[PATCH V11 2/4] nvmet: add ZBD over ZNS backend support
Damien Le Moal
Damien.LeMoal at wdc.com
Thu Mar 11 05:39:02 GMT 2021
On 2021/03/11 14:29, Chaitanya Kulkarni wrote:
>>> +
>>> +bool nvmet_bdev_zns_enable(struct nvmet_ns *ns)
>>> +{
>>> + if (nvmet_bdev_has_conv_zones(ns->bdev))
>>> + return false;
>>> +
>>> + ns->blksize_shift = blksize_bits(bdev_physical_block_size(ns->bdev));
>> Shouldn't this be using logical block size ? Otherwise, zsze calculation in
>> nvmet_execute_identify_cns_cs_ns() could be wrong.
>
> Okay, will send out V12 with that fix.
>
>>> +
>>> + if (!nvmet_zns_update_zasl(ns))
>>> + return false;
>>> +
>>> + return !(get_capacity(ns->bdev->bd_disk) &
>>> + (bdev_zone_sectors(ns->bdev) - 1));
>>> +}
>>> +
>>> +void nvmet_execute_identify_cns_cs_ctrl(struct nvmet_req *req)
>>>
>>> +
>>> +void nvmet_bdev_execute_zone_mgmt_recv(struct nvmet_req *req)
>>> +{
>>> + sector_t sect = nvmet_lba_to_sect(req->ns, req->cmd->zmr.slba);
>>> + u32 bufsize = (le32_to_cpu(req->cmd->zmr.numd) + 1) << 2;
>>> + struct nvmet_report_zone_data data = { .ns = req->ns };
>>> + unsigned int nr_zones;
>>> + int reported_zones;
>>> + u16 status;
>>> +
>>> + status = nvmet_bdev_zns_checks(req);
>>> + if (status)
>>> + goto out;
>>> +
>>> + data.rz = __vmalloc(bufsize, GFP_KERNEL | __GFP_NORETRY | __GFP_ZERO);
>>> + if (!data.rz) {
>>> + status = NVME_SC_INTERNAL;
>>> + goto out;
>>> + }
>>> +
>>> + nr_zones = (bufsize - sizeof(struct nvme_zone_report)) /
>>> + sizeof(struct nvme_zone_descriptor);
>> What if nr_zones is 0 ? This should be failed.
>
> blkdev_report_zones() already handles that check. I thinkthe error condition
> below blkdev_report_zones() should include<= 0 case instead of just <.
Reporting 0 zones with a valid buffer size (nr_zones > 0) is a valid reply, not
an error. This can happen depending on reporting options. Even though
blkdev_report_zones() does not allow reporting options, it would be strange to
fail empty reports.
>
> Will send V12 with the <= fix.
>
>>
>>> + reported_zones = blkdev_report_zones(req->ns->bdev, sect, nr_zones,
>>> + nvmet_bdev_report_zone_cb, &data);
>>> + if (reported_zones < 0) {
>>> + status = NVME_SC_INTERNAL;
>>> + goto out_free_report_zones;
>>> + }
>>> +
>>> + data.rz->nr_zones = cpu_to_le64(reported_zones);
>>> +
>>> + status = nvmet_copy_to_sgl(req, 0, data.rz, bufsize);
>>> +
>>> +out_free_report_zones:
>>> + kvfree(data.rz);
>>> +out:
>>> + nvmet_req_complete(req, status);
>>> +}
>>> +
>>>
>
>
--
Damien Le Moal
Western Digital Research
More information about the Linux-nvme
mailing list