svn commit: r216375 - head/sys/mips/rmi/dev/xlr

Jayachandran C. jchandra at freebsd.org
Mon Dec 13 17:05:07 UTC 2010


On Mon, Dec 13, 2010 at 6:17 PM, John Baldwin <jhb at freebsd.org> wrote:
> On Saturday, December 11, 2010 5:21:38 am Andriy Gapon wrote:
>> Author: avg
>> Date: Sat Dec 11 10:21:38 2010
>> New Revision: 216375
>> URL: http://svn.freebsd.org/changeset/base/216375
>>
>> Log:
>>   fix atomic_set_xxx misuse in rge
>>
>>   It seems that atomic_set_xxx and atomic_store_xxx were confused.
>>
>>   Reviewed by:        jhb (general issue)
>>   MFC after:  3 weeks
>>
>> Modified:
>>   head/sys/mips/rmi/dev/xlr/rge.c
>>
>> Modified: head/sys/mips/rmi/dev/xlr/rge.c
>> ==============================================================================
>> --- head/sys/mips/rmi/dev/xlr/rge.c   Sat Dec 11 10:18:05 2010        (r216374)
>> +++ head/sys/mips/rmi/dev/xlr/rge.c   Sat Dec 11 10:21:38 2010        (r216375)
>> @@ -170,7 +170,7 @@ extern uint32_t cpu_ltop_map[32];
>>  static int port_counters[4][8] __aligned(XLR_CACHELINE_SIZE);
>>
>>  #define port_inc_counter(port, counter)      atomic_add_int(&port_counters[port][(counter)], 1)
>> -#define port_set_counter(port, counter, value)       atomic_set_int(&port_counters[port][(counter)], (value))
>> +#define port_set_counter(port, counter, value)       atomic_store_int(&port_counters[port][(counter)], (value))
>>  #else
>>  #define port_inc_counter(port, counter)      /* Nothing */
>>  #define port_set_counter(port, counter, value)       /* Nothing */
>> @@ -2281,7 +2281,7 @@ rmi_xlr_mac_open(struct rge_softc *sc)
>>       mtx_unlock_spin(&priv->lock);
>>
>>       for (i = 0; i < 8; i++) {
>> -             atomic_set_int(&(priv->frin_to_be_sent[i]), 0);
>> +             atomic_store_int(&(priv->frin_to_be_sent[i]), 0);
>
> Err, there is no such op for this.  If you don't need a memory barrier then a
> simple assignment to 0 should suffice.

This is the deprecated  driver for the XLR's on-chip network
interface. There is a new driver for this, so you can ignore problems
in this file - it should be removed soon.

Thanks,
JC


More information about the svn-src-all mailing list