Re: 2.6.14 kernels and above copy_to_user stupidity with IRQ disabled check

[jmerkey]

On Fri, Jan 27, 2006 at 01:10:58PM -0700, jmerkey@xxxxxxxxxxxxxxxx wrote:


Also, allowing page faults in kernel to swap in user space memory 
seems dangerous.  Now I understand why I see the page fault handler
recurse in some cases with other code.  

W2K does the same thing for performance reasons, so I guess it doesn't really
matter.  I would assume there would be a safer place to fault in the memory, but
given people rolling their own ioctl requests, I can see where determining
this would be difficult.

Jeff

> 
> 
> OK.  Got it.  I guess I need to restructure.  And BTW, This was a code fragment
> only, the spinlock gets released when -EFAULT is called -- was just an example.
> 
> Jeff
> 
> On Fri, Jan 27, 2006 at 03:18:06PM -0500, linux-os (Dick Johnson) wrote:
> > 
> > On Fri, 27 Jan 2006, Jeff V. Merkey wrote:
> > 
> > >
> > > Is there a good reason someone set a disabled_irq() check on 2.6.14 and
> > > above for copy_to_user to barf out
> > > tons of bogus stack dump messages if the function is called from within
> > > a spinlock:
> > >
> > 
> > This is a joke, right????
> > 
> > > i.e.
> > >
> > > spin_lock_irqsave(&regen_lock, regen_flags);
> > >    v = regen_head;
> > >    while (v)
> > >    {
> > >       if (i >= count)
> > >          return -EFAULT;
> > 
> > ** BUG **  return with spin-lock held!
> > 
> > >
> > >
> > >       err = copy_to_user(&s[i++], v, sizeof(VIRTUAL_SETUP));
> > 
> > ** BUG ** copy to user with spinlock held!
> > 
> > >       if (err)
> > >          return err;
> > >
> > 
> > ** BUG ** Return with spin-lock held!
> > >
> > >       v = v->next;
> > >    }
> > >    spin_unlock_irqrestore(&regen_lock, regen_flags);
> > >
> > > is now busted and worked in kernels up to this point.  The error message
> > > is annoying but non-fatal.
> > >
> > > Jeff
> > 
> > It was NEVER supposed to work! The only reason it worked is because
> > your page(s) copied to, were not swapped out. If they were swapped
> > out, you are stuck, the page-fault won't occur.
> > 
> > Cheers,
> > Dick Johnson
> > Penguin : Linux version 2.6.13.4 on an i686 machine (5589.66 BogoMips).
> > Warning : 98.36% of all statistics are fiction.
> > .
> > 
> > ****************************************************************
> > The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@xxxxxxxxxxxx - and destroy all copies of this information, including any attachments, without reading or disclosing them.
> > 
> > Thank you.
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/