[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <55eeec88-63ea-4838-865a-17493dfb847d@ti.com>
Date: Mon, 13 May 2024 11:56:17 +0530
From: MD Danish Anwar <danishanwar@...com>
To: Paolo Abeni <pabeni@...hat.com>, Dan Carpenter <dan.carpenter@...aro.org>,
Andrew Lunn <andrew@...n.ch>, Jan Kiszka <jan.kiszka@...mens.com>,
Simon
Horman <horms@...nel.org>,
Niklas Schnelle <schnelle@...ux.ibm.com>,
Randy
Dunlap <rdunlap@...radead.org>,
Diogo Ivo <diogo.ivo@...mens.com>,
Wolfram
Sang <wsa+renesas@...g-engineering.com>,
Vignesh Raghavendra
<vigneshr@...com>,
Richard Cochran <richardcochran@...il.com>,
Roger Quadros
<rogerq@...nel.org>, Jakub Kicinski <kuba@...nel.org>,
Eric Dumazet
<edumazet@...gle.com>,
"David S. Miller" <davem@...emloft.net>
CC: <linux-arm-kernel@...ts.infradead.org>, <netdev@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <srk@...com>, <r-gunasekaran@...com>,
Roger
Quadros <rogerq@...com>,
Vinicius Costa Gomes <vinicius.gomes@...el.com>,
Vladimir Oltean <vladimir.oltean@....com>
Subject: Re: [PATCH net-next v5] net: ti: icssg_prueth: add TAPRIO offload
support
Hi Paolo,
On 02/05/24 5:29 pm, Paolo Abeni wrote:
> On Mon, 2024-04-29 at 16:00 +0530, MD Danish Anwar wrote:
>> +static int emac_taprio_replace(struct net_device *ndev,
>> + struct tc_taprio_qopt_offload *taprio)
>> +{
>> + struct prueth_emac *emac = netdev_priv(ndev);
>> + struct tc_taprio_qopt_offload *est_new;
>> + int ret;
>> +
>> + if (taprio->cycle_time_extension) {
>> + NL_SET_ERR_MSG_MOD(taprio->extack, "Cycle time extension not supported");
>> + return -EOPNOTSUPP;
>> + }
>> +
>> + if (taprio->cycle_time < TAS_MIN_CYCLE_TIME) {
>> + NL_SET_ERR_MSG_FMT_MOD(taprio->extack, "cycle_time %llu is less than min supported cycle_time %d",
>> + taprio->cycle_time, TAS_MIN_CYCLE_TIME);
>> + return -EINVAL;
>> + }
>> +
>> + if (taprio->num_entries > TAS_MAX_CMD_LISTS) {
>> + NL_SET_ERR_MSG_FMT_MOD(taprio->extack, "num_entries %lu is more than max supported entries %d",
>> + taprio->num_entries, TAS_MAX_CMD_LISTS);
>> + return -EINVAL;
>> + }
>> +
>> + if (emac->qos.tas.taprio_admin)
>> + devm_kfree(&ndev->dev, emac->qos.tas.taprio_admin);
>
> it looks like 'qos.tas.taprio_admin' is initialized from
> taprio_offload_get(), so it should be free with taprio_offload_free(),
> right?
>
'qos.tas.taprio_admin' is assigned by "emac->qos.tas.taprio_admin =
taprio_offload_get(taprio);". Here I will free it with taprio_offload_free()
>> +
>> + est_new = devm_kzalloc(&ndev->dev,
>> + struct_size(est_new, entries, taprio->num_entries),
>> + GFP_KERNEL);
>> + if (!est_new)
>> + return -ENOMEM;
>
> Why are you allocating 'est_new'? it looks like it's not used
> anywhere?!?
>
Sorry my bad. Forgot to remove est_new. I will remove it.
>> +
>> + emac->qos.tas.taprio_admin = taprio_offload_get(taprio);
>> + ret = tas_update_oper_list(emac);
>> + if (ret)
>> + return ret;
>
> Should the above clear 'taprio_admin' on error, as well?
>
Yes here also we should clear taprio_admin and taprio on error. I will
add a goto label and clear taprio on both errors.
Below is the diff to address handling of taprio and taprio_admin.
diff --git a/drivers/net/ethernet/ti/icssg/icssg_qos.c
b/drivers/net/ethernet/ti/icssg/icssg_qos.c
index 459463ea6c20..c7cadab0edec 100644
--- a/drivers/net/ethernet/ti/icssg/icssg_qos.c
+++ b/drivers/net/ethernet/ti/icssg/icssg_qos.c
@@ -210,7 +210,7 @@ static int emac_taprio_replace(struct net_device *ndev,
}
if (emac->qos.tas.taprio_admin)
- devm_kfree(&ndev->dev, emac->qos.tas.taprio_admin);
+ taprio_offload_free(emac->qos.tas.taprio_admin);
est_new = devm_kzalloc(&ndev->dev,
struct_size(est_new, entries, taprio->num_entries),
@@ -221,13 +221,15 @@ static int emac_taprio_replace(struct net_device
*ndev,
emac->qos.tas.taprio_admin = taprio_offload_get(taprio);
ret = tas_update_oper_list(emac);
if (ret)
- return ret;
+ goto clear_taprio;
ret = tas_set_state(emac, TAS_STATE_ENABLE);
- if (ret) {
- emac->qos.tas.taprio_admin = NULL;
- taprio_offload_free(taprio);
- }
+ if (ret)
+ goto clear_taprio;
+
+clear_taprio:
+ emac->qos.tas.taprio_admin = NULL;
+ taprio_offload_free(taprio);
return ret;
}
Please have a look and let me know if this looks ok to you.
>>
> Thanks,
>
> Paolo
>
--
Thanks and Regards,
Danish
Powered by blists - more mailing lists