public inbox for linux-btrfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Anand Jain <anand.jain@oracle.com>
To: Josef Bacik <josef@toxicpanda.com>
Cc: linux-btrfs@vger.kernel.org
Subject: Re: [PATCH v2 1/3] btrfs: add readmirror type framework
Date: Fri, 3 Jan 2020 17:57:16 +0800	[thread overview]
Message-ID: <4abd93cc-eb99-d977-ba48-2754b15390bd@oracle.com> (raw)
In-Reply-To: <e6585b93-a585-15e1-ea2a-de6279c317a8@toxicpanda.com>



On 3/1/20 12:24 AM, Josef Bacik wrote:
> On 1/2/20 5:12 AM, Anand Jain wrote:
>> As of now we use %pid method to read stripped mirrored data. So
>> application's process id determines the stripe id to be read. This type
>> of read IO routing typically helps in a system with many small
>> independent applications tying to read random data. On the other hand
>> the %pid based read IO distribution policy is inefficient if there is a
>> single application trying to read large data and the overall disk
>> bandwidth remains under utilized.
>>
>> So this patch introduces a framework where we could add more readmirror
>> policies, such as routing the IO based on device's wait-queue or manual
>> when we have a read-preferred device or a policy based on the target
>> storage caching.
>>
>> Signed-off-by: Anand Jain <anand.jain@oracle.com>
>> ---
>> v2: Declare fs_devices::readmirror as u8 instead of atomic_t
>>      A small change in comment and change log wordings.
>>
>>   fs/btrfs/volumes.c | 16 +++++++++++++++-
>>   fs/btrfs/volumes.h |  8 ++++++++
>>   2 files changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>> index c95e47aa84f8..e26af766f2b9 100644
>> --- a/fs/btrfs/volumes.c
>> +++ b/fs/btrfs/volumes.c
>> @@ -1162,6 +1162,8 @@ static int open_fs_devices(struct 
>> btrfs_fs_devices *fs_devices,
>>       fs_devices->opened = 1;
>>       fs_devices->latest_bdev = latest_dev->bdev;
>>       fs_devices->total_rw_bytes = 0;
>> +    /* Set the default readmirror policy */
>> +    fs_devices->readmirror = BTRFS_READMIRROR_DEFAULT;
>>   out:
>>       return ret;
>>   }
>> @@ -5300,7 +5302,19 @@ static int find_live_mirror(struct 
>> btrfs_fs_info *fs_info,
>>       else
>>           num_stripes = map->num_stripes;
>> -    preferred_mirror = first + current->pid % num_stripes;
>> +    switch (fs_info->fs_devices->readmirror) {
>> +    case BTRFS_READMIRROR_BY_PID:
>> +        preferred_mirror = first + current->pid % num_stripes;
>> +        break;
>> +    default:
>> +        /*
>> +         * Shouln't happen, just warn and use by_pid instead of failing.
>> +         */
>> +        btrfs_warn_rl(fs_info,
>> +                  "unknown readmirror type %u, fallback to by_pid",
>> +                  fs_info->fs_devices->readmirror);
>> +        preferred_mirror = first + current->pid % num_stripes;
>> +    }
>>       if (dev_replace_is_ongoing &&
>>           fs_info->dev_replace.cont_reading_from_srcdev_mode ==
>> diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
>> index 68021d1ee216..f5f091f3c72b 100644
>> --- a/fs/btrfs/volumes.h
>> +++ b/fs/btrfs/volumes.h
>> @@ -209,6 +209,12 @@ struct btrfs_device {
>>   BTRFS_DEVICE_GETSET_FUNCS(disk_total_bytes);
>>   BTRFS_DEVICE_GETSET_FUNCS(bytes_used);
>> +/* readmirror_policy types */
>> +#define BTRFS_READMIRROR_DEFAULT    BTRFS_READMIRROR_BY_PID
>> +enum btrfs_readmirror_policy_type {
>> +    BTRFS_READMIRROR_BY_PID,
>> +};
>> +
>>   struct btrfs_fs_devices {
>>       u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
>>       u8 metadata_uuid[BTRFS_FSID_SIZE];
>> @@ -260,6 +266,8 @@ struct btrfs_fs_devices {
>>       struct kobject *devices_kobj;
>>       struct kobject *devinfo_kobj;
>>       struct completion kobj_unregister;
>> +
>> +    u8 readmirror;
> 
> The only valid values for this are the enum, so make this
> 
> enum btrfs_readmirror_policy_type readmirror;

  Oh. Ok.

Thanks, Anand


> Thanks,
> 
> Josef

  reply	other threads:[~2020-01-03  9:57 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-02 10:12 [PATCH v2 0/3] readmirror feature (sysfs and in-memory only approach) Anand Jain
2020-01-02 10:12 ` [PATCH v2 1/3] btrfs: add readmirror type framework Anand Jain
2020-01-02 16:24   ` Josef Bacik
2020-01-03  9:57     ` Anand Jain [this message]
2020-01-02 19:32   ` Steven Davies
2020-01-03 10:28     ` Anand Jain
2020-01-03 14:51       ` Steven Davies
2020-01-03 10:31   ` [PATCH v3 " Anand Jain
2020-01-02 10:12 ` [PATCH v2 2/3] btrfs: sysfs, add readmirror kobject Anand Jain
2020-01-02 10:12 ` [PATCH v2 3/3] btrfs: sysfs, create by_pid readmirror attribute Anand Jain

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4abd93cc-eb99-d977-ba48-2754b15390bd@oracle.com \
    --to=anand.jain@oracle.com \
    --cc=josef@toxicpanda.com \
    --cc=linux-btrfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox