lkml.org 
[lkml]   [2023]   [Sep]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH net-next v4 03/18] net/smc: extract v2 check helper from SMC-D device registration
From


On 2023/9/28 11:08, Jan Karcher wrote:
>
>
> On 24/09/2023 17:16, Wen Gu wrote:
>> This patch extracts v2-capable logic from the process of registering the
>> ISM device as an SMC-D device, so that the registration process of other
>> underlying devices can reuse it.
>>
>> Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
>> ---
>>   net/smc/smc_ism.c | 29 ++++++++++++++++++-----------
>>   net/smc/smc_ism.h |  1 +
>>   2 files changed, 19 insertions(+), 11 deletions(-)
>>
>> diff --git a/net/smc/smc_ism.c b/net/smc/smc_ism.c
>> index 455ae0a..8f1ba74 100644
>> --- a/net/smc/smc_ism.c
>> +++ b/net/smc/smc_ism.c
>> @@ -69,6 +69,22 @@ bool smc_ism_is_v2_capable(void)
>>       return smc_ism_v2_capable;
>>   }
>> +/* must be called under smcd_dev_list.mutex lock */
>> +void smc_ism_check_v2_capable(struct smcd_dev *smcd)
>> +{
>> +    u8 *system_eid = NULL;
>> +
>> +    if (smc_ism_v2_capable)
>> +        return;
>> +
>> +    system_eid = smcd->ops->get_system_eid();
>> +    if (smcd->ops->supports_v2()) {
>> +        smc_ism_v2_capable = true;
>> +        memcpy(smc_ism_v2_system_eid, system_eid,
>> +               SMC_MAX_EID_LEN);
>> +    }
>> +}
>> +
>>   /* Set a connection using this DMBE. */
>>   void smc_ism_set_conn(struct smc_connection *conn)
>>   {
>> @@ -423,16 +439,7 @@ static void smcd_register_dev(struct ism_dev *ism)
>>           smc_pnetid_by_table_smcd(smcd);
>>       mutex_lock(&smcd_dev_list.mutex);
>> -    if (list_empty(&smcd_dev_list.list)) {
>> -        u8 *system_eid = NULL;
>> -
>> -        system_eid = smcd->ops->get_system_eid();
>> -        if (smcd->ops->supports_v2()) {
>> -            smc_ism_v2_capable = true;
>> -            memcpy(smc_ism_v2_system_eid, system_eid,
>> -                   SMC_MAX_EID_LEN);
>> -        }
>> -    }
>> +    smc_ism_check_v2_capable(smcd);
>
> The list_empty check is omitted here which means the smc_ism_check_v2_capable does not touch the list.
> So i think the call could be placed prior the mutex_lock.
>

Good catch. I omitted the list_empty check in this version but forget to remove 'the
lock comments' and place the helper prior to the mutex_lock. It will be fixed.

Thank you.

>>       /* sort list: devices without pnetid before devices with pnetid */
>>       if (smcd->pnetid[0])
>>           list_add_tail(&smcd->list, &smcd_dev_list.list);
>> @@ -535,10 +542,10 @@ int smc_ism_init(void)
>>   {
>>       int rc = 0;
>> -#if IS_ENABLED(CONFIG_ISM)
>>       smc_ism_v2_capable = false;
>>       memset(smc_ism_v2_system_eid, 0, SMC_MAX_EID_LEN);
>> +#if IS_ENABLED(CONFIG_ISM)
>>       rc = ism_register_client(&smc_ism_client);
>>   #endif
>>       return rc;
>> diff --git a/net/smc/smc_ism.h b/net/smc/smc_ism.h
>> index 832b2f4..14d2e77 100644
>> --- a/net/smc/smc_ism.h
>> +++ b/net/smc/smc_ism.h
>> @@ -42,6 +42,7 @@ int smc_ism_register_dmb(struct smc_link_group *lgr, int buf_size,
>>   void smc_ism_get_system_eid(u8 **eid);
>>   u16 smc_ism_get_chid(struct smcd_dev *dev);
>>   bool smc_ism_is_v2_capable(void);
>> +void smc_ism_check_v2_capable(struct smcd_dev *dev);
>>   int smc_ism_init(void);
>>   void smc_ism_exit(void);
>>   int smcd_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb);

\
 
 \ /
  Last update: 2023-09-30 10:42    [W:0.079 / U:2.032 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site