lkml.org 
[lkml]   [2021]   [Jun]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v4 for-next 3/3] IB/core: Obtain subnet_prefix from cache in IB devices
From
Date
On 6/16/2021 1:12 PM, Anand Khoje wrote:
> On 6/16/2021 12:57 PM, Leon Romanovsky wrote:
>> On Wed, Jun 16, 2021 at 12:22:13PM +0530, Anand Khoje wrote:
>>> ib_query_port() calls device->ops.query_port() to get the port
>>> attributes. The method of querying is device driver specific.
>>> The same function calls device->ops.query_gid() to get the GID and
>>> extract the subnet_prefix (gid_prefix).
>>>
>>> The GID and subnet_prefix are stored in a cache. But they do not get
>>> read from the cache if the device is an Infiniband device. The
>>> following change takes advantage of the cached subnet_prefix.
>>> Testing with RDBMS has shown a significant improvement in performance
>>> with this change.
>>>
>>> The function ib_cache_is_initialised() is introduced because
>>> ib_query_port() gets called early in the stage when the cache is not
>>> built while reading port immutable property.
>>>
>>> In that case, the default GID still gets read from HCA for IB link-
>>> layer devices.
>>>
>>> In the situation of an event causing cache update, the subnet_prefix
>>> will get retrieved from newly updated GID cache in ib_cache_update(),
>>> so that we do not end up reading a stale value from cache via
>>> ib_query_port().
>>>
>>> Fixes: fad61ad ("IB/core: Add subnet prefix to port info")
>>> Suggested-by: Leon Romanovsky <leonro@nvidia.com>
>>> Suggested-by: Aru Kolappan <aru.kolappan@oracle.com>
>>> Signed-off-by: Anand Khoje <anand.a.khoje@oracle.com>
>>> Signed-off-by: Haakon Bugge <haakon.bugge@oracle.com>
>>> ---
>>>
>>> v1 -> v2:
>>>      -   Split the v1 patch in 3 patches as per Leon's suggestion.
>>>
>>> v2 -> v3:
>>>      -   Added changes as per Mark Zhang's suggestion of clearing
>>>          flags in git_table_cleanup_one().
>>> v3 -> v4:
>>>      -   Removed the enum ib_port_data_flags and 8 byte flags from
>>>          struct ib_port_data, and the set_bit()/clear_bit() API
>>>          used to update this flag as that was not necessary.
>>>          Done to keep the code simple.
>>>      -   Added code to read subnet_prefix from updated GID cache in the
>>>          event of cache update. Prior to this change, ib_cache_update
>>>          was reading the value for subnet_prefix via ib_query_port(),
>>>          due to this patch, we ended up reading a stale cached value of
>>>          subnet_prefix.
>>>
>>> ---
>>>   drivers/infiniband/core/cache.c  | 18 +++++++++++++++---
>>>   drivers/infiniband/core/device.c |  9 +++++++++
>>>   include/rdma/ib_cache.h          |  5 +++++
>>>   include/rdma/ib_verbs.h          |  1 +
>>>   4 files changed, 30 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/infiniband/core/cache.c
>>> b/drivers/infiniband/core/cache.c
>>> index 2325171..cd99c46 100644
>>> --- a/drivers/infiniband/core/cache.c
>>> +++ b/drivers/infiniband/core/cache.c
>>> @@ -917,9 +917,11 @@ static void gid_table_cleanup_one(struct
>>> ib_device *ib_dev)
>>>   {
>>>       u32 p;
>>> -    rdma_for_each_port (ib_dev, p)
>>> +    rdma_for_each_port (ib_dev, p) {
>>> +        ib_dev->port_data[p].cache_is_initialized = 0;
>>
>> I think that this line is not needed, we are removing device anyway and
>> and query_port is not allowed at this stage.
>>
> We have kept this for code completeness purposes. Just as we did with
> set_bit() and clear_bit() APIs.
>
>>>           cleanup_gid_table_port(ib_dev, p,
>>>                          ib_dev->port_data[p].cache.gid);
>>> +    }
>>>   }
>>>   static int gid_table_setup_one(struct ib_device *ib_dev)
>>> @@ -1466,6 +1468,7 @@ static int config_non_roce_gid_cache(struct
>>> ib_device *device,
>>>       struct ib_port_attr       *tprops = NULL;
>>>       struct ib_pkey_cache      *pkey_cache = NULL;
>>>       struct ib_pkey_cache      *old_pkey_cache = NULL;
>>> +    union ib_gid               gid;
>>>       int                        i;
>>>       int                        ret;
>>> @@ -1523,13 +1526,21 @@ static int config_non_roce_gid_cache(struct
>>> ib_device *device,
>>>       device->port_data[port].cache.lmc = tprops->lmc;
>>>       device->port_data[port].cache.port_state = tprops->state;
>>> -    device->port_data[port].cache.subnet_prefix =
>>> tprops->subnet_prefix;
>>> +    ret = rdma_query_gid(device, port, 0, &gid);
>>> +    if (ret) {
>>> +        write_unlock_irq(&device->cache.lock);
>>> +        goto err;
>>> +    }
>>> +
>>> +    device->port_data[port].cache.subnet_prefix =
>>> +            be64_to_cpu(gid.global.subnet_prefix);
>>> +
>>>       write_unlock_irq(&device->cache_lock);
>>>       if (enforce_security)
>>>           ib_security_cache_change(device,
>>>                        port,
>>> -                     tprops->subnet_prefix);
>>> +                     be64_to_cpu(gid.global.subnet_prefix));
>>>       kfree(old_pkey_cache);
>>>       kfree(tprops);
>>> @@ -1629,6 +1640,7 @@ int ib_cache_setup_one(struct ib_device *device)
>>>           err = ib_cache_update(device, p, true, true, true);
>>>           if (err)
>>>               return err;
>>> +        device->port_data[p].cache_is_initialized = 1;
>>>       }
>>>       return 0;
>>> diff --git a/drivers/infiniband/core/device.c
>>> b/drivers/infiniband/core/device.c
>>> index 7a617e4..57b9039 100644
>>> --- a/drivers/infiniband/core/device.c
>>> +++ b/drivers/infiniband/core/device.c
>>> @@ -2057,6 +2057,15 @@ static int __ib_query_port(struct ib_device
>>> *device,
>>>           IB_LINK_LAYER_INFINIBAND)
>>>           return 0;
>>> +    if (!ib_cache_is_initialised(device, port_num))
>>> +        goto query_gid_from_device;
>>
>> IMHO, we don't need this new function and can access ib_port_data
>> directly. In device.c, we have plenty of places that does it.
>>
>> Not critical.
>>
> Added this function to have a way to check validity of cache, such that
> it could be used in future for the same check in areas to which
> ib_port_data is opaque.
>
Also, there was a review comment from Mark Zhang for patch version 2,
where he had suggested to shift clear_bit(IB_PORT_CACHE_INITIALIZED,
&device->port_data[p].flags);
to gid_table_cleanup_one().

