From: Steve French <smfrench@gmail.com>
To: Ronnie Sahlberg <lsahlber@redhat.com>
Cc: linux-cifs <linux-cifs@vger.kernel.org>
Subject: Re: [PATCH] cifs: improve fallocate emulation
Date: Thu, 8 Apr 2021 22:55:26 -0500 [thread overview]
Message-ID: <CAH2r5mtcXGPBZWkfCQwpFg6DgSoYL-_SfStbt17=-zmoFQxxTw@mail.gmail.com> (raw)
In-Reply-To: <20210408224054.649656-1-lsahlber@redhat.com>
tentatively merged into cifs-2.6.git for-next pending more testing
On Thu, Apr 8, 2021 at 5:41 PM Ronnie Sahlberg <lsahlber@redhat.com> wrote:
>
> RHBZ: 1866684
>
> We don't have a real fallocate in the SMB2 protocol so we used to emulate fallocate
> by simply switching the file to become non-sparse. But as that could potantially
> consume a lot more data than we intended to fallocate (large sparse file and fallocating a thin
> slice in the middle) we would only do this IFF the fallocate request was for virtually the entire file.
>
> This patch improves this and starts allowing us to fallocate smaller chunks of a file by
> overwriting the region with 0, for the parts that are unallocated.
>
> The method used is to first query the server for FSCTL_QUERY_ALLOCATED_RANGES to find what
> is unallocated in teh fallocate range and then to only overwrite-with-zero the unallocated ranges to fill
> in the holes.
> As overwriting-with-zero is different from just allocating blocks, and potentially much more expensive,
> we limit this to only allow fallocate ranges up to 1Mb in size.
>
> Reported-by: kernel test robot <lkp@intel.com>
> Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
> ---
> fs/cifs/smb2ops.c | 133 ++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 133 insertions(+)
>
> diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
> index f703204fb185..1eecaeb4beb4 100644
> --- a/fs/cifs/smb2ops.c
> +++ b/fs/cifs/smb2ops.c
> @@ -3563,6 +3563,119 @@ static long smb3_punch_hole(struct file *file, struct cifs_tcon *tcon,
> return rc;
> }
>
> +static int smb3_simple_fallocate_write_range(unsigned int xid,
> + struct cifs_tcon *tcon,
> + struct cifsFileInfo *cfile,
> + loff_t off, loff_t len,
> + char *buf)
> +{
> + struct cifs_io_parms io_parms = {0};
> + int nbytes;
> + struct kvec iov[2];
> +
> + io_parms.netfid = cfile->fid.netfid;
> + io_parms.pid = current->tgid;
> + io_parms.tcon = tcon;
> + io_parms.persistent_fid = cfile->fid.persistent_fid;
> + io_parms.volatile_fid = cfile->fid.volatile_fid;
> + io_parms.offset = off;
> + io_parms.length = len;
> +
> + /* iov[0] is reserved for smb header */
> + iov[1].iov_base = buf;
> + iov[1].iov_len = io_parms.length;
> + return SMB2_write(xid, &io_parms, &nbytes, iov, 1);
> +}
> +
> +static int smb3_simple_fallocate_range(unsigned int xid,
> + struct cifs_tcon *tcon,
> + struct cifsFileInfo *cfile,
> + loff_t off, loff_t len)
> +{
> + struct file_allocated_range_buffer in_data, *out_data = NULL, *tmp_data;
> + u32 out_data_len;
> + char *buf = NULL;
> + loff_t l;
> + int rc;
> +
> + in_data.file_offset = cpu_to_le64(off);
> + in_data.length = cpu_to_le64(len);
> + rc = SMB2_ioctl(xid, tcon, cfile->fid.persistent_fid,
> + cfile->fid.volatile_fid,
> + FSCTL_QUERY_ALLOCATED_RANGES, true,
> + (char *)&in_data, sizeof(in_data),
> + 1024 * sizeof(struct file_allocated_range_buffer),
> + (char **)&out_data, &out_data_len);
> + if (rc)
> + goto out;
> + /*
> + * It is already all allocated
> + */
> + if (out_data_len == 0)
> + goto out;
> +
> + buf = kzalloc(1024 * 1024, GFP_KERNEL);
> + if (buf == NULL) {
> + rc = -ENOMEM;
> + goto out;
> + }
> +
> + tmp_data = out_data;
> + while (len) {
> + /*
> + * The rest of the region is unmapped so write it all.
