[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Trivial Russell]

[ Removes the unused warning, and since qla1280_do_dpc is only called when
  version < 2.1.95 anyway, makes the whole thing conditional ]

From:  Eric Sandeen <sandeen@sgi.com>


--- trivial-2.4.20-pre1/drivers/scsi/qla1280.c.orig	2002-08-07 10:57:25.000000000 +1000
+++ trivial-2.4.20-pre1/drivers/scsi/qla1280.c	2002-08-07 10:57:25.000000000 +1000
@@ -312,7 +312,9 @@
 STATIC int qla1280_return_status( sts_entry_t *sts, Scsi_Cmnd       *cp);
 STATIC void qla1280_removeq(scsi_lu_t *q, srb_t *sp);
 STATIC void qla1280_mem_free(scsi_qla_host_t *ha);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p);
+#endif
 #ifdef  QLA1280_UNUSED 
 static void qla1280_set_flags(char * s);
 #endif
@@ -1611,17 +1613,12 @@
  * "host->can_queue". This can cause a panic if we were in our interrupt
  * code .
  **************************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p)
 {
     scsi_qla_host_t *ha = (scsi_qla_host_t *) p;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    unsigned long cpu_flags = 0;
-#endif
 
     COMTRACE('p')  
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_lock_irqsave(&io_request_lock, cpu_flags);
-#endif
     if (ha->flags.isp_abort_needed)
         qla1280_abort_isp(ha);
 
@@ -1631,10 +1628,8 @@
     if (ha->done_q_first)
         qla1280_done(ha, (srb_t **)&ha->done_q_first, (srb_t **)&ha->done_q_last);
     ha->flags.dpc_sched = FALSE;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_unlock_irqrestore(&io_request_lock, cpu_flags);
-#endif
 }
+#endif /* LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95) */
 
 /**************************************************************************
  *   qla1280_device_queue_depth
-- 
  Don't blame me: the Monkey is driving

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Steve Ralston]

On 08/07/2002 06:37 AM, Alan Cox wrote:

>General policy - rip out version ifdefs unless the driver has a
>maintainer who actively objects ?
>

Hi Alan,

When I worked for LSI, and was actively maintaining 
drivers/message/fusion/...
code, we had a single set of driver source code that compiled and ran
across lots of linux kernels (2.2.5 thru 2.4.19).  Most of the 
version-specific
#ifdef's were isolated to one header file (linux_compat.h).  We then 
also had
a set of patch files (each pretty small) that could be applied against 
any explicit
kernel (in the supported range).  A customer could then (in theory:-) 
download
one set of driver source (including the patch file sets), untar/overlay it,
apply a single patch file, and then compile+run the drivers on any of 
those kernels.

So just curious, are you saying that you'd rather have driver maintainers
maintain a separate set of driver source for each explicit kernel (with no
version-specific #ifdef's) vs. the way we did it above?

Thanks,
-SteveR

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Alan Cox]

On Wed, 2002-08-07 at 04:59, Rusty Trivial Russell wrote:
> [ Removes the unused warning, and since qla1280_do_dpc is only called when
>   version < 2.1.95 anyway, makes the whole thing conditional ]
> 

General policy - rip out version ifdefs unless the driver has a
maintainer who actively objects ?

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Alan Cox]

On Wed, 2002-08-07 at 12:12, Steve Ralston wrote:
> On 08/07/2002 06:37 AM, Alan Cox wrote:
> 
> >General policy - rip out version ifdefs unless the driver has a
> >maintainer who actively objects ?
> >
> 
> So just curious, are you saying that you'd rather have driver maintainers
> maintain a separate set of driver source for each explicit kernel (with no
> version-specific #ifdef's) vs. the way we did it above?

If someone is maintaining them all then there are cases it makes sense
to leave the ifdefs in. Thats why I said about maintainers who actively
object. I suspect Deanna would also want the MPT fusion ifdefs left in
for example

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Russell]

In message <1028720225.18156.246.camel@irongate.swansea.linux.org.uk> you write
:
> On Wed, 2002-08-07 at 04:59, Rusty Trivial Russell wrote:
> > [ Removes the unused warning, and since qla1280_do_dpc is only called when
> >   version < 2.1.95 anyway, makes the whole thing conditional ]
> > 
> 
> General policy - rip out version ifdefs unless the driver has a
> maintainer who actively objects ?

Hm, I wouldn't accept a patch to Trivial which just ripped them out.
If it was part of a larger cleanup, then sure.  In this case it
doesn't change code, it just simplifies it as well as the easiest way
of fixing the warning.

Rusty.
--
  Anyone who quotes me in their sig is an idiot. -- Rusty Russell.

