[PATCH] kill old EH constants

[PATCH] kill old EH constants

[Christoph Hellwig]

Fix up two drivers that incorrectly were using the old return values for
their new-style EH methods and kill off scsi_obsolete.h that defined the
constants.  The initio driver has all these constansts defined locally
and uses them internally, I'll fix that up some time later.


--- 1.95/drivers/scsi/scsi.h	2005-03-29 19:28:04 +02:00
+++ edited/drivers/scsi/scsi.h	2005-04-04 16:09:48 +02:00
@@ -60,12 +60,6 @@
 #define scsi_to_pci_dma_dir(scsi_dir)	((int)(scsi_dir))
 #define scsi_to_sbus_dma_dir(scsi_dir)	((int)(scsi_dir))
 
-/*
- * This is the crap from the old error handling code.  We have it in a special
- * place so that we can more easily delete it later on.
- */
-#include "scsi_obsolete.h"
-
 /* obsolete typedef junk. */
 #include "scsi_typedefs.h"
 
--- 1.1/drivers/scsi/scsi_obsolete.h	2002-02-05 18:40:08 +01:00
+++ edited/drivers/scsi/scsi_obsolete.h	2005-04-04 16:09:33 +02:00
@@ -1,106 +0,0 @@
-/*
- *  scsi_obsolete.h Copyright (C) 1997 Eric Youngdale
- *
- */
-
-#ifndef _SCSI_OBSOLETE_H
-#define _SCSI_OBSOLETE_H
-
-/*
- * These are the return codes for the abort and reset functions.  The mid-level
- * code uses these to decide what to do next.  Each of the low level abort
- * and reset functions must correctly indicate what it has done.
- * The descriptions are written from the point of view of the mid-level code,
- * so that the return code is telling the mid-level drivers exactly what
- * the low level driver has already done, and what remains to be done.
- */
-
-/* We did not do anything.  
- * Wait some more for this command to complete, and if this does not work, 
- * try something more serious. */
-#define SCSI_ABORT_SNOOZE 0
-
-/* This means that we were able to abort the command.  We have already
- * called the mid-level done function, and do not expect an interrupt that 
- * will lead to another call to the mid-level done function for this command */
-#define SCSI_ABORT_SUCCESS 1
-
-/* We called for an abort of this command, and we should get an interrupt 
- * when this succeeds.  Thus we should not restore the timer for this
- * command in the mid-level abort function. */
-#define SCSI_ABORT_PENDING 2
-
-/* Unable to abort - command is currently on the bus.  Grin and bear it. */
-#define SCSI_ABORT_BUSY 3
-
-/* The command is not active in the low level code. Command probably
- * finished. */
-#define SCSI_ABORT_NOT_RUNNING 4
-
-/* Something went wrong.  The low level driver will indicate the correct
- * error condition when it calls scsi_done, so the mid-level abort function
- * can simply wait until this comes through */
-#define SCSI_ABORT_ERROR 5
-
-/* We do not know how to reset the bus, or we do not want to.  Bummer.
- * Anyway, just wait a little more for the command in question, and hope that
- * it eventually finishes.  If it never finishes, the SCSI device could
- * hang, so use this with caution. */
-#define SCSI_RESET_SNOOZE 0
-
-/* We do not know how to reset the bus, or we do not want to.  Bummer.
- * We have given up on this ever completing.  The mid-level code will
- * request sense information to decide how to proceed from here. */
-#define SCSI_RESET_PUNT 1
-
-/* This means that we were able to reset the bus.  We have restarted all of
- * the commands that should be restarted, and we should be able to continue
- * on normally from here.  We do not expect any interrupts that will return
- * DID_RESET to any of the other commands in the host_queue, and the mid-level
- * code does not need to do anything special to keep the commands alive. 
- * If a hard reset was performed then all outstanding commands on the
- * bus have been restarted. */
-#define SCSI_RESET_SUCCESS 2
-
-/* We called for a reset of this bus, and we should get an interrupt 
- * when this succeeds.  Each command should get its own status
- * passed up to scsi_done, but this has not happened yet. 
- * If a hard reset was performed, then we expect an interrupt
- * for *each* of the outstanding commands that will have the
- * effect of restarting the commands.
- */
-#define SCSI_RESET_PENDING 3
-
-/* We did a reset, but do not expect an interrupt to signal DID_RESET.
- * This tells the upper level code to request the sense info, and this
- * should keep the command alive. */
-#define SCSI_RESET_WAKEUP 4
-
-/* The command is not active in the low level code. Command probably
-   finished. */
-#define SCSI_RESET_NOT_RUNNING 5
-
-/* Something went wrong, and we do not know how to fix it. */
-#define SCSI_RESET_ERROR 6
-
-#define SCSI_RESET_SYNCHRONOUS		0x01
-#define SCSI_RESET_ASYNCHRONOUS		0x02
-#define SCSI_RESET_SUGGEST_BUS_RESET	0x04
-#define SCSI_RESET_SUGGEST_HOST_RESET	0x08
-/*
- * This is a bitmask that is ored with one of the above codes.
- * It tells the mid-level code that we did a hard reset.
- */
-#define SCSI_RESET_BUS_RESET 0x100
-/*
- * This is a bitmask that is ored with one of the above codes.
- * It tells the mid-level code that we did a host adapter reset.
- */
-#define SCSI_RESET_HOST_RESET 0x200
-/*
- * Used to mask off bits and to obtain the basic action that was
- * performed.  
- */
-#define SCSI_RESET_ACTION   0xff
-
-#endif				/* SCSI_OBSOLETE_H */
===== drivers/scsi/ultrastor.c 1.25 vs edited =====
--- 1.25/drivers/scsi/ultrastor.c	2005-03-11 21:32:31 +01:00
+++ edited/drivers/scsi/ultrastor.c	2005-04-04 16:09:02 +02:00
@@ -945,7 +945,7 @@
 	       config.mscp[mscp_index].SCint, SCpnt);
 #endif
     if (config.mscp[mscp_index].SCint == 0)
-	return SCSI_ABORT_NOT_RUNNING;
+	return FAILURE;
 
