Re: [PATCH] scsi: remove CDROM not ready printk

Re: [PATCH] scsi: remove CDROM not ready printk

[Andrew Morton]

On Tue, 27 May 2008 15:02:43 -0400 (EDT) Jerry Stralko <gerb.stralko@gmail.com> wrote:

> Hello,

let's cc a scsi list which exists ;)

>    This printk was spamming my dmesg and /var/log/message.  Is there a 
> reason we have this printk?  Can we simply remove it?
> 
> 
> Signed-off-by: Jerry Stralko <gerb.stralko@gmail.com>
> ---
> 
> 
> diff --git a/drivers/scsi/sr_ioctl.c b/drivers/scsi/sr_ioctl.c
> index ae87d08..c37fb1c 100644
> --- a/drivers/scsi/sr_ioctl.c
> +++ b/drivers/scsi/sr_ioctl.c
> @@ -238,8 +238,6 @@ int sr_do_ioctl(Scsi_CD *cd, struct packet_command *cgc)
>   					break;
>   				}
>   			}
> -			if (!cgc->quiet)
> -				printk(KERN_INFO "%s: CDROM not ready.  Make sure there is a disc in the drive.\n", cd->cdi.name);
>   #ifdef DEBUG
>   			scsi_print_sense_hdr("sr", &sshdr);
>   #endif

The answer may be that you need to enable cgc->quiet.  If that is
user-enableable - I can't work out how from a quick grep.

Re: [PATCH] scsi: remove CDROM not ready printk

[James Bottomley]

On Thu, 2008-05-29 at 00:41 -0700, Andrew Morton wrote:
> On Tue, 27 May 2008 15:02:43 -0400 (EDT) Jerry Stralko <gerb.stralko@gmail.com> wrote:
> 
> > Hello,
> 
> let's cc a scsi list which exists ;)
> 
> >    This printk was spamming my dmesg and /var/log/message.  Is there a 
> > reason we have this printk?  Can we simply remove it?
> > 
> > 
> > Signed-off-by: Jerry Stralko <gerb.stralko@gmail.com>
> > ---
> > 
> > 
> > diff --git a/drivers/scsi/sr_ioctl.c b/drivers/scsi/sr_ioctl.c
> > index ae87d08..c37fb1c 100644
> > --- a/drivers/scsi/sr_ioctl.c
> > +++ b/drivers/scsi/sr_ioctl.c
> > @@ -238,8 +238,6 @@ int sr_do_ioctl(Scsi_CD *cd, struct packet_command *cgc)
> >   					break;
> >   				}
> >   			}
> > -			if (!cgc->quiet)
> > -				printk(KERN_INFO "%s: CDROM not ready.  Make sure there is a disc in the drive.\n", cd->cdi.name);
> >   #ifdef DEBUG
> >   			scsi_print_sense_hdr("sr", &sshdr);
> >   #endif
> 
> The answer may be that you need to enable cgc->quiet.  If that is
> user-enableable - I can't work out how from a quick grep.

That's correct.  cgc is the data from the packet command that was sent
in from the ioctl.  The user application is actually the one that
decides whether you get to see the message or not.  Modern apps like hal
set it because they take care of all state updates and notifications
themselves.  Which is the application that's doing this (because it
probably needs updating).

James

Re: [PATCH] scsi: remove CDROM not ready printk

[Gerb Stralko]

