merge megaraid 2.20.x in kernel 2.6?

merge megaraid 2.20.x in kernel 2.6?

[Matt Domsch]

James, is there anything you're still waiting on before merging the
LSI megaraid 2.20.x driver into your scsi-misc-2.6 tree?  I thought
the agreement by you, Christoph, and Atul was to get it merged, and
then fix up the few outstanding concerns post-merge.

Thanks,
Matt

-- 
Matt Domsch
Sr. Software Engineer, Lead Engineer
Dell Linux Solutions linux.dell.com & www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com

Re: merge megaraid 2.20.x in kernel 2.6?

[James Bottomley]

On Tue, 2004-08-17 at 12:01, Matt Domsch wrote:
> James, is there anything you're still waiting on before merging the
> LSI megaraid 2.20.x driver into your scsi-misc-2.6 tree?  I thought
> the agreement by you, Christoph, and Atul was to get it merged, and
> then fix up the few outstanding concerns post-merge.

OK, I forgot in the rush of conferences.

I did put it in, but it fails to compile on parisc:

In file included from drivers/scsi/megaraid/megaraid_mm.c:18:
drivers/scsi/megaraid/megaraid_mm.h:51:25: asm/ioctl32.h: No such file
or directory
drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_init':
drivers/scsi/megaraid/megaraid_mm.c:1069: warning: implicit declaration
of function `register_ioctl32_conversion'
drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_exit':
drivers/scsi/megaraid/megaraid_mm.c:1098: warning: implicit declaration
of function `unregister_ioctl32_conversion'
make[3]: *** [drivers/scsi/megaraid/megaraid_mm.o] Error 1
make[2]: *** [drivers/scsi/megaraid] Error 2
make[1]: *** [drivers/scsi] Error 2
make: *** [drivers] Error 2

It looks like the compat ioctl infrastructure is incorrect.

James

Re: merge megaraid 2.20.x in kernel 2.6?

[Christoph Hellwig]

On Tue, Aug 17, 2004 at 12:40:33PM -0400, James Bottomley wrote:
> On Tue, 2004-08-17 at 12:01, Matt Domsch wrote:
> > James, is there anything you're still waiting on before merging the
> > LSI megaraid 2.20.x driver into your scsi-misc-2.6 tree?  I thought
> > the agreement by you, Christoph, and Atul was to get it merged, and
> > then fix up the few outstanding concerns post-merge.
> 
> OK, I forgot in the rush of conferences.
> 
> I did put it in, but it fails to compile on parisc:
> 
> In file included from drivers/scsi/megaraid/megaraid_mm.c:18:
> drivers/scsi/megaraid/megaraid_mm.h:51:25: asm/ioctl32.h: No such file
> or directory

No one should use <asm/ioctl32.h> but always <linux/ioctl32.h>

> drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_init':
> drivers/scsi/megaraid/megaraid_mm.c:1069: warning: implicit declaration
> of function `register_ioctl32_conversion'
> drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_exit':
> drivers/scsi/megaraid/megaraid_mm.c:1098: warning: implicit declaration
> of function `unregister_ioctl32_conversion'

That should also fix this one I guess, not having the current source
tree near me.  Does this driver release have the queue removal James
wanted?

Re: merge megaraid 2.20.x in kernel 2.6?

[Matt Domsch]

On Tue, Aug 17, 2004 at 05:47:55PM +0100, Christoph Hellwig wrote:
> Does this driver release have the queue removal James wanted?

Not yet, that was one of the changes Atul requested he be allowed to
make post-merge.

Thanks,
Matt

-- 
Matt Domsch
Sr. Software Engineer, Lead Engineer
Dell Linux Solutions linux.dell.com & www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com

Re: merge megaraid 2.20.x in kernel 2.6?

[James Bottomley]

On Tue, 2004-08-17 at 12:51, Matt Domsch wrote:
> Not yet, that was one of the changes Atul requested he be allowed to
> make post-merge.

OK, this seems to be what fixes the COMPAT_IOCTL problem.

James

===== drivers/scsi/megaraid/megaraid_mm.c 1.1 vs edited =====
--- 1.1/drivers/scsi/megaraid/megaraid_mm.c	2004-06-23 14:36:31 -05:00
+++ edited/drivers/scsi/megaraid/megaraid_mm.c	2004-08-17 11:56:03 -05:00
@@ -1075,7 +1075,7 @@
 /**
  * mraid_mm_compat_ioctl	: 32bit to 64bit ioctl conversion routine
  */
-#ifdef LSI_CONFIG_COMPAT
+#ifdef CONFIG_COMPAT
 static int
 mraid_mm_compat_ioctl(unsigned int fd, unsigned int cmd,
 			unsigned long arg, struct file *filep)
===== drivers/scsi/megaraid/megaraid_mm.h 1.1 vs edited =====
--- 1.1/drivers/scsi/megaraid/megaraid_mm.h	2004-06-23 10:53:53 -05:00
+++ edited/drivers/scsi/megaraid/megaraid_mm.h	2004-08-17 11:54:14 -05:00
@@ -41,18 +41,7 @@
 /*
  * Localizing ioctl32 differences
  */
-#if defined (CONFIG_COMPAT) || defined(__x86_64__) || defined(IA32_EMULATION)
-#if !defined(CONFIG_IA64)
-#define LSI_CONFIG_COMPAT
-#endif
-#endif
-
-#ifdef LSI_CONFIG_COMPAT
-#include <asm/ioctl32.h>
-#else
-#define register_ioctl32_conversion(a,b)	do{}while(0)
-#define unregister_ioctl32_conversion(a)	do{}while(0)
-#endif /* LSI_CONFIG_COMPAT */
+#include <linux/ioctl32.h>
 