     if (config.mscp[mscp_index].SCint != SCpnt) panic("Bad abort");
     config.mscp[mscp_index].SCint = NULL;
@@ -1020,7 +1020,7 @@
 #endif
 
     spin_unlock_irqrestore(host->host_lock, flags);
-    return SCSI_RESET_SUCCESS;
+    return SUCCESS;
 
 }
 
--- 1.44/drivers/usb/image/microtek.c	2005-03-05 16:13:09 +01:00
+++ edited/drivers/usb/image/microtek.c	2005-04-04 16:09:02 +02:00
@@ -335,7 +335,7 @@
 
 	mts_urb_abort(desc);
 
-	return SCSI_ABORT_PENDING;
+	return FAILURE;
 }
 
 static int mts_scsi_host_reset (Scsi_Cmnd *srb)

Re: [PATCH] kill old EH constants

[Geert Uytterhoeven]

On Wed, 20 Apr 2005, Linux Kernel Mailing List wrote:
> tree 932c6f9689fd08a7a9d689cfbec8682ccde8175d
> parent 84011ae88da62a20b3ae7b48e2ae3b1ef0fc810a
> author <hch@lst.de> Mon, 11 Apr 2005 08:19:25 -0500
> committer James Bottomley <jejb@titanic> Sun, 17 Apr 2005 06:14:52 -0500
> 
> [PATCH] kill old EH constants
> 
> Fix up two drivers that incorrectly were using the old return values for
> their new-style EH methods and kill off scsi_obsolete.h that defined the
> constants.  The initio driver has all these constansts defined locally
> and uses them internally, I'll fix that up some time later.
> 
> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

sun3_NCR5380.c still uses the following:

  - SCSI_ABORT_SUCCESS
  - SCSI_ABORT_ERROR
  - SCSI_ABORT_SNOOZE
  - SCSI_ABORT_BUSY
  - SCSI_ABORT_NOT_RUNNING
  - SCSI_RESET_SUCCESS
  - SCSI_RESET_BUS_RESET

causing the driver to fail to build in 2.6.12-rc3. What should I replace them
by?

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

Re: [PATCH] kill old EH constants

[Christoph Hellwig]

On Thu, Apr 21, 2005 at 11:58:12AM +0200, Geert Uytterhoeven wrote:
> sun3_NCR5380.c still uses the following:
> 
>   - SCSI_ABORT_SUCCESS
>   - SCSI_ABORT_ERROR
>   - SCSI_ABORT_SNOOZE
>   - SCSI_ABORT_BUSY
>   - SCSI_ABORT_NOT_RUNNING
>   - SCSI_RESET_SUCCESS
>   - SCSI_RESET_BUS_RESET
> 
> causing the driver to fail to build in 2.6.12-rc3. What should I replace them
> by?

You must replace NCR5380_abort and NCR5380_bus_reset with real new-style
EH routines.  I'd suggest copying them from NCR5380.c or even better
scrapping sun3_NCR5380.c in favour of that one completely.

Re: [PATCH] kill old EH constants

[Sam Creasey]

On Thu, 21 Apr 2005, Christoph Hellwig wrote:

> On Thu, Apr 21, 2005 at 11:58:12AM +0200, Geert Uytterhoeven wrote:
> > sun3_NCR5380.c still uses the following:
> >
> >   - SCSI_ABORT_SUCCESS
> >   - SCSI_ABORT_ERROR
> >   - SCSI_ABORT_SNOOZE
> >   - SCSI_ABORT_BUSY
> >   - SCSI_ABORT_NOT_RUNNING
> >   - SCSI_RESET_SUCCESS
> >   - SCSI_RESET_BUS_RESET
> >
> > causing the driver to fail to build in 2.6.12-rc3. What should I replace them
> > by?
>
> You must replace NCR5380_abort and NCR5380_bus_reset with real new-style
> EH routines.  I'd suggest copying them from NCR5380.c or even better
> scrapping sun3_NCR5380.c in favour of that one completely.

Trust me, there's reasons we don't use NCR5380.c.... (primarily, this has
to due with the order of operations necessary to keep the DMA controller
happpy, which required changing the flow of several functions
(admittedly, I haven't looked into this again since the "new-style"
driver change)).

That being said, it seems the first option needs done.  I suppose I'll
add fixing those two routines to my long list of backlogged sun3 stuff
which needs doing.

-- Sam