2.4.28-pre3: broken ips update

2.4.28-pre3: broken ips update

[Adrian Bunk]

On Sat, Sep 11, 2004 at 07:01:17PM -0300, Marcelo Tosatti wrote:
>...
> Summary of changes from v2.4.28-pre2 to v2.4.28-pre3
> ============================================
>...
> Jack Hammer:
>   o ServeRAID driver (ips) Version 7.10.18
>...

<--  snip  -->

...
gcc-3.4 -D__KERNEL__ 
-I/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include -Wall 
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common 
-fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=athlon 
-fno-unit-at-a-time   -nostdinc -iwithprefix include 
-DKBUILD_BASENAME=ips  -c -o ips.o ips.c
In file included from ips.c:190:
ips.h:101: error: redefinition of typedef 'irqreturn_t'
/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include/linux/interrupt.h:16: 
error: previous declaration of 'irqreturn_t' was here
make[3]: *** [ips.o] Error 1
make[3]: Leaving directory `/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/drivers/scsi'

<--  snip  -->


This update was obviously submitted without even testing the compilation 
of the driver (not to mention testing whether it actually works).


Even worse:

We had _exactly the same problem_ with the 7.00.15 driver in 
2.4.27-pre3, but 7.10.18 now simply reverts my (trivial) fix for this 
compile error.


Marcelo, can you please refuse patches from people who don't bother to 
do even the simplest compile testing when blindly submitting the latest 
version of a driver?


TIA
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: 2.4.28-pre3: broken ips update

[Mikael Pettersson]

Adrian Bunk writes:
 > On Sat, Sep 11, 2004 at 07:01:17PM -0300, Marcelo Tosatti wrote:
 > >...
 > > Summary of changes from v2.4.28-pre2 to v2.4.28-pre3
 > > ============================================
 > >...
 > > Jack Hammer:
 > >   o ServeRAID driver (ips) Version 7.10.18
 > >...
 > 
 > <--  snip  -->
 > 
 > ...
 > gcc-3.4 -D__KERNEL__ 
 > -I/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include -Wall 
 > -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common 
 > -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=athlon 
 > -fno-unit-at-a-time   -nostdinc -iwithprefix include 
 > -DKBUILD_BASENAME=ips  -c -o ips.o ips.c
 > In file included from ips.c:190:
 > ips.h:101: error: redefinition of typedef 'irqreturn_t'
 > /home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include/linux/interrupt.h:16: 
 > error: previous declaration of 'irqreturn_t' was here
 > make[3]: *** [ips.o] Error 1
 > make[3]: Leaving directory `/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/drivers/scsi'

I posted a patch yesterday fixing this very issue.
The compat definitions in ips.h are in conflict with the
kernel's own compat definitions in interrupt.h. The
workaround is to disable ips.h's irqreturn_t etc emulation
for kernels >= 2.4.23.

/Mikael

RE: 2.4.28-pre3: broken ips update

[Hammer, Jack]

[-- Attachment #1: Type: text/plain, Size: 3248 bytes --]

Marcelo,

This trivial one line patch corrects my mistake in ips.h. I assure you
that the previous patch was compiled and tested extensively - over a
several month test cycle.  However, our testing concentrates on existing
customers and shipping distro's.  Most of those changed to the 2.6
kernel on or before 2.4.20, so we had no 2.4 kernel in our test matrix
beyond that. I apologize for the oversight - we must update our test
matrix to always check out latest 2.4 development kernel.

The concerns expressed below about this driver being "blindly submitted"
or is "untested" are completely unwarranted.

Thanks.

Jack



--- linux.orig/drivers/scsi/ips.h	Mon Sep 13 09:40:04 2004
+++ linux/drivers/scsi/ips.h	Mon Sep 13 09:40:27 2004
@@ -97,7 +97,7 @@
 
    #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
    
-      #ifndef irqreturn_t
+      #if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,23)
          typedef void irqreturn_t;
       #endif 







 

-----Original Message-----
From: linux-scsi-owner@vger.kernel.org
[mailto:linux-scsi-owner@vger.kernel.org] On Behalf Of Adrian Bunk
Sent: Monday, September 13, 2004 3:04 AM
To: Marcelo Tosatti; IpsLinux
Cc: linux-kernel@vger.kernel.org; linux-scsi@vger.kernel.org
Subject: 2.4.28-pre3: broken ips update

On Sat, Sep 11, 2004 at 07:01:17PM -0300, Marcelo Tosatti wrote:
>...
> Summary of changes from v2.4.28-pre2 to v2.4.28-pre3  
>============================================
>...
> Jack Hammer:
>   o ServeRAID driver (ips) Version 7.10.18 ...

<--  snip  -->

...
gcc-3.4 -D__KERNEL__
-I/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=athlon 
-fno-unit-at-a-time   -nostdinc -iwithprefix include 
-DKBUILD_BASENAME=ips  -c -o ips.o ips.c In file included from
ips.c:190:
ips.h:101: error: redefinition of typedef 'irqreturn_t'
/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/include/linux/interru
pt.h:16: 
error: previous declaration of 'irqreturn_t' was here
make[3]: *** [ips.o] Error 1
make[3]: Leaving directory
`/home/bunk/linux/kernel-2.4/linux-2.4.28-pre3-full/drivers/scsi'

<--  snip  -->


This update was obviously submitted without even testing the compilation
of the driver (not to mention testing whether it actually works).


