On Thu, May 25, 2023 at 07:49:32PM +0530, Chandan Babu R wrote:
> On Tue, May 23, 2023 at 10:25:13 AM -0700, Darrick J. Wong wrote:
> > On Tue, May 23, 2023 at 02:30:35PM +0530, Chandan Babu R wrote:
> >> This commit moves functionality associated with writing metadump to disk into
> >> a new function. It also renames metadump initialization, write and release
> >> functions to reflect the fact that they work with v1 metadump files.
> >>
> >> The metadump initialization, write and release functions are now invoked via
> >> metadump_ops->init_metadump(), metadump_ops->write_metadump() and
> >> metadump_ops->release_metadump() respectively.
> >>
> >> Signed-off-by: Chandan Babu R <chandan.babu@xxxxxxxxxx>
> >> ---
> >> db/metadump.c | 124 +++++++++++++++++++++++++-------------------------
> >> 1 file changed, 61 insertions(+), 63 deletions(-)
> >>
> >> diff --git a/db/metadump.c b/db/metadump.c
> >> index 56d8c3bdf..7265f73ec 100644
> >> --- a/db/metadump.c
> >> +++ b/db/metadump.c
> >> @@ -135,59 +135,6 @@ print_progress(const char *fmt, ...)
> >> metadump.progress_since_warning = 1;
> >> }
> >>
> >> -/*
> >> - * A complete dump file will have a "zero" entry in the last index block,
> >> - * even if the dump is exactly aligned, the last index will be full of
> >> - * zeros. If the last index entry is non-zero, the dump is incomplete.
> >> - * Correspondingly, the last chunk will have a count < num_indices.
> >> - *
> >> - * Return 0 for success, -1 for failure.
> >> - */
> >> -
> >> -static int
> >> -write_index(void)
> >> -{
> >> - struct xfs_metablock *metablock = metadump.metablock;
> >> - /*
> >> - * write index block and following data blocks (streaming)
> >> - */
> >> - metablock->mb_count = cpu_to_be16(metadump.cur_index);
> >> - if (fwrite(metablock, (metadump.cur_index + 1) << BBSHIFT, 1,
> >> - metadump.outf) != 1) {
> >> - print_warning("error writing to target file");
> >> - return -1;
> >> - }
> >> -
> >> - memset(metadump.block_index, 0, metadump.num_indices * sizeof(__be64));
> >> - metadump.cur_index = 0;
> >> - return 0;
> >> -}
> >> -
> >> -/*
> >> - * Return 0 for success, -errno for failure.
> >> - */
> >> -static int
> >> -write_buf_segment(
> >> - char *data,
> >> - int64_t off,
> >> - int len)
> >> -{
> >> - int i;
> >> - int ret;
> >> -
> >> - for (i = 0; i < len; i++, off++, data += BBSIZE) {
> >> - metadump.block_index[metadump.cur_index] = cpu_to_be64(off);
> >> - memcpy(&metadump.block_buffer[metadump.cur_index << BBSHIFT],
> >> - data, BBSIZE);
> >> - if (++metadump.cur_index == metadump.num_indices) {
> >> - ret = write_index();
> >> - if (ret)
> >> - return -EIO;
> >> - }
> >> - }
> >> - return 0;
> >> -}
> >> -
> >> /*
> >> * we want to preserve the state of the metadata in the dump - whether it is
> >> * intact or corrupt, so even if the buffer has a verifier attached to it we
> >> @@ -224,15 +171,16 @@ write_buf(
> >>
> >> /* handle discontiguous buffers */
> >> if (!buf->bbmap) {
> >> - ret = write_buf_segment(buf->data, buf->bb, buf->blen);
> >> + ret = metadump.mdops->write_metadump(buf->typ->typnm, buf->data,
> >> + buf->bb, buf->blen);
> >> if (ret)
> >> return ret;
> >> } else {
> >> int len = 0;
> >> for (i = 0; i < buf->bbmap->nmaps; i++) {
> >> - ret = write_buf_segment(buf->data + BBTOB(len),
> >> - buf->bbmap->b[i].bm_bn,
> >> - buf->bbmap->b[i].bm_len);
> >> + ret = metadump.mdops->write_metadump(buf->typ->typnm,
> >> + buf->data + BBTOB(len), buf->bbmap->b[i].bm_bn,
> >> + buf->bbmap->b[i].bm_len);
> >> if (ret)
> >> return ret;
> >> len += buf->bbmap->b[i].bm_len;
> >> @@ -2994,7 +2942,7 @@ done:
> >> }
> >>
> >> static int
> >> -init_metadump(void)
> >> +init_metadump_v1(void)
> >> {
> >> metadump.metablock = (xfs_metablock_t *)calloc(BBSIZE + 1, BBSIZE);
> >> if (metadump.metablock == NULL) {
> >> @@ -3035,12 +2983,60 @@ init_metadump(void)
> >> return 0;
> >> }
> >>
> >> +static int
> >> +end_write_metadump_v1(void)
> >> +{
> >> + /*
> >> + * write index block and following data blocks (streaming)
> >> + */
> >> + metadump.metablock->mb_count = cpu_to_be16(metadump.cur_index);
> >> + if (fwrite(metadump.metablock, (metadump.cur_index + 1) << BBSHIFT, 1, metadump.outf) != 1) {
> >> + print_warning("error writing to target file");
> >> + return -1;
> >> + }
> >> +
> >> + memset(metadump.block_index, 0, metadump.num_indices * sizeof(__be64));
> >> + metadump.cur_index = 0;
> >> + return 0;
> >> +}
> >> +
> >> +static int
> >> +write_metadump_v1(
> >> + enum typnm type,
> >> + char *data,
> >> + int64_t off,
> >
> > This really ought to be an xfs_daddr_t, right?