[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Trivial Russell]

[ Hasn't been touched in years, but this cleans up an "unused" warning
  the simplest way (make the whole function dependent on kernel
  version, since its only call is ]


From:  Eric Sandeen <sandeen@sgi.com>


--- trivial-2.4.20-pre4/drivers/scsi/qla1280.c.orig	2002-08-20 18:00:34.000000000 +1000
+++ trivial-2.4.20-pre4/drivers/scsi/qla1280.c	2002-08-20 18:00:34.000000000 +1000
@@ -312,7 +312,9 @@
 STATIC int qla1280_return_status( sts_entry_t *sts, Scsi_Cmnd       *cp);
 STATIC void qla1280_removeq(scsi_lu_t *q, srb_t *sp);
 STATIC void qla1280_mem_free(scsi_qla_host_t *ha);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p);
+#endif
 #ifdef  QLA1280_UNUSED 
 static void qla1280_set_flags(char * s);
 #endif
@@ -1611,17 +1613,12 @@
  * "host->can_queue". This can cause a panic if we were in our interrupt
  * code .
  **************************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p)
 {
     scsi_qla_host_t *ha = (scsi_qla_host_t *) p;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    unsigned long cpu_flags = 0;
-#endif
 
     COMTRACE('p')  
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_lock_irqsave(&io_request_lock, cpu_flags);
-#endif
     if (ha->flags.isp_abort_needed)
         qla1280_abort_isp(ha);
 
@@ -1631,10 +1628,8 @@
     if (ha->done_q_first)
         qla1280_done(ha, (srb_t **)&ha->done_q_first, (srb_t **)&ha->done_q_last);
     ha->flags.dpc_sched = FALSE;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_unlock_irqrestore(&io_request_lock, cpu_flags);
-#endif
 }
+#endif /* LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95) */
 
 /**************************************************************************
  *   qla1280_device_queue_depth
-- 
  Don't blame me: the Monkey is driving
  File: Eric Sandeen <sandeen@sgi.com>: warning cleanup for drivers_scsi_qla1280.c

[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Trivial Russell]

[ Indeed, the only call to qla1280_do_dpc is already inside 
  #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,1,95), so this changes nothing ]

From:  Eric Sandeen <sandeen@sgi.com>


--- trivial-2.4.21-pre1/drivers/scsi/qla1280.c.orig	2002-12-16 17:22:08.000000000 +1100
+++ trivial-2.4.21-pre1/drivers/scsi/qla1280.c	2002-12-16 17:22:08.000000000 +1100
@@ -312,7 +312,9 @@
 STATIC int qla1280_return_status( sts_entry_t *sts, Scsi_Cmnd       *cp);
 STATIC void qla1280_removeq(scsi_lu_t *q, srb_t *sp);
 STATIC void qla1280_mem_free(scsi_qla_host_t *ha);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p);
+#endif
 #ifdef  QLA1280_UNUSED 
 static void qla1280_set_flags(char * s);
 #endif
@@ -1611,17 +1613,12 @@
  * "host->can_queue". This can cause a panic if we were in our interrupt
  * code .
  **************************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p)
 {
     scsi_qla_host_t *ha = (scsi_qla_host_t *) p;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    unsigned long cpu_flags = 0;
-#endif
 
     COMTRACE('p')  
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_lock_irqsave(&io_request_lock, cpu_flags);
-#endif
     if (ha->flags.isp_abort_needed)
         qla1280_abort_isp(ha);
 
@@ -1631,10 +1628,8 @@
     if (ha->done_q_first)
         qla1280_done(ha, (srb_t **)&ha->done_q_first, (srb_t **)&ha->done_q_last);
     ha->flags.dpc_sched = FALSE;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_unlock_irqrestore(&io_request_lock, cpu_flags);
-#endif
 }
+#endif /* LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95) */
 
 /**************************************************************************
  *   qla1280_device_queue_depth
-- 
  Don't blame me: the Monkey is driving
  File: Eric Sandeen <sandeen@sgi.com>: warning cleanup for drivers_scsi_qla1280.c

[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Trivial Russell]

From:  Eric Sandeen <sandeen@sgi.com>


--- trivial-2.4.21-pre4/drivers/scsi/qla1280.c.orig	2003-02-06 13:53:42.000000000 +1100
+++ trivial-2.4.21-pre4/drivers/scsi/qla1280.c	2003-02-06 13:53:42.000000000 +1100
@@ -312,7 +312,9 @@
 STATIC int qla1280_return_status( sts_entry_t *sts, Scsi_Cmnd       *cp);
 STATIC void qla1280_removeq(scsi_lu_t *q, srb_t *sp);
 STATIC void qla1280_mem_free(scsi_qla_host_t *ha);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p);
+#endif
 #ifdef  QLA1280_UNUSED 
 static void qla1280_set_flags(char * s);
 #endif
@@ -1611,17 +1613,12 @@
  * "host->can_queue". This can cause a panic if we were in our interrupt
  * code .
  **************************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p)
 {
     scsi_qla_host_t *ha = (scsi_qla_host_t *) p;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    unsigned long cpu_flags = 0;
-#endif
 