Even worse:

We had _exactly the same problem_ with the 7.00.15 driver in 
2.4.27-pre3, but 7.10.18 now simply reverts my (trivial) fix for this 
compile error.


Marcelo, can you please refuse patches from people who don't bother to 
do even the simplest compile testing when blindly submitting the latest 
version of a driver?


TIA
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[-- Attachment #2: patch_ips.h --]
[-- Type: application/octet-stream, Size: 330 bytes --]

--- linux.orig/drivers/scsi/ips.h	Mon Sep 13 09:40:04 2004
+++ linux/drivers/scsi/ips.h	Mon Sep 13 09:40:27 2004
@@ -97,7 +97,7 @@
 
    #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
    
-      #ifndef irqreturn_t
+      #if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,23)
          typedef void irqreturn_t;
       #endif 
       

Re: 2.4.28-pre3: broken ips update

[Matthew Wilcox]

On Mon, Sep 13, 2004 at 10:28:44AM -0400, Hammer, Jack wrote:
> This trivial one line patch corrects my mistake in ips.h. I assure you
> that the previous patch was compiled and tested extensively - over a
> several month test cycle.  However, our testing concentrates on existing
> customers and shipping distro's.  Most of those changed to the 2.6
> kernel on or before 2.4.20, so we had no 2.4 kernel in our test matrix
> beyond that. I apologize for the oversight - we must update our test
> matrix to always check out latest 2.4 development kernel.
> 
> The concerns expressed below about this driver being "blindly submitted"
> or is "untested" are completely unwarranted.

I disagree.  Your testing, while I'm sure is long drawn out and massively
complicated, is missing the crucial element of testing what you're
submitting against.  I don't think you need to check the latest bleeding
unstable kernel in your test runs, partly because it changes so fast, and
partly because you'll start to hit bugs that are unrelated to your driver.

But you absolutely must at least test compiling your driver against the
tree you submit it against, and a quick boot of the resulting kernel
wouldn't add much time to your patch submission process.

-- 
"Next the statesmen will invent cheap lies, putting the blame upon 
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince 
himself that the war is just, and will thank God for the better sleep 
he enjoys after this process of grotesque self-deception." -- Mark Twain

RE: 2.4.28-pre3: broken ips update

[Arjan van de Ven]

[-- Attachment #1: Type: text/plain, Size: 654 bytes --]

On Mon, 2004-09-13 at 16:28, Hammer, Jack wrote:
> Marcelo,

> --- linux.orig/drivers/scsi/ips.h	Mon Sep 13 09:40:04 2004
> +++ linux/drivers/scsi/ips.h	Mon Sep 13 09:40:27 2004
> @@ -97,7 +97,7 @@
>  
>     #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
>     
> -      #ifndef irqreturn_t
> +      #if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,23)
>           typedef void irqreturn_t;
>        #endif 
> 

your fix is wrong.
you should nuke all version codes, and ONLY ifndef for  IRQ_NONE and
nothing else. That's not just cosmetics, that's also to make keep
working with those exact vendor kernels  you claim to have tested on :)


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

RE: 2.4.28-pre3: broken ips update

[Hammer, Jack]

I concur with Arjan's critique.  

Please disregard this patch and I will create a new one, with Arjan's
recommendation, shortly.

Jack 



-----Original Message-----
From: Arjan van de Ven [mailto:arjanv@redhat.com] 
Sent: Monday, September 13, 2004 10:46 AM
To: Hammer, Jack
Cc: Adrian Bunk; Marcelo Tosatti; linux-kernel@vger.kernel.org;
linux-scsi@vger.kernel.org
Subject: RE: 2.4.28-pre3: broken ips update

On Mon, 2004-09-13 at 16:28, Hammer, Jack wrote:
> Marcelo,

> --- linux.orig/drivers/scsi/ips.h	Mon Sep 13 09:40:04 2004
> +++ linux/drivers/scsi/ips.h	Mon Sep 13 09:40:27 2004
> @@ -97,7 +97,7 @@
>  
>     #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
>     
> -      #ifndef irqreturn_t
> +      #if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,23)
>           typedef void irqreturn_t;
>        #endif
> 

your fix is wrong.
you should nuke all version codes, and ONLY ifndef for  IRQ_NONE and
nothing else. That's not just cosmetics, that's also to make keep
working with those exact vendor kernels  you claim to have tested on :)

2.4.28-pre3: broken ips update

[Hammer, Jack]

This patch replaces the patch I submitted ( and then rescinded ) earlier
today. It fixes the compilation problem in 2.4.28 and includes the
changes recommended by Arjan on my earlier proposal for this.

Jack 




--- linux.orig/drivers/scsi/ips.h	Mon Sep 13 12:51:04 2004
+++ linux/drivers/scsi/ips.h	Mon Sep 13 12:55:15 2004
@@ -95,15 +95,14 @@
       #define scsi_set_pci_device(sh,dev) (0)
    #endif
 
-   #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
-   
-      #ifndef irqreturn_t
-         typedef void irqreturn_t;
-      #endif 
-      
+   #ifndef IRQ_NONE
+      typedef void irqreturn_t;
       #define IRQ_NONE
       #define IRQ_HANDLED
       #define IRQ_RETVAL(x)
+   #endif
+   
+   #if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
       #define IPS_REGISTER_HOSTS(SHT)
scsi_register_module(MODULE_SCSI_HA,SHT)
       #define IPS_UNREGISTER_HOSTS(SHT)
scsi_unregister_module(MODULE_SCSI_HA,SHT)
       #define IPS_ADD_HOST(shost,device)