Re: [PATCH 1/5] dynamic logical partitioning infrastructure

[Nathan Fontenot <nfont@austin.ibm.com>]

Brian King wrote:
> Nathan Fontenot wrote:
>> +#include <linux/kernel.h>
>> +#include <linux/kref.h>
>> +#include <linux/notifier.h>
>> +#include <linux/proc_fs.h>
>> +#include <linux/spinlock.h>
>> +
>> +#include <asm/prom.h>
>> +#include <asm/machdep.h>
>> +#include <asm/uaccess.h>
>> +#include <asm/rtas.h>
>> +#include <asm/pSeries_reconfig.h>
>> +
>> +#define CFG_CONN_WORK_SIZE    4096
>> +static char workarea[CFG_CONN_WORK_SIZE];
>> +spinlock_t workarea_lock;
> 
> This can be:
> 
> static DEFINE_SPINLOCK(workarea_lock);
> 
> Then you can get rid of the runtime initializer.

Good catch,  I will fix it in the updated patches.

> 
>> +
>> +int release_drc(u32 drc_index)
>> +{
>> +    int dr_status, rc;
>> +
>> +    rc = rtas_call(rtas_token("get-sensor-state"), 2, 2, &dr_status,
>> +               DR_ENTITY_SENSE, drc_index);
>> +    if (rc || dr_status != DR_ENTITY_PRESENT)
>> +        return -1;
>> +
>> +    rc = rtas_set_indicator(ISOLATION_STATE, drc_index, ISOLATE);
>> +    if (rc)
>> +        return -1;
>> +
>> +    rc = rtas_set_indicator(ALLOCATION_STATE, drc_index, ALLOC_UNUSABLE);
>> +    if (rc) {
>> +        rtas_set_indicator(ISOLATION_STATE, drc_index, UNISOLATE);
>> +        return -1;
>> +    }
> 
> Is there a better return value here that might be more descriptive than -1?

Yes,  I could return the rtas error code to the user to allow the caller to
evaluate it if they wanted to.  For what I am doing I am only concerned with
success/failure so I did not deal with returning anything other than -1.

I'll update the next patch to return the rtas error for failures and 0
for success.

> 
> 
>> +
>> +    return 0;
>> +}
>> +
>> +static int pseries_dlpar_init(void)
>> +{
>> +    spin_lock_init(&workarea_lock);
>> +
>> +    if (!machine_is(pseries))
>> +        return 0;
> 
> What's the point of this if check if you return 0 either way?

Yes, it seems a bit odd here, but in patches later in this series I
add additional initialization steps after the machine_is() check
such that it makes sense to bail out here if the check fails.

> 
>> +
>> +    return 0;
>> +}
>> +__initcall(pseries_dlpar_init);
> 
> 
>> Index: powerpc/arch/powerpc/platforms/pseries/reconfig.c
>> ===================================================================
>> --- powerpc.orig/arch/powerpc/platforms/pseries/reconfig.c    2009-09-11
>> 12:43:39.000000000 -0500
>> +++ powerpc/arch/powerpc/platforms/pseries/reconfig.c    2009-09-11
>> 12:51:52.000000000 -0500
>> @@ -95,7 +95,7 @@
>>     return parent;
>> }
>>
>> -static BLOCKING_NOTIFIER_HEAD(pSeries_reconfig_chain);
>> +struct blocking_notifier_head pSeries_reconfig_chain =
>> BLOCKING_NOTIFIER_INIT(pSeries_reconfig_chain);
> 
> Can't this just be?
> 
> BLOCKING_NOTIFIER_HEAD(pSeries_reconfig_chain);
> 

I think I tried this and was having issues,  I don't remember what they
were though.  I will try to fix this in the updated patch.

-Nathan Fontenot