Re: Nuke remnants of /dev/io

[Crystal Kolipe]

Ping?

On Thu, Feb 23, 2023 at 08:05:07PM -0800, Greg Steuck wrote:
> Thanks Crystal. If somebody wants to commit this, it is OK gnezdo@
> 
> Crystal Kolipe  writes:
> 
> > The iskmemdev function checks for minor number 14 in addition to 0 and 1 on
> > the following archs:
> >
> > amd64, arm64, i386, and riscv64
> >
> > Device 2, 14 was traditionally /dev/io, which we don't support and so 
> > opening
> > it will always return ENXIO from mmopen anyway.
> >
> > We only use iskmemdev in one place in the tree, to return EPERM when trying
> > to access /dev/kmem or /dev/mem when securelevel >= 1.
> >
> > This patch removes the check for minor(dev) == 14 on the four above 
> > mentioned
> > architectures.
> >
> > --- sys/arch/amd64/amd64/conf.c.distMon Feb 20 18:17:44 2023
> > +++ sys/arch/amd64/amd64/conf.c Mon Feb 20 18:29:28 2023
> > @@ -313,7 +313,7 @@
> >  iskmemdev(dev_t dev)
> >  {
> >  
> > -   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> > +   return (major(dev) == mem_no && minor(dev) < 2);
> >  }
> >  
> >  /*
> > --- sys/arch/arm64/arm64/conf.c.distMon Feb 20 18:18:20 2023
> > +++ sys/arch/arm64/arm64/conf.c Mon Feb 20 18:29:14 2023
> > @@ -255,7 +255,7 @@
> >  iskmemdev(dev_t dev)
> >  {
> >  
> > -   return (major(dev) == CMAJ_MM && (minor(dev) < 2 || minor(dev) == 14));
> > +   return (major(dev) == CMAJ_MM && minor(dev) < 2);
> >  }
> >  
> >  /*
> > --- sys/arch/i386/i386/conf.c.dist  Mon Feb 20 18:18:35 2023
> > +++ sys/arch/i386/i386/conf.c   Mon Feb 20 18:28:51 2023
> > @@ -309,7 +309,7 @@
> >  int
> >  iskmemdev(dev_t dev)
> >  {
> > -   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> > +   return (major(dev) == mem_no && minor(dev) < 2);
> >  }
> >  
> >  /*
> > --- sys/arch/riscv64/riscv64/conf.c.distMon Feb 20 18:18:48 2023
> > +++ sys/arch/riscv64/riscv64/conf.c Mon Feb 20 18:28:35 2023
> > @@ -253,7 +253,7 @@
> >  iskmemdev(dev_t dev)
> >  {
> >  
> > -   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> > +   return (major(dev) == mem_no && minor(dev) < 2);
> >  }
> >  
> >  /*
> 

Re: Nuke remnants of /dev/io

[Greg Steuck]

Thanks Crystal. If somebody wants to commit this, it is OK gnezdo@

Crystal Kolipe  writes:

> The iskmemdev function checks for minor number 14 in addition to 0 and 1 on
> the following archs:
>
> amd64, arm64, i386, and riscv64
>
> Device 2, 14 was traditionally /dev/io, which we don't support and so opening
> it will always return ENXIO from mmopen anyway.
>
> We only use iskmemdev in one place in the tree, to return EPERM when trying
> to access /dev/kmem or /dev/mem when securelevel >= 1.
>
> This patch removes the check for minor(dev) == 14 on the four above mentioned
> architectures.
>
> --- sys/arch/amd64/amd64/conf.c.dist  Mon Feb 20 18:17:44 2023
> +++ sys/arch/amd64/amd64/conf.c   Mon Feb 20 18:29:28 2023
> @@ -313,7 +313,7 @@
>  iskmemdev(dev_t dev)
>  {
>  
> - return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> + return (major(dev) == mem_no && minor(dev) < 2);
>  }
>  
>  /*
> --- sys/arch/arm64/arm64/conf.c.dist  Mon Feb 20 18:18:20 2023
> +++ sys/arch/arm64/arm64/conf.c   Mon Feb 20 18:29:14 2023
> @@ -255,7 +255,7 @@
>  iskmemdev(dev_t dev)
>  {
>  
> - return (major(dev) == CMAJ_MM && (minor(dev) < 2 || minor(dev) == 14));
> + return (major(dev) == CMAJ_MM && minor(dev) < 2);
>  }
>  
>  /*
> --- sys/arch/i386/i386/conf.c.distMon Feb 20 18:18:35 2023
> +++ sys/arch/i386/i386/conf.c Mon Feb 20 18:28:51 2023
> @@ -309,7 +309,7 @@
>  int
>  iskmemdev(dev_t dev)
>  {
> - return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> + return (major(dev) == mem_no && minor(dev) < 2);
>  }
>  
>  /*
> --- sys/arch/riscv64/riscv64/conf.c.dist  Mon Feb 20 18:18:48 2023
> +++ sys/arch/riscv64/riscv64/conf.c   Mon Feb 20 18:28:35 2023
> @@ -253,7 +253,7 @@
>  iskmemdev(dev_t dev)
>  {
>  
> - return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
> + return (major(dev) == mem_no && minor(dev) < 2);
>  }
>  
>  /*

Nuke remnants of /dev/io

[Crystal Kolipe]

The iskmemdev function checks for minor number 14 in addition to 0 and 1 on
the following archs:

amd64, arm64, i386, and riscv64

Device 2, 14 was traditionally /dev/io, which we don't support and so opening
it will always return ENXIO from mmopen anyway.

We only use iskmemdev in one place in the tree, to return EPERM when trying
to access /dev/kmem or /dev/mem when securelevel >= 1.

This patch removes the check for minor(dev) == 14 on the four above mentioned
architectures.

--- sys/arch/amd64/amd64/conf.c.distMon Feb 20 18:17:44 2023
+++ sys/arch/amd64/amd64/conf.c Mon Feb 20 18:29:28 2023
@@ -313,7 +313,7 @@
 iskmemdev(dev_t dev)
 {
 
-   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
+   return (major(dev) == mem_no && minor(dev) < 2);
 }
 
 /*
--- sys/arch/arm64/arm64/conf.c.distMon Feb 20 18:18:20 2023
+++ sys/arch/arm64/arm64/conf.c Mon Feb 20 18:29:14 2023
@@ -255,7 +255,7 @@
 iskmemdev(dev_t dev)
 {
 
-   return (major(dev) == CMAJ_MM && (minor(dev) < 2 || minor(dev) == 14));
+   return (major(dev) == CMAJ_MM && minor(dev) < 2);
 }
 
 /*
--- sys/arch/i386/i386/conf.c.dist  Mon Feb 20 18:18:35 2023
+++ sys/arch/i386/i386/conf.c   Mon Feb 20 18:28:51 2023
@@ -309,7 +309,7 @@
 int
 iskmemdev(dev_t dev)
 {
-   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
+   return (major(dev) == mem_no && minor(dev) < 2);
 }
 
 /*
--- sys/arch/riscv64/riscv64/conf.c.distMon Feb 20 18:18:48 2023
+++ sys/arch/riscv64/riscv64/conf.c Mon Feb 20 18:28:35 2023
@@ -253,7 +253,7 @@
 iskmemdev(dev_t dev)
 {
 
-   return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
+   return (major(dev) == mem_no && minor(dev) < 2);
 }
 
 /*