[PATCH v3] nvme_core: scan namespaces asynchronously

stuart hayes stuart.w.hayes at gmail.com
Mon Jul 15 18:23:29 PDT 2024



On 7/15/2024 5:28 PM, Sagi Grimberg wrote:
> 
> 
> On 15/07/2024 23:34, Stuart Hayes wrote:
>> Use async function calls to make namespace scanning happen in parallel.
>>
>> Without the patch, NVME namespaces are scanned serially, so it can take
>> a long time for all of a controller's namespaces to become available,
>> especially with a slower (TCP) interface with large number of
>> namespaces.
>>
>> It is not uncommon to have large numbers (hundreds or thousands) of
>> namespaces on nvme-of with storage servers.
>>
>> The time it took for all namespaces to show up after connecting (via
>> TCP) to a controller with 1002 namespaces was measured on one system:
>>
>> network latency   without patch   with patch
>>       0                 6s            1s
>>      50ms             210s           10s
>>     100ms             417s           18s
>>
>> Measurements taken on another system show the effect of the patch on the
>> time nvme_scan_work() took to complete, when connecting to a linux
>> nvme-of target with varying numbers of namespaces, on a network of
>> 400us.
>>
>> namespaces    without patch   with patch
>>       1            16ms           14ms
>>       2            24ms           16ms
>>       4            49ms           22ms
>>       8           101ms           33ms
>>      16           207ms           56ms
>>     100           1.4s           0.6s
>>    1000          12.9s           2.0s
> 
> Not sure how common is the 1000 namespaces use-case, but the dozens of namespaces
> seems compelling enough.
> 
>>
>> On the same system, connecting to a local PCIe NVMe drive (a Samsung
>> PM1733) instead of a network target:
>>
>> namespaces    without patch   with patch
>>       1            13ms           12ms
>>       2            41ms           13ms
>>
>> Signed-off-by: Stuart Hayes <stuart.w.hayes at gmail.com>
>> ---
>> changes from V2:
>>    * make a separate function nvme_scan_ns_async() that calls
>>      nvme_scan_ns(), instead of modifying nvme_scan_ns()
>>    * only scan asynchronously from nvme_scan_ns_list(), not from
>>      nvme_scan_ns_sequential()
>>    * provide more timing data in the commit message
>>
>> changes from V1:
>>    * remove module param to enable/disable async scanning
>>    * add scan time measurements to commit message
>>
>>
>>   drivers/nvme/host/core.c | 48 +++++++++++++++++++++++++++++++++-------
>>   1 file changed, 40 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
>> index 782090ce0bc1..dbf05cfea063 100644
>> --- a/drivers/nvme/host/core.c
>> +++ b/drivers/nvme/host/core.c
>> @@ -4,6 +4,7 @@
>>    * Copyright (c) 2011-2014, Intel Corporation.
>>    */
>> +#include <linux/async.h>
>>   #include <linux/blkdev.h>
>>   #include <linux/blk-mq.h>
>>   #include <linux/blk-integrity.h>
>> @@ -3952,6 +3953,30 @@ static void nvme_scan_ns(struct nvme_ctrl *ctrl, unsigned nsid)
>>       }
>>   }
>> +/*
>> + * struct async_scan_info - keeps track of controller & NSIDs to scan
>> + * @ctrl:    Controller on which namespaces are being scanned
>> + * @next_idx:    Index of next NSID to scan in ns_list
>> + * @ns_list:    Pointer to list of NSIDs to scan
>> + */
>> +struct async_scan_info {
>> +    struct nvme_ctrl *ctrl;
>> +    atomic_t next_idx;
> 
> next_nsid ?
> 

OK!

>> +    __le32 *ns_list;
>> +};
>> +
>> +static void nvme_scan_ns_async(void *data, async_cookie_t cookie)
>> +{
>> +    struct async_scan_info *scan_info = data;
>> +    int idx;
>> +    u32 nsid;
>> +
>> +    idx = (u32)atomic_fetch_add(1, &scan_info->next_idx);
>> +    nsid = le32_to_cpu(scan_info->ns_list[idx]);
>> +
>> +    nvme_scan_ns(scan_info->ctrl, nsid);
>> +}
>> +
>>   static void nvme_remove_invalid_namespaces(struct nvme_ctrl *ctrl,
>>                       unsigned nsid)
>>   {
>> @@ -3975,12 +4000,14 @@ static void nvme_remove_invalid_namespaces(struct nvme_ctrl *ctrl,
>>   static int nvme_scan_ns_list(struct nvme_ctrl *ctrl)
>>   {
>>       const int nr_entries = NVME_IDENTIFY_DATA_SIZE / sizeof(__le32);
>> -    __le32 *ns_list;
>> +    struct async_scan_info scan_info;
> 
> What initializes next_idx?

See below--there's an atomic_set().  It is inside of the outer "for" loop because there can
be multiple lists that have to be scanned and it has to reset to 0 each time.

> 
>>       u32 prev = 0;
>>       int ret = 0, i;
>> +    ASYNC_DOMAIN(domain);
>> -    ns_list = kzalloc(NVME_IDENTIFY_DATA_SIZE, GFP_KERNEL);
>> -    if (!ns_list)
>> +    scan_info.ctrl = ctrl;
>> +    scan_info.ns_list = kzalloc(NVME_IDENTIFY_DATA_SIZE, GFP_KERNEL);
>> +    if (!scan_info.ns_list)
>>           return -ENOMEM;
> 
> I think you can leave the local variable ns_list as is, and just assign it to scan_info
> after, its common practice to allocate to a local pointer and use it to init a struct member.
> 
> Plus it will make the patch diff simpler.
> 

No problem, I agree.  I think someone suggested the opposite last time I submitted this.  :)

>>       for (;;) {
>> @@ -3990,28 +4017,33 @@ static int nvme_scan_ns_list(struct nvme_ctrl *ctrl)
>>               .identify.nsid        = cpu_to_le32(prev),
>>           };
>> -        ret = nvme_submit_sync_cmd(ctrl->admin_q, &cmd, ns_list,
>> -                        NVME_IDENTIFY_DATA_SIZE);
>> +        ret = nvme_submit_sync_cmd(ctrl->admin_q, &cmd,
>> +                       scan_info.ns_list,
>> +                       NVME_IDENTIFY_DATA_SIZE);
>>           if (ret) {
>>               dev_warn(ctrl->device,
>>                   "Identify NS List failed (status=0x%x)\n", ret);
>>               goto free;
>>           }
>> +        atomic_set(&scan_info.next_idx, 0);

This atomic_set is what initializes next_idx.

>>           for (i = 0; i < nr_entries; i++) {
>> -            u32 nsid = le32_to_cpu(ns_list[i]);
>> +            u32 nsid = le32_to_cpu(scan_info.ns_list[i]);
>>               if (!nsid)    /* end of the list? */
>>                   goto out;
>> -            nvme_scan_ns(ctrl, nsid);
>> +            async_schedule_domain(nvme_scan_ns_async, &scan_info,
>> +                        &domain);
>>               while (++prev < nsid)
>>                   nvme_ns_remove_by_nsid(ctrl, prev);
>>           }
>> +        async_synchronize_full_domain(&domain);
>>       }
>>    out:
>>       nvme_remove_invalid_namespaces(ctrl, prev);
>>    free:
>> -    kfree(ns_list);
>> +    async_synchronize_full_domain(&domain);
>> +    kfree(scan_info.ns_list);
>>       return ret;
>>   }
> 

Thank you for the feedback!



More information about the Linux-nvme mailing list