Thanks,
Anand
>>> +
>>> +    ib_get_cached_subnet_prefix(device, port_num,
>>> +                    &port_attr->subnet_prefix);
>>> +
>>> +    return 0;
>>> +
>>> +query_gid_from_device:
>>>       err = device->ops.query_gid(device, port_num, 0, &gid);
>>>       if (err)
>>>           return err;
>>> diff --git a/include/rdma/ib_cache.h b/include/rdma/ib_cache.h
>>> index 226ae37..46b43a7 100644
>>> --- a/include/rdma/ib_cache.h
>>> +++ b/include/rdma/ib_cache.h
>>> @@ -114,4 +114,9 @@ ssize_t rdma_query_gid_table(struct ib_device
>>> *device,
>>>                    struct ib_uverbs_gid_entry *entries,
>>>                    size_t max_entries);
>>> +static inline bool ib_cache_is_initialised(struct ib_device *device,
>>> +                    u32 port_num)
>>> +{
>>> +    return device->port_data[port_num].cache_is_initialized;
>>> +}
>>>   #endif /* _IB_CACHE_H */
>>> diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
>>> index c96d601..405f7da 100644
>>> --- a/include/rdma/ib_verbs.h
>>> +++ b/include/rdma/ib_verbs.h
>>> @@ -2177,6 +2177,7 @@ struct ib_port_data {
>>>       spinlock_t netdev_lock;
>>> +    u8 cache_is_initialized:1;
>>>       struct list_head pkey_list;
>>>       struct ib_port_cache cache;
>>> --
>>> 1.8.3.1
>>>
>

\
 
 \ /
  Last update: 2021-06-16 10:07    [W:2.196 / U:1.904 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site