Re: linux-next: Tree for Jun 1 (scsi)

Re: linux-next: Tree for Jun 1 (scsi)

[Randy Dunlap]

On 05/31/2012 08:22 PM, Stephen Rothwell wrote:

> Hi all,
> 
> Changes since 20120531:



already fixed?

on i386 and x86_64, loadable module or builtin:


ERROR: "scsi_sd_probe_domain" [drivers/scsi/scsi_mod.ko] undefined!
or
(.text+0x4f6c77): undefined reference to `scsi_sd_probe_domain'

-- 

~Randy
(sorry about the delay, was out for 3 days)

Re: linux-next: Tree for Jun 1 (scsi)

[James Bottomley]

On Sun, 2012-06-03 at 20:31 -0700, Randy Dunlap wrote:
> On 05/31/2012 08:22 PM, Stephen Rothwell wrote:
> 
> > Hi all,
> > 
> > Changes since 20120531:
> 
> 
> 
> already fixed?
> 
> on i386 and x86_64, loadable module or builtin:
> 
> 
> ERROR: "scsi_sd_probe_domain" [drivers/scsi/scsi_mod.ko] undefined!
> or
> (.text+0x4f6c77): undefined reference to `scsi_sd_probe_domain'

This looks to be an obvious thinko:  the symbol is conditionally
defined, but unconditionally used in
scsi_scan.c:scsi_complete_async_scans().  We can fix just by removing
the wrong conditional.

James

---

diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index 61c82a3..bbbc9c9 100644
--- a/drivers/scsi/scsi.c
+++ b/drivers/scsi/scsi.c
@@ -90,11 +90,9 @@ unsigned int scsi_logging_level;
 EXPORT_SYMBOL(scsi_logging_level);
 #endif
 
-#if IS_ENABLED(CONFIG_PM) || IS_ENABLED(CONFIG_BLK_DEV_SD)
-/* sd and scsi_pm need to coordinate flushing async actions */
+/* sd, scsi core and power management need to coordinate flushing async actions */
 LIST_HEAD(scsi_sd_probe_domain);
 EXPORT_SYMBOL(scsi_sd_probe_domain);
                                -#endif
 
 /* NB: These are exposed through /proc/scsi/scsi and form part of the ABI.
  * You may not alter any existing entry (although adding new ones is

Re: linux-next: Tree for Jun 1 (scsi)

[Randy Dunlap]

On 06/05/2012 10:54 PM, James Bottomley wrote:

> On Sun, 2012-06-03 at 20:31 -0700, Randy Dunlap wrote:
>> On 05/31/2012 08:22 PM, Stephen Rothwell wrote:
>>
>>> Hi all,
>>>
>>> Changes since 20120531:
>>
>>
>>
>> already fixed?
>>
>> on i386 and x86_64, loadable module or builtin:
>>
>>
>> ERROR: "scsi_sd_probe_domain" [drivers/scsi/scsi_mod.ko] undefined!
>> or
>> (.text+0x4f6c77): undefined reference to `scsi_sd_probe_domain'
> 
> This looks to be an obvious thinko:  the symbol is conditionally
> defined, but unconditionally used in
> scsi_scan.c:scsi_complete_async_scans().  We can fix just by removing
> the wrong conditional.
> 
> James
> 
> ---
> 
> diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
> index 61c82a3..bbbc9c9 100644
> --- a/drivers/scsi/scsi.c
> +++ b/drivers/scsi/scsi.c
> @@ -90,11 +90,9 @@ unsigned int scsi_logging_level;
>  EXPORT_SYMBOL(scsi_logging_level);
>  #endif
>  
> -#if IS_ENABLED(CONFIG_PM) || IS_ENABLED(CONFIG_BLK_DEV_SD)
> -/* sd and scsi_pm need to coordinate flushing async actions */
> +/* sd, scsi core and power management need to coordinate flushing async actions */
>  LIST_HEAD(scsi_sd_probe_domain);
>  EXPORT_SYMBOL(scsi_sd_probe_domain);
>                                 -#endif


Yes, this patch works after I move that "-#endif" to column 1 (or 0).

Thanks.

Acked-by: Randy Dunlap <rdunlap@xenotime.net>


>  

>  /* NB: These are exposed through /proc/scsi/scsi and form part of the ABI.
>   * You may not alter any existing entry (although adding new ones is
> 
> 



-- 
~Randy