 /**
  * mimd_t	: Old style ioctl packet structure (deprecated)
@@ -137,7 +126,7 @@
 static void mraid_mm_free_adp_resources(mraid_mmadp_t *);
 static void mraid_mm_teardown_dma_pools(mraid_mmadp_t *);
 
-#ifdef LSI_CONFIG_COMPAT
+#ifdef CONFIG_COMPAT
 static int mraid_mm_compat_ioctl(unsigned int, unsigned int, unsigned long,
 		struct file *);
 #endif

Re: merge megaraid 2.20.x in kernel 2.6?

[Arjan van de Ven]

[-- Attachment #1: Type: text/plain, Size: 457 bytes --]

On Tue, 2004-08-17 at 18:51, Matt Domsch wrote:
> On Tue, Aug 17, 2004 at 05:47:55PM +0100, Christoph Hellwig wrote:
> > Does this driver release have the queue removal James wanted?
> 
> Not yet, that was one of the changes Atul requested he be allowed to
> make post-merge.

if that was a few weeks ago.. that should be completed by now, right ?
I fear that when we merge this new driver we'll never get the queueing
removed once it's let in.


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

RE: merge megaraid 2.20.x in kernel 2.6?

[Mukker, Atul]

Arjan,

It will be :-) but the decision should not be made in a haste since two
other subsystems also use the same pending queue structure, the common
management module and the upcoming /sysfs. James idea of pushing those
packets to the mid-layer queues hasn't gotten much exposure yet. IMHO, that
isn't very clean. The respective sub-system should re-try in case the host
is busy.

There is another change pending, as suggested by Matt and Christoph,
function re-ordering. This also is pushed for post-merge since otherwise it
will be difficult to get it at all in the kernel. Remember, the first
submission was made sometime in May, when the stable kernel was 2.6!

Thanks
-Atul

> -----Original Message-----
> From: Arjan van de Ven [mailto:arjanv@redhat.com]
> Sent: Tuesday, August 17, 2004 1:07 PM
> To: Matt Domsch
> Cc: Christoph Hellwig; James Bottomley; atulm@lsil.com; SCSI Mailing
> List
> Subject: Re: merge megaraid 2.20.x in kernel 2.6?
> 
> 
> On Tue, 2004-08-17 at 18:51, Matt Domsch wrote:
> > On Tue, Aug 17, 2004 at 05:47:55PM +0100, Christoph Hellwig wrote:
> > > Does this driver release have the queue removal James wanted?
> > 
> > Not yet, that was one of the changes Atul requested he be allowed to
> > make post-merge.
> 
> if that was a few weeks ago.. that should be completed by now, right ?
> I fear that when we merge this new driver we'll never get the queueing
> removed once it's let in.
> 
> 

RE: merge megaraid 2.20.x in kernel 2.6?

[Mukker, Atul]

James,

I have released the 2.20.3.1 drive, which should fix this issue. Did you get
a chance to merge that driver?

Thanks
-Atul

> -----Original Message-----
> From: James Bottomley [mailto:James.Bottomley@SteelEye.com]
> Sent: Tuesday, August 17, 2004 12:41 PM
> To: Matt Domsch; atulm@lsil.com
> Cc: SCSI Mailing List
> Subject: Re: merge megaraid 2.20.x in kernel 2.6?
> 
> 
> On Tue, 2004-08-17 at 12:01, Matt Domsch wrote:
> > James, is there anything you're still waiting on before merging the
> > LSI megaraid 2.20.x driver into your scsi-misc-2.6 tree?  I thought
> > the agreement by you, Christoph, and Atul was to get it merged, and
> > then fix up the few outstanding concerns post-merge.
> 
> OK, I forgot in the rush of conferences.
> 
> I did put it in, but it fails to compile on parisc:
> 
> In file included from drivers/scsi/megaraid/megaraid_mm.c:18:
> drivers/scsi/megaraid/megaraid_mm.h:51:25: asm/ioctl32.h: No such file
> or directory
> drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_init':
> drivers/scsi/megaraid/megaraid_mm.c:1069: warning: implicit 
> declaration
> of function `register_ioctl32_conversion'
> drivers/scsi/megaraid/megaraid_mm.c: In function `mraid_mm_exit':
> drivers/scsi/megaraid/megaraid_mm.c:1098: warning: implicit 
> declaration
> of function `unregister_ioctl32_conversion'
> make[3]: *** [drivers/scsi/megaraid/megaraid_mm.o] Error 1
> make[2]: *** [drivers/scsi/megaraid] Error 2
> make[1]: *** [drivers/scsi] Error 2
> make: *** [drivers] Error 2
> 
> It looks like the compat ioctl infrastructure is incorrect.
> 
> James
> 
>