> >
>
> Yes, you are right. I will make the required change.
>
> >> + int len)
> >> +{
> >> + int i;
> >> + int ret;
> >> +
> >> + for (i = 0; i < len; i++, off++, data += BBSIZE) {
> >> + metadump.block_index[metadump.cur_index] = cpu_to_be64(off);
> >> + memcpy(&metadump.block_buffer[metadump.cur_index << BBSHIFT],
> >> + data, BBSIZE);
> >
> > Wondering if this ought to be called ->record_segment or something, since
> > it's not really writing anything to disk, merely copying it to the index
> > buffer.
> >
> >> + if (++metadump.cur_index == metadump.num_indices) {
> >> + ret = end_write_metadump_v1();
> >> + if (ret)
> >> + return -EIO;
> >
> > This is "generic" code for "Have we filled up the index table? If so,
> > then write the index block the indexed data". Shouldn't it go in
> > write_buf? And then write_buf does something like:
> >
> > while (len > 0) {
> > segment_len = min(len, metadump.num_indices - metadump.cur_index);
> >
> > metadump.ops->record_segment(type, buf, daddr, segment_len);
> >
> > metadump.cur_index += segment_len;
> > if (metadump.cur_index == metadump.num_indices) {
> > metadump.ops->write_index(...);
> > metadump.cur_index = 0;
> > }
> >
> > len -= segment_len;
> > daddr += segment_len;
> > buf += (segment_len << 9);
> > }
> >
> > if (metadump.cur_index)
> > metadump.ops->write_index(...);
> > metadump.cur_index = 0;
> >
>
> The above change would require write_buf() to know about the internals of
> metadump v1 format. This change can be performed as long as the metadump
> supports only the v1 format. However, supporting the v2 format later requires
> that write_buf() invoke version specific functions via function pointers and
> to not perform any other version specific operation (e.g. checking to see if
> the v1 format's in-memory index is filled) on its own.
Ah, right. I made that comment before you pointed out that v2
interleaves meta_extent records with dump data instead of having
separate index blocks. Question withdrawn.
--D
> >> + }
> >> + }
> >> +
> >> + return 0;
> >> +}
> >> +
> >> static void
> >> -release_metadump(void)
> >> +release_metadump_v1(void)
> >> {
> >> free(metadump.metablock);
> >> }
> >>
> >> +static struct metadump_ops metadump1_ops = {
> >> + .init_metadump = init_metadump_v1,
> >> + .write_metadump = write_metadump_v1,
> >> + .end_write_metadump = end_write_metadump_v1,
> >> + .release_metadump = release_metadump_v1,
> >> +};
> >> +
> >> static int
> >> metadump_f(
> >> int argc,
> >> @@ -3182,9 +3178,11 @@ metadump_f(
> >> }
> >> }
> >>
> >> - ret = init_metadump();
> >> + metadump.mdops = &metadump1_ops;
> >> +
> >> + ret = metadump.mdops->init_metadump();
> >> if (ret)
> >> - return 0;
> >> + goto out;
> >>
> >> exitcode = 0;
> >>
> >> @@ -3206,7 +3204,7 @@ metadump_f(
> >>
> >> /* write the remaining index */
> >> if (!exitcode)
> >> - exitcode = write_index() < 0;
> >> + exitcode = metadump.mdops->end_write_metadump() < 0;
> >>
> >> if (metadump.progress_since_warning)
> >> fputc('\n', metadump.stdout_metadump ? stderr : stdout);
> >> @@ -3225,7 +3223,7 @@ metadump_f(
> >> while (iocur_sp > start_iocur_sp)
> >> pop_cur();
> >>
> >> - release_metadump();
> >> + metadump.mdops->release_metadump();
> >>
> >> out:
> >> return 0;
> >> --
> >> 2.39.1
> >>
>
>
> --
> chandan