On Thu, May 29, 2008 at 10:05 AM, James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
> On Thu, 2008-05-29 at 00:41 -0700, Andrew Morton wrote:
>> On Tue, 27 May 2008 15:02:43 -0400 (EDT) Jerry Stralko <gerb.stralko@gmail.com> wrote:
>>
>> > Hello,
>>
>> let's cc a scsi list which exists ;)
>>
>> >    This printk was spamming my dmesg and /var/log/message.  Is there a
>> > reason we have this printk?  Can we simply remove it?
>> >
>> >
>> > Signed-off-by: Jerry Stralko <gerb.stralko@gmail.com>
>> > ---
>> >
>> >
>> > diff --git a/drivers/scsi/sr_ioctl.c b/drivers/scsi/sr_ioctl.c
>> > index ae87d08..c37fb1c 100644
>> > --- a/drivers/scsi/sr_ioctl.c
>> > +++ b/drivers/scsi/sr_ioctl.c
>> > @@ -238,8 +238,6 @@ int sr_do_ioctl(Scsi_CD *cd, struct packet_command *cgc)
>> >                                     break;
>> >                             }
>> >                     }
>> > -                   if (!cgc->quiet)
>> > -                           printk(KERN_INFO "%s: CDROM not ready.  Make sure there is a disc in the drive.\n", cd->cdi.name);
>> >   #ifdef DEBUG
>> >                     scsi_print_sense_hdr("sr", &sshdr);
>> >   #endif
>>
>> The answer may be that you need to enable cgc->quiet.  If that is
>> user-enableable - I can't work out how from a quick grep.
>
> That's correct.  cgc is the data from the packet command that was sent
> in from the ioctl.  The user application is actually the one that
> decides whether you get to see the message or not.  Modern apps like hal
> set it because they take care of all state updates and notifications
> themselves.  Which is the application that's doing this (because it
> probably needs updating).
>
> James

So is that printk really needed?  Should the kernel even need to print
a message like that, esp. if user-space is handling state updates and
notifications.  Or do i need to configure hald to be quietier?   FWIW
I'm using fedora core 9 and hald version:
-bash-3.2$ /usr/sbin/hald --version
HAL package version: 0.5.11

thanks,

Jerry

Re: [PATCH] scsi: remove CDROM not ready printk

[Matthew Wilcox]

On Thu, May 29, 2008 at 10:35:08AM -0400, Gerb Stralko wrote:
> On Thu, May 29, 2008 at 10:05 AM, James Bottomley
> <James.Bottomley@hansenpartnership.com> wrote:
> > That's correct.  cgc is the data from the packet command that was sent
> > in from the ioctl.  The user application is actually the one that
> > decides whether you get to see the message or not.  Modern apps like hal
> > set it because they take care of all state updates and notifications
> > themselves.  Which is the application that's doing this (because it
> > probably needs updating).
> 
> So is that printk really needed?  Should the kernel even need to print
> a message like that, esp. if user-space is handling state updates and
> notifications.  Or do i need to configure hald to be quietier?   FWIW
> I'm using fedora core 9 and hald version:
> -bash-3.2$ /usr/sbin/hald --version
> HAL package version: 0.5.11

I think you misunderstood.  James is saying that an app like HAL will
set the ->quiet flag.  You must be using some other application that
isn't setting the ->quiet flag.  Any idea what this other application
is?

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."

Re: [PATCH] scsi: remove CDROM not ready printk

[James Bottomley]

On Thu, 2008-05-29 at 10:35 -0400, Gerb Stralko wrote:
> So is that printk really needed?  Should the kernel even need to print
> a message like that, esp. if user-space is handling state updates and
> notifications.  Or do i need to configure hald to be quietier?   FWIW
> I'm using fedora core 9 and hald version:
> -bash-3.2$ /usr/sbin/hald --version
> HAL package version: 0.5.11

Actually, I misspoke; it's set internally via sets of ioctls.  However,
it's designed only to show on ioctls that do medium requiring things.

I'm also using FC9 and I see no such messages (and for noisy cgc,
they're printed out for every action on the CD).  You have some
application that's poking the CD wrongly (probably not hal), so you
really need to find out what it is.  It may be pointing to some other
error in the kernel that needs fixing, but simply removing the message
is covering up the issue.

To help track the application, you could update the printk to print out
current->comm and current->pid.  That would tell you who is responsible

James

Re: [PATCH] scsi: remove CDROM not ready printk

[Gerb Stralko]

On Thu, May 29, 2008 at 10:55 AM, James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
> On Thu, 2008-05-29 at 10:35 -0400, Gerb Stralko wrote:
>> So is that printk really needed?  Should the kernel even need to print
>> a message like that, esp. if user-space is handling state updates and
>> notifications.  Or do i need to configure hald to be quietier?   FWIW
>> I'm using fedora core 9 and hald version:
>> -bash-3.2$ /usr/sbin/hald --version
>> HAL package version: 0.5.11
>
> Actually, I misspoke; it's set internally via sets of ioctls.  However,
> it's designed only to show on ioctls that do medium requiring things.
>
> I'm also using FC9 and I see no such messages (and for noisy cgc,
> they're printed out for every action on the CD).  You have some
> application that's poking the CD wrongly (probably not hal), so you
> really need to find out what it is.  It may be pointing to some other
> error in the kernel that needs fixing, but simply removing the message
> is covering up the issue.
>
> To help track the application, you could update the printk to print out
> current->comm and current->pid.  That would tell you who is responsible
>
> James
>
>
>
Okay, i'll do some investigation and see if I can track this down,
I'll let you know.   Thanks for the info.

