[PATCH net-next 5/6] net: ti: icss-iep: Move icss_iep structure
MD Danish Anwar
danishanwar at ti.com
Sun Aug 11 22:41:21 PDT 2024
Hi Dan,
On 10/08/24 1:40 am, Dan Carpenter wrote:
> On Thu, Aug 08, 2024 at 04:37:59PM +0530, MD Danish Anwar wrote:
>> - struct ptp_clock *ptp_clock;
>> - struct mutex ptp_clk_mutex; /* PHC access serializer */
>> - u32 def_inc;
>> - s16 slow_cmp_inc;
>
> [ cut ]
>
>> + struct ptp_clock *ptp_clock;
>> + struct mutex ptp_clk_mutex; /* PHC access serializer */
>> + spinlock_t irq_lock; /* CMP IRQ vs icss_iep_ptp_enable access */
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> The patch adds this new struct member. When you're moving code around, please
> just move the code. Don't fix checkpatch warnings or do any other cleanups.
>
My bad. I didn't notice this new struct member was introduced. I will
take care of it.
Also apart from doing the code movement, this patch also does the
following change. Instead of hardcoding the value 4, the patch uses
emac->iep->def_inc. Since the iep->def_inc is now accessible from
drivers/net/ethernet/ti/icssg/icssg_prueth.c
@@ -384,7 +384,8 @@ static void prueth_iep_settime(void *clockops_data,
u64 ns)
sc_desc.cyclecounter0_set = cyclecount & GENMASK(31, 0);
sc_desc.cyclecounter1_set = (cyclecount & GENMASK(63, 32)) >> 32;
sc_desc.iepcount_set = ns % cycletime;
- sc_desc.CMP0_current = cycletime - 4; //Count from 0 to (cycle time)-4
+ /* Count from 0 to (cycle time) - emac->iep->def_inc */
+ sc_desc.CMP0_current = cycletime - emac->iep->def_inc;
memcpy_toio(sc_descp, &sc_desc, sizeof(sc_desc));
Should I keep the above change as it is or should I split it into a
separate patch and make this patch strictly for code movement. I kept
the above change as part of this patch because it is related with the
code movement. Moving the iep structure from icss_iep.c to icss_iep.h
makes it accessible to icssg_prueth.c so I thought keeping them together
will be a better idea.
Please let me know if this is okay.
>> + u32 def_inc;
>> + s16 slow_cmp_inc;
>> + u32 slow_cmp_count;
>
> regards,
> dan carpenter
--
Thanks and Regards,
Danish
More information about the linux-arm-kernel
mailing list