Message ID | 20200430101918.30351-1-dplotnikov@virtuozzo.com |
---|---|
Headers | show |
Series | implement zstd cluster compression method | expand |
On 30.04.20 12:19, Denis Plotnikov wrote: > v23: > Undecided: whether to add zstd(zlib) compression > details to the qcow2 spec > 03: tighten assertion on zstd decompression [Eric] > 04: use _rm_test_img appropriately [Max] Thanks, applied to my block branch: https://git.xanclic.moe/XanClic/qemu/commits/branch/block Max
On 05.05.20 12:26, Max Reitz wrote: > On 30.04.20 12:19, Denis Plotnikov wrote: >> v23: >> Undecided: whether to add zstd(zlib) compression >> details to the qcow2 spec >> 03: tighten assertion on zstd decompression [Eric] >> 04: use _rm_test_img appropriately [Max] > > Thanks, applied to my block branch: I’m afraid I have to unqueue this series again, because it makes many iotests fail due to an additional “compression_type=zlib” output when images are created, an additional “compression type” line in qemu-img info output where format-specific information is not suppressed, and an additional line in qemu-img create -f qcow2 -o help. Max
On 05.05.2020 15:03, Max Reitz wrote: > On 05.05.20 12:26, Max Reitz wrote: >> On 30.04.20 12:19, Denis Plotnikov wrote: >>> v23: >>> Undecided: whether to add zstd(zlib) compression >>> details to the qcow2 spec >>> 03: tighten assertion on zstd decompression [Eric] >>> 04: use _rm_test_img appropriately [Max] >> Thanks, applied to my block branch: > I’m afraid I have to unqueue this series again, because it makes many > iotests fail due to an additional “compression_type=zlib” output when > images are created, an additional “compression type” line in > qemu-img info output where format-specific information is not > suppressed, and an additional line in qemu-img create -f qcow2 -o help. > > Max > Hmm, this is strange. I made some modifications for the tests in 0001 of the series (qcow2: introduce compression type feature). Among the other test related changes, the patch contains the hunk: +++ b/tests/qemu-iotests/common.filter @@ -152,7 +152,8 @@ _filter_img_create() -e "s# refcount_bits=[0-9]\\+##g" \ -e "s# key-secret=[a-zA-Z0-9]\\+##g" \ -e "s# iter-time=[0-9]\\+##g" \ - -e "s# force_size=\\(on\\|off\\)##g" + -e "s# force_size=\\(on\\|off\\)##g" \ + -e "s# compression_type=[a-zA-Z0-9]\\+##g" } which has to filter out "compression_type" on image creation. But you say that you can see the "compression_type" on the image creation. May be the patch wasn't fully applied? Or the test related modification were omitted? I've just re-based the series on top of: 681b07f4e76dbb700c16918d (vanilla/master, mainstream) Merge: a2261b2754 714eb0dbc5 Author: Peter Maydell <peter.maydell@linaro.org> Date: Tue May 5 15:47:44 2020 +0100 and run the tests with 'make check-block' and got the following: Not run: 071 099 184 220 259 267 Some cases not run in: 030 040 041 Passed all 113 iotests May be I do something wrong? Denis
06.05.2020 11:01, Denis Plotnikov wrote: > > > On 05.05.2020 15:03, Max Reitz wrote: >> On 05.05.20 12:26, Max Reitz wrote: >>> On 30.04.20 12:19, Denis Plotnikov wrote: >>>> v23: >>>> Undecided: whether to add zstd(zlib) compression >>>> details to the qcow2 spec >>>> 03: tighten assertion on zstd decompression [Eric] >>>> 04: use _rm_test_img appropriately [Max] >>> Thanks, applied to my block branch: >> I’m afraid I have to unqueue this series again, because it makes many >> iotests fail due to an additional “compression_type=zlib” output when >> images are created, an additional “compression type” line in >> qemu-img info output where format-specific information is not >> suppressed, and an additional line in qemu-img create -f qcow2 -o help. >> >> Max >> > > Hmm, this is strange. I made some modifications for the tests > in 0001 of the series (qcow2: introduce compression type feature). > > Among the other test related changes, the patch contains the hunk: > > +++ b/tests/qemu-iotests/common.filter > @@ -152,7 +152,8 @@ _filter_img_create() > -e "s# refcount_bits=[0-9]\\+##g" \ > -e "s# key-secret=[a-zA-Z0-9]\\+##g" \ > -e "s# iter-time=[0-9]\\+##g" \ > - -e "s# force_size=\\(on\\|off\\)##g" > + -e "s# force_size=\\(on\\|off\\)##g" \ > + -e "s# compression_type=[a-zA-Z0-9]\\+##g" > } > > which has to filter out "compression_type" on image creation. > > But you say that you can see the "compression_type" on the image creation. > May be the patch wasn't fully applied? Or the test related modification were omitted? > > I've just re-based the series on top of: > > 681b07f4e76dbb700c16918d (vanilla/master, mainstream) > Merge: a2261b2754 714eb0dbc5 > Author: Peter Maydell <peter.maydell@linaro.org> > Date: Tue May 5 15:47:44 2020 +0100 > > and run the tests with 'make check-block' > > and got the following: > > Not run: 071 099 184 220 259 267 > Some cases not run in: 030 040 041 > Passed all 113 iotests > Hmm, we definitely have a lot more iotests. So, I assume make check-block doesn't run all of them. To run all: cd tests/qemu-iotests ./check -qcow2 ./check -raw ... and so for any format you want to test I also recommend to do export TEST_DIR=/something before running tests, where something is tmpfs or ssd, which will decrease testing time a lot. Still, some (very small subset) of tests doesn't run on tmpfs, you can rerun them with TEST_DIR=/normal/hdd/directory