[PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Douglas Gilbert]

[-- Attachment #1: Type: text/plain, Size: 445 bytes --]

Here is a simple patch to replace the single occurrence
of "u8" with "u_char" in include/scsi/scsi.h .

While working on an application (to test sgbind) that
needs to include the kernel rather than the glibc
headers, it is a nuisance to have to add:
    typedef unsigned char u8;
before including /usr/src/linux/include/scsi/scsi.h

Perhaps we should be moving toward the official C99
typedef: uint8_t which is defined in stdint.h

Doug Gilbert

[-- Attachment #2: scsi_h_2569.diff --]
[-- Type: text/plain, Size: 274 bytes --]

--- linux/include/scsi/scsi.h	2003-02-11 08:39:21.000000000 +1000
+++ linux/include/scsi/scsi.h2569fix	2003-05-14 12:41:48.000000000 +1000
@@ -201,7 +201,7 @@
  * ScsiLun: 8 byte LUN.
  */
 typedef struct scsi_lun {
-	u8 scsi_lun[8];
+	u_char scsi_lun[8];
 } ScsiLun;
 
 /*

Re: [PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Christoph Hellwig]

On Wed, May 14, 2003 at 04:39:43PM +1000, Douglas Gilbert wrote:
> Here is a simple patch to replace the single occurrence
> of "u8" with "u_char" in include/scsi/scsi.h .

That's wrong.  No userland should include kernwel headers.
(and u_char is a non-standard type, too).

Re: [PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Jeff Garzik]

Christoph Hellwig wrote:
> On Wed, May 14, 2003 at 04:39:43PM +1000, Douglas Gilbert wrote:
> 
>>Here is a simple patch to replace the single occurrence
>>of "u8" with "u_char" in include/scsi/scsi.h .
> 
> 
> That's wrong.  No userland should include kernwel headers.
> (and u_char is a non-standard type, too).


Agreed.  And like Douglas guessed, uint8_t is preferred over u_char :)

	Jeff

Re: [PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Douglas Gilbert]

Christoph Hellwig wrote:
> On Wed, May 14, 2003 at 04:39:43PM +1000, Douglas Gilbert wrote:
> 
>>Here is a simple patch to replace the single occurrence
>>of "u8" with "u_char" in include/scsi/scsi.h .
> 
> 
> That's wrong.  No userland should include kernwel headers.
> (and u_char is a non-standard type, too).

Christoph,
That's naive. In some case glibc's headers and libraries
(e.g. mknod) are an impediment to development.

If I add one define to ~linux/include/scsi/sg.h just
how long do you think it takes for it to appear in
/usr/include/scsi/sg.h?
It can take over a year in my experience.

Put another way, the sg.h header that applications use
should match the sg driver in the kernel not the glibc
version. The glibc maintainers add nothing to the
interface (apart from inertia and some bugs once in a
while).

Applications that include sg.h typically include scsi.h
as well. Only when I change sg.h does this problem arise.

It is one thing to state a policy (e.g. "no user program
should include kernel headers") it is another thing to
gratuitously break those few programs that have good
reason not to follow that policy.

Doug Gilbert

Re: [PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Christoph Hellwig]

On Wed, May 14, 2003 at 11:10:27PM +1000, Douglas Gilbert wrote:
> Christoph,
> That's naive. In some case glibc's headers and libraries
> (e.g. mknod) are an impediment to development.
> 
> If I add one define to ~linux/include/scsi/sg.h just
> how long do you think it takes for it to appear in
> /usr/include/scsi/sg.h?

As long as it takes you to send the same diff to Ullrich.

> It can take over a year in my experience.
> 
> Put another way, the sg.h header that applications use
> should match the sg driver in the kernel

What kernel?  I run everything from 2.2.2x + local hacks to latest
2.5 on one box.

> It is one thing to state a policy (e.g. "no user program
> should include kernel headers") it is another thing to
> gratuitously break those few programs that have good
> reason not to follow that policy.

So get declarations for u8 into your program.  It's a kernel
header for a reason and if you ignore the advise _you_ have
to work around the problems arising from it not the kernel.

Re: [PATCH] include/scsi/scsi.h replace u8 in 2.5.69

[Andries Brouwer]

On Wed, May 14, 2003 at 11:10:27PM +1000, Douglas Gilbert wrote:

> >That's wrong.  No userland should include kernwel headers.
> >(and u_char is a non-standard type, too).
> 
> Christoph,
> That's naive. In some case glibc's headers and libraries
> (e.g. mknod) are an impediment to development.
> 
> If I add one define to ~linux/include/scsi/sg.h just
> how long do you think it takes for it to appear in
> /usr/include/scsi/sg.h?
> It can take over a year in my experience.

Consider: the same binaries are used on a wide range of kernels.
Consider: the same sources are compiled on machines with a
wide range of installed kernel sources.

It is undesirable that recompilation is needed when
a kernel is upgraded (or downgraded).

Thus, the robust setup uses its own copy of the headers,
or perhaps uses something like

#include <header.h>
#ifndef MY_CONSTANT
#define MY_CONSTANT 12345
#endif

to make sure that things compile with old and new *libc, with
old and new kernels.

Next, this robust setup must make sure that things not only
compile but also run, i.e., must be prepared to deal with EINVAL
or ENOSYS etc in case it is run on a kernel that does not yet
know about this new stuff.

Andries


[There is an different case for a private copy of the headers:
the kernel headers change all the time, and sometimes such
changes cause obscure and difficult-to-debug errors in utilities.
I still recall the case, many years ago, of a struct field that
was made unsigned, causing a change in semantics for a comparison.
User space software is easier to debug when the corresponding
binary does not depend in obscure ways on the kernel source
installed when it was compiled.]