> + */
> + if (out_data_len == 0) {
> + rc = smb3_simple_fallocate_write_range(xid, tcon,
> + cfile, off, len, buf);
> + goto out;
> + }
> +
> + if (out_data_len < sizeof(struct file_allocated_range_buffer)) {
> + rc = -EINVAL;
> + goto out;
> + }
> +
> + if (off < le64_to_cpu(tmp_data->file_offset)) {
> + /*
> + * We are at a hole. Write until the end of the region
> + * or until the next allocated data,
> + * whichever comes next.
> + */
> + l = le64_to_cpu(tmp_data->file_offset) - off;
> + if (len < l)
> + l = len;
> + rc = smb3_simple_fallocate_write_range(xid, tcon,
> + cfile, off, l, buf);
> + if (rc)
> + goto out;
> + off = off + l;
> + len = len - l;
> + if (len == 0)
> + goto out;
> + }
> + /*
> + * We are at a section of allocated data, just skip forward
> + * until the end of the data or the end of the region
> + * we are supposed to fallocate, whichever comes first.
> + */
> + l = le64_to_cpu(tmp_data->length);
> + if (len < l)
> + l = len;
> + off += l;
> + len -= l;
> +
> + tmp_data = &tmp_data[1];
> + out_data_len -= sizeof(struct file_allocated_range_buffer);
> + }
> +
> + out:
> + kfree(out_data);
> + kfree(buf);
> + return rc;
> +}
> +
> +
> static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
> loff_t off, loff_t len, bool keep_size)
> {
> @@ -3623,6 +3736,26 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
> }
>
> if ((keep_size == true) || (i_size_read(inode) >= off + len)) {
> + /*
> + * At this point, we are trying to fallocate an internal
> + * regions of a sparse file. Since smb2 does not have a
> + * fallocate command we have two otions on how to emulate this.
> + * We can either turn the entire file to become non-sparse
> + * which we only do if the fallocate is for virtually
> + * the whole file, or we can overwrite the region with zeroes
> + * using SMB2_write, which could be prohibitevly expensive
> + * if len is large.
> + */
> + /*
> + * We are only trying to fallocate a small region so
> + * just write it with zero.
> + */
> + if (len <= 1024 * 1024) {
> + rc = smb3_simple_fallocate_range(xid, tcon, cfile,
> + off, len);
> + goto out;
> + }
> +
> /*
> * Check if falloc starts within first few pages of file
> * and ends within a few pages of the end of file to
> --
> 2.30.2
>
--
Thanks,
Steve
next prev parent reply other threads:[~2021-04-09 3:55 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-08 22:40 [PATCH] cifs: improve fallocate emulation Ronnie Sahlberg
2021-04-09 3:55 ` Steve French [this message]
[not found] <20210603053101.1229297-1-lsahlber@redhat.com>
2021-06-03 5:31 ` Ronnie Sahlberg
2021-06-05 20:59 ` Steve French
-- strict thread matches above, loose matches on Subject: below --
2021-04-08 7:46 Ronnie Sahlberg
2021-04-08 10:36 ` kernel test robot
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='CAH2r5mtcXGPBZWkfCQwpFg6DgSoYL-_SfStbt17=-zmoFQxxTw@mail.gmail.com' \
--to=smfrench@gmail.com \
--cc=linux-cifs@vger.kernel.org \
--cc=lsahlber@redhat.com \
/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;
as well as URLs for NNTP newsgroup(s).