| Message ID | 20160707084249.29084-1-fullmanet@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, 07/07 10:42, Reda Sallahi wrote: > Commit "cdeaf1f vmdk: add bdrv_co_write_zeroes" causes a regression on > writes. It writes metadata after every write instead of doing it only once > for each cluster. > > vmdk_pwritev() writes metadata whenever m_data is set as valid so this patch > sets m_data as valid only when we have a new cluster which hasn't been > allocated before or a zero grain. > > Signed-off-by: Reda Sallahi <fullmanet@gmail.com> > --- > v2: Corrected the commit id referenced in the commit message. > > block/vmdk.c | 14 +++++++------- > 1 file changed, 7 insertions(+), 7 deletions(-) > > diff --git a/block/vmdk.c b/block/vmdk.c > index d73f431..1cbd487 100644 > --- a/block/vmdk.c > +++ b/block/vmdk.c > @@ -1202,13 +1202,6 @@ static int get_cluster_offset(BlockDriverState *bs, > l2_index = ((offset >> 9) / extent->cluster_sectors) % extent->l2_size; > cluster_sector = le32_to_cpu(l2_table[l2_index]); > > - if (m_data) { > - m_data->valid = 1; > - m_data->l1_index = l1_index; > - m_data->l2_index = l2_index; > - m_data->l2_offset = l2_offset; > - m_data->l2_cache_entry = &l2_table[l2_index]; > - } > if (extent->has_zero_grain && cluster_sector == VMDK_GTE_ZEROED) { > zeroed = true; > } > @@ -1231,6 +1224,13 @@ static int get_cluster_offset(BlockDriverState *bs, > if (ret) { > return ret; > } > + if (m_data) { > + m_data->valid = 1; > + m_data->l1_index = l1_index; > + m_data->l2_index = l2_index; > + m_data->l2_offset = l2_offset; > + m_data->l2_cache_entry = &l2_table[l2_index]; > + } > } > *cluster_offset = cluster_sector << BDRV_SECTOR_BITS; > return VMDK_OK; > -- > 2.9.0 > Looks good, nice catch! Reviewed-by: Fam Zheng <famz@redhat.com>
On 07.07.2016 10:42, Reda Sallahi wrote: > Commit "cdeaf1f vmdk: add bdrv_co_write_zeroes" causes a regression on > writes. It writes metadata after every write instead of doing it only once > for each cluster. > > vmdk_pwritev() writes metadata whenever m_data is set as valid so this patch > sets m_data as valid only when we have a new cluster which hasn't been > allocated before or a zero grain. > > Signed-off-by: Reda Sallahi <fullmanet@gmail.com> > --- > v2: Corrected the commit id referenced in the commit message. > > block/vmdk.c | 14 +++++++------- > 1 file changed, 7 insertions(+), 7 deletions(-) Thanks Reda, I've applied the patch to my block branch: https://github.com/XanClic/qemu/commits/block Max
diff --git a/block/vmdk.c b/block/vmdk.c index d73f431..1cbd487 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -1202,13 +1202,6 @@ static int get_cluster_offset(BlockDriverState *bs, l2_index = ((offset >> 9) / extent->cluster_sectors) % extent->l2_size; cluster_sector = le32_to_cpu(l2_table[l2_index]); - if (m_data) { - m_data->valid = 1; - m_data->l1_index = l1_index; - m_data->l2_index = l2_index; - m_data->l2_offset = l2_offset; - m_data->l2_cache_entry = &l2_table[l2_index]; - } if (extent->has_zero_grain && cluster_sector == VMDK_GTE_ZEROED) { zeroed = true; } @@ -1231,6 +1224,13 @@ static int get_cluster_offset(BlockDriverState *bs, if (ret) { return ret; } + if (m_data) { + m_data->valid = 1; + m_data->l1_index = l1_index; + m_data->l2_index = l2_index; + m_data->l2_offset = l2_offset; + m_data->l2_cache_entry = &l2_table[l2_index]; + } } *cluster_offset = cluster_sector << BDRV_SECTOR_BITS; return VMDK_OK;
Commit "cdeaf1f vmdk: add bdrv_co_write_zeroes" causes a regression on writes. It writes metadata after every write instead of doing it only once for each cluster. vmdk_pwritev() writes metadata whenever m_data is set as valid so this patch sets m_data as valid only when we have a new cluster which hasn't been allocated before or a zero grain. Signed-off-by: Reda Sallahi <fullmanet@gmail.com> --- v2: Corrected the commit id referenced in the commit message. block/vmdk.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)