     COMTRACE('p')  
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_lock_irqsave(&io_request_lock, cpu_flags);
-#endif
     if (ha->flags.isp_abort_needed)
         qla1280_abort_isp(ha);
 
@@ -1631,10 +1628,8 @@
     if (ha->done_q_first)
         qla1280_done(ha, (srb_t **)&ha->done_q_first, (srb_t **)&ha->done_q_last);
     ha->flags.dpc_sched = FALSE;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_unlock_irqrestore(&io_request_lock, cpu_flags);
-#endif
 }
+#endif /* LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95) */
 
 /**************************************************************************
  *   qla1280_device_queue_depth
-- 
  Don't blame me: the Monkey is driving
  File: Eric Sandeen <sandeen@sgi.com>: warning cleanup for drivers_scsi_qla1280.c

[TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Trivial Russell]

[ This function is unused if kernel > 2.1.95, so put #ifdef around whole thing
  to avoid warning. --RR ]

From:  Eric Sandeen <sandeen@sgi.com>


--- trivial-2.4.22-pre1/drivers/scsi/qla1280.c.orig	2003-06-23 16:39:22.000000000 +1000
+++ trivial-2.4.22-pre1/drivers/scsi/qla1280.c	2003-06-23 16:39:22.000000000 +1000
@@ -312,7 +312,9 @@
 STATIC int qla1280_return_status( sts_entry_t *sts, Scsi_Cmnd       *cp);
 STATIC void qla1280_removeq(scsi_lu_t *q, srb_t *sp);
 STATIC void qla1280_mem_free(scsi_qla_host_t *ha);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p);
+#endif
 #ifdef  QLA1280_UNUSED 
 static void qla1280_set_flags(char * s);
 #endif
@@ -1611,17 +1613,12 @@
  * "host->can_queue". This can cause a panic if we were in our interrupt
  * code .
  **************************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95)
 static void qla1280_do_dpc(void *p)
 {
     scsi_qla_host_t *ha = (scsi_qla_host_t *) p;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    unsigned long cpu_flags = 0;
-#endif
 
     COMTRACE('p')  
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_lock_irqsave(&io_request_lock, cpu_flags);
-#endif
     if (ha->flags.isp_abort_needed)
         qla1280_abort_isp(ha);
 
@@ -1631,10 +1628,8 @@
     if (ha->done_q_first)
         qla1280_done(ha, (srb_t **)&ha->done_q_first, (srb_t **)&ha->done_q_last);
     ha->flags.dpc_sched = FALSE;
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,1,95)
-    spin_unlock_irqrestore(&io_request_lock, cpu_flags);
-#endif
 }
+#endif /* LINUX_VERSION_CODE < KERNEL_VERSION(2,1,95) */
 
 /**************************************************************************
  *   qla1280_device_queue_depth
-- 
  What is this? http://www.kernel.org/pub/linux/kernel/people/rusty/trivial/
  Don't blame me: the Monkey is driving
  File: Eric Sandeen <sandeen@sgi.com>: warning cleanup for drivers_scsi_qla1280.c

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Christoph Hellwig]

On Mon, Jun 23, 2003 at 04:47:25PM +1000, Rusty Trivial Russell wrote:
> [ This function is unused if kernel > 2.1.95, so put #ifdef around whole thing
>   to avoid warning. --RR ]

I'd say kill it completly.  qlogic has stopped maintaining the driver
and there's other places in the driver that won't work anymore with 2.0.

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Jes Sorensen]

>>>>> "Christoph" == Christoph Hellwig <hch@infradead.org> writes:

Christoph> On Mon, Jun 23, 2003 at 04:47:25PM +1000, Rusty Trivial
Christoph> Russell wrote:
>> [ This function is unused if kernel > 2.1.95, so put #ifdef around
>> whole thing to avoid warning. --RR ]

Christoph> I'd say kill it completly.  qlogic has stopped maintaining
Christoph> the driver and there's other places in the driver that
Christoph> won't work anymore with 2.0.

Hi Christoph and Rusty

The current qla1280 codebase found in 2.4 is utter rubbish. I have a
major cleanup of it almost ready in which I already killed all the
< 2.4.x support.

I hope to get this new version out shortly, so this patch is probably
obsolete.

Cheers,
Jes

Re: [TRIVIAL] warning cleanup for drivers_scsi_qla1280.c

[Rusty Russell]

In message <m38yrtnpmy.fsf@trained-monkey.org> you write:
> I hope to get this new version out shortly, so this patch is probably
> obsolete.

Thanks Jes, and Christoph!

I've closed the patch.
Rusty.
--
  Anyone who quotes me in their sig is an idiot. -- Rusty Russell.