Are you using GNOME or KDE?  I'm currently using KDE.

thanks

Re: [PATCH] scsi: remove CDROM not ready printk

[James Bottomley]

On Thu, 2008-05-29 at 11:04 -0400, Gerb Stralko wrote:
> On Thu, May 29, 2008 at 10:55 AM, James Bottomley
> <James.Bottomley@hansenpartnership.com> wrote:
> > On Thu, 2008-05-29 at 10:35 -0400, Gerb Stralko wrote:
> >> So is that printk really needed?  Should the kernel even need to print
> >> a message like that, esp. if user-space is handling state updates and
> >> notifications.  Or do i need to configure hald to be quietier?   FWIW
> >> I'm using fedora core 9 and hald version:
> >> -bash-3.2$ /usr/sbin/hald --version
> >> HAL package version: 0.5.11
> >
> > Actually, I misspoke; it's set internally via sets of ioctls.  However,
> > it's designed only to show on ioctls that do medium requiring things.
> >
> > I'm also using FC9 and I see no such messages (and for noisy cgc,
> > they're printed out for every action on the CD).  You have some
> > application that's poking the CD wrongly (probably not hal), so you
> > really need to find out what it is.  It may be pointing to some other
> > error in the kernel that needs fixing, but simply removing the message
> > is covering up the issue.
> >
> > To help track the application, you could update the printk to print out
> > current->comm and current->pid.  That would tell you who is responsible
> >
> > James
> >
> >
> >
> Okay, i'll do some investigation and see if I can track this down,
> I'll let you know.   Thanks for the info.
> 
> Are you using GNOME or KDE?  I'm currently using KDE.

I'm using gnome ... there are intermittent reports of kde problems
mainly with k3b though (the burner program), but if you put the printk
in it should identify the application.

James

Re: [PATCH] scsi: remove CDROM not ready printk

[Gerb Stralko]

On Thu, May 29, 2008 at 11:22 AM, James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
> On Thu, 2008-05-29 at 11:04 -0400, Gerb Stralko wrote:
>> On Thu, May 29, 2008 at 10:55 AM, James Bottomley
>> <James.Bottomley@hansenpartnership.com> wrote:
>> > On Thu, 2008-05-29 at 10:35 -0400, Gerb Stralko wrote:
>> >> So is that printk really needed?  Should the kernel even need to print
>> >> a message like that, esp. if user-space is handling state updates and
>> >> notifications.  Or do i need to configure hald to be quietier?   FWIW
>> >> I'm using fedora core 9 and hald version:
>> >> -bash-3.2$ /usr/sbin/hald --version
>> >> HAL package version: 0.5.11
>> >
>> > Actually, I misspoke; it's set internally via sets of ioctls.  However,
>> > it's designed only to show on ioctls that do medium requiring things.
>> >
>> > I'm also using FC9 and I see no such messages (and for noisy cgc,
>> > they're printed out for every action on the CD).  You have some
>> > application that's poking the CD wrongly (probably not hal), so you
>> > really need to find out what it is.  It may be pointing to some other
>> > error in the kernel that needs fixing, but simply removing the message
>> > is covering up the issue.
>> >
>> > To help track the application, you could update the printk to print out
>> > current->comm and current->pid.  That would tell you who is responsible
>> >
>> > James
>> >
>> >
>> >
>> Okay, i'll do some investigation and see if I can track this down,
>> I'll let you know.   Thanks for the info.
>>
>> Are you using GNOME or KDE?  I'm currently using KDE.
>
> I'm using gnome ... there are intermittent reports of kde problems
> mainly with k3b though (the burner program), but if you put the printk
> in it should identify the application.
>
> James
>
>
>
I download the fedora updates and I rebooted and the messages went
away from my dmesg.   I left my computer on for a few hours and no
message. So I guess I problem was fixed in one of the updates.

Thanks for you help.  Sorry for the noise.

thanks again,

Jerry