RE: [01/17]PATCH Add API for allocating dynamic TR resouce. V8

["Zhang, Xiantao" <xiantao.zhang@xxxxxxxxx>]

Carsten Otte wrote:
> Zhang, Xiantao wrote:
>> +/*  mca_insert_tr
>> + *
>> + *  Switch rid when TR reload and needed!
>> + *  iord: 1: itr, 2: itr;
>> + *
>> +*/
>> +static void mca_insert_tr(u64 iord)
>> +{
>> +
>> +	int i;
>> +	u64 old_rr;
>> +	struct ia64_tr_entry *p;
>> +	unsigned long psr;
>> +	int cpu = smp_processor_id();
> What if CONFIG_PREEMPT is on, and we're being preempted and scheduled
> to a different CPU here? Are we running preempt disabled here? If so,
> the function header should state that this function needs to be called
> preempt_disabled.

The function insert one TR to local TLB, and doesn't allow preempt
before and after the call, so the caller should be with preempt_disable
before calling into this routine. 
Maybe the descripiton of this function should contain "Called with
preempt disabled!".  Does it make sense ?
Xiantao


>> +/*
>> + * ia64_insert_tr in virtual mode. Allocate a TR slot + *
>> + * target_mask : 0x1 : itr, 0x2 : dtr, 0x3 : idtr
>> + *
>> + * va 	: virtual address.
>> + * pte 	: pte entries inserted.
>> + * log_size: range to be covered.
>> + *
>> + * Return value:  <0 :  error No.
>> + *
>> + *		  >=0 : slot number allocated for TR.
>> + */
>> +int ia64_itr_entry(u64 target_mask, u64 va, u64 pte, u64 log_size)
>> +{ +	int i, r;
>> +	unsigned long psr;
>> +	struct ia64_tr_entry *p;
>> +	int cpu = smp_processor_id();
> Same here.
> 
>> +/*
>> + * ia64_purge_tr
>> + *
>> + * target_mask: 0x1: purge itr, 0x2 : purge dtr, 0x3 purge idtr. + *
>> + * slot: slot number to be freed.
>> + */
>> +void ia64_ptr_entry(u64 target_mask, int slot)
>> +{
>> +	int cpu = smp_processor_id();
>> +	int i;
>> +	struct ia64_tr_entry *p;
> Here again.

--
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html