmware()
>> > doesn't allow us to change the level of its messages on demand.
>>
>> Can we try this :
>>
>>       if (smp_processor_id())
>>               dev_dbg(..);
>>       else
>>               dev_info(..);
>
> Well, it doesn't look particularly nice, does it?
>
> Besides, say we're requesting firmware for a non-CPU device which happens
> to run on CPU1.  Then, dev_dbg() will be used, which most likely is not what we
> want.
>

Yes, you are right. But atleast can you try it once and show the
output with and without this.

Thanks,
--
Jaswinder.

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Greg KH]

On Mon, Mar 01, 2010 at 01:43:54AM +0530, Jaswinder Singh Rajput wrote:
> Hello Rafael,
> 
> On Mon, Mar 1, 2010 at 1:24 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> >>
> >> On Sun, 2010-02-28 at 13:13 +0100, Rafael J. Wysocki wrote:
> >> > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> >> > > On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >> > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> >> > > >
> >> > > > The messages from _request_firmware() informing that firmware is
> >> > > > being requested or built-in firmware is going to be used are printed
> >> > > > at KERN_INFO, which produces lots of noise on systems with huge
> >> > > > numbers of AMD CPUs.  Reduce the level of these messages to
> >> > > > KERN_DEBUG to get rid of that noise.
> >> > > >
> >> > >
> >> > > Which firmware we are using is very useful information. Because of
> >> > > huge numbers of CPUs it seems noise then better provide the
> >> > > information for first cpu and for the rest of the CPUs you can show by
> >> > > KERN_DEBUG.
> >> >
> >> > That would have been better indeed, but the problem is _request_firmware()
> >> > doesn't allow us to change the level of its messages on demand.
> >>
> >> Can we try this :
> >>
> >>       if (smp_processor_id())
> >>               dev_dbg(..);
> >>       else
> >>               dev_info(..);
> >
> > Well, it doesn't look particularly nice, does it?
> >
> > Besides, say we're requesting firmware for a non-CPU device which happens
> > to run on CPU1.  Then, dev_dbg() will be used, which most likely is not what we
> > want.
> >
> 
> Yes, you are right. But atleast can you try it once and show the
> output with and without this.

Was there ever a resolution to this?  Rafael, do you want me to apply
your original patch now?

thanks,

greg k-h

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

On Monday 15 March 2010, Greg KH wrote:
> On Mon, Mar 01, 2010 at 01:43:54AM +0530, Jaswinder Singh Rajput wrote:
> > Hello Rafael,
> > 
> > On Mon, Mar 1, 2010 at 1:24 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> > >>
> > >> On Sun, 2010-02-28 at 13:13 +0100, Rafael J. Wysocki wrote:
> > >> > On Sunday 28 February 2010, Jaswinder Singh Rajput wrote:
> > >> > > On Sun, Feb 28, 2010 at 2:13 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > >> > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > >> > > >
> > >> > > > The messages from _request_firmware() informing that firmware is
> > >> > > > being requested or built-in firmware is going to be used are printed
> > >> > > > at KERN_INFO, which produces lots of noise on systems with huge
> > >> > > > numbers of AMD CPUs.  Reduce the level of these messages to
> > >> > > > KERN_DEBUG to get rid of that noise.
> > >> > > >
> > >> > >
> > >> > > Which firmware we are using is very useful information. Because of
> > >> > > huge numbers of CPUs it seems noise then better provide the
> > >> > > information for first cpu and for the rest of the CPUs you can show by
> > >> > > KERN_DEBUG.
> > >> >
> > >> > That would have been better indeed, but the problem is _request_firmware()
> > >> > doesn't allow us to change the level of its messages on demand.
> > >>
> > >> Can we try this :
> > >>
> > >>       if (smp_processor_id())
> > >>               dev_dbg(..);
> > >>       else
> > >>               dev_info(..);
> > >
> > > Well, it doesn't look particularly nice, does it?
> > >
> > > Besides, say we're requesting firmware for a non-CPU device which happens
> > > to run on CPU1.  Then, dev_dbg() will be used, which most likely is not what we
> > > want.
> > >
> > 
> > Yes, you are right. But atleast can you try it once and show the
> > output with and without this.
> 
> Was there ever a resolution to this?

Nope.

> Rafael, do you want me to apply your original patch now?

I still think it's a good idea.  The information about what firmware is being
used doesn't seem to be very useful for anything but debugging, so I think
KERN_DEBUG is actually the right level for these messages.

Rafael

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Jaswinder Singh Rajput]

Hello Rafael,

On Tue, Mar 16, 2010 at 2:57 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>
> I still think it's a good idea.  The information about what firmware is being
> used doesn't seem to be very useful for anything but debugging, so I think
> KERN_DEBUG is actually the right level for these messages.
>

Can we use printk_once()

Thanks,
--
Jaswinder Singh.

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

On Tuesday 16 March 2010, Jaswinder Singh Rajput wrote:
> Hello Rafael,
> 
> On Tue, Mar 16, 2010 at 2:57 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >
> > I still think it's a good idea.  The information about what firmware is being
> > used doesn't seem to be very useful for anything but debugging, so I think
> > KERN_DEBUG is actually the right level for these messages.
> >
> 
> Can we use printk_once()

Yes, I guess so.

OK, I'll rework the patch.

Thanks,
Rafael

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Greg KH]

On Tue, Mar 16, 2010 at 05:04:09AM +0530, Jaswinder Singh Rajput wrote:
> Hello Rafael,
> 
> On Tue, Mar 16, 2010 at 2:57 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >
> > I still think it's a good idea.  The information about what firmware is being
> > used doesn't seem to be very useful for anything but debugging, so I think
> > KERN_DEBUG is actually the right level for these messages.
> >
> 
> Can we use printk_once()

No, I'd rather use dev_dbg() that way we can dynamically turn it on and
off as needed.  I took Rafael's original patch.

Funny thing is, someone else also sent the same patch, so there must be
others wanting this change :)

thanks,

greg k-h

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Jaswinder Singh Rajput]

Hello Rafael,

On Tue, Mar 16, 2010 at 5:10 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> On Tuesday 16 March 2010, Jaswinder Singh Rajput wrote:
>> Hello Rafael,
>>
>> On Tue, Mar 16, 2010 at 2:57 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>> >
>> > I still think it's a good idea.  The information about what firmware is being
>> > used doesn't seem to be very useful for anything but debugging, so I think
>> > KERN_DEBUG is actually the right level for these messages.
>> >
>>
>> Can we use printk_once()
>
> Yes, I guess so.
>
> OK, I'll rework the patch.
>

Good, so you need to prepare printk_once() with condition like in our
case so that it prints once for a particular name.

Thanks,
--
Jaswinder Singh.

Re: [PATCH] Driver core: Reduce the level of request_firmware() messages

[Rafael J. Wysocki]

On Tuesday 16 March 2010, Greg KH wrote:
> On Tue, Mar 16, 2010 at 05:04:09AM +0530, Jaswinder Singh Rajput wrote:
> > Hello Rafael,
> > 
> > On Tue, Mar 16, 2010 at 2:57 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > >
> > > I still think it's a good idea.  The information about what firmware is being
> > > used doesn't seem to be very useful for anything but debugging, so I think
> > > KERN_DEBUG is actually the right level for these messages.
> > >
> > 
> > Can we use printk_once()
> 
> No, I'd rather use dev_dbg() that way we can dynamically turn it on and
> off as needed.  I took Rafael's original patch.

OK, thanks!

> Funny thing is, someone else also sent the same patch, so there must be
> others wanting this change :)

I know SGI wants it at least.

Rafael