From mboxrd@z Thu Jan 1 00:00:00 1970 From: minwoo.im.dev@gmail.com (Minwoo Im) Date: Thu, 25 Jul 2019 22:43:32 +0900 Subject: [PATCH 4/5] lnvm: introduce chunk-log command for chunk info In-Reply-To: References: <20190723183601.29370-1-minwoo.im.dev@gmail.com> <20190723183601.29370-5-minwoo.im.dev@gmail.com> Message-ID: <20190725134332.GC3018@minwoo-desktop> Hi Matias, On 19-07-24 14:27:10, Matias Bj?rling wrote: > On 23/07/2019 20.36, Minwoo Im wrote: > > To retrieve the chunk information from the nvme namespae for the given > > OCSSD, we can just do like: > > nvme lnvm chunk-log /dev/nvme0n1 --output-format=normal > > > > This will calculate the data length from the geometry data structure > > which might be retrieved by a Geometry command(Identity for 1.2 spec.). > > Then it will request get log page API for 1.3 NVMe spec to get the > > entries which indicate chunk information. > > > > Cc: Keith Busch > > Cc: Matias Bjorling > > Signed-off-by: Minwoo Im > > --- > > diff --git a/plugins/lnvm/lnvm-nvme.h b/plugins/lnvm/lnvm-nvme.h > > index 3d5cbc5..f091f7b 100644 > > --- a/plugins/lnvm/lnvm-nvme.h > > +++ b/plugins/lnvm/lnvm-nvme.h > > @@ -12,6 +12,7 @@ PLUGIN(NAME("lnvm", "LightNVM specific extensions"), > > ENTRY("list", "List available LightNVM devices", lnvm_list) > > ENTRY("info", "List general information and available target engines", lnvm_info) > > ENTRY("id-ns", "List geometry for LightNVM device", lnvm_id_ns) > > + ENTRY("chunk-log", "Chunk information by Get Log Page", lnvm_chunk_log) > > Chunk Information Log Page ? Sure, I'm fine with that also. > > ENTRY("init", "Initialize media manager on LightNVM device", lnvm_init) > > ENTRY("create", "Create target on top of a LightNVM device", lnvm_create_tgt) > > ENTRY("remove", "Remove target from device", lnvm_remove_tgt) > > Hi Minwoo, > > Could you squash patch 2, 3 and 4 together (4 being the main patch) - They > belong together to implement one feature. > > As a side-note, we have the same command for ZNS (that will be pushed when > the ZNS TP is ratified) - In that, we've also added support for supplying > start lba and number of chunks (zones) to return. Could you add that as well > to this? Then there is coherency between the two? Thanks for the review. Yes I will if the TP has been ratified. Also I'll prepare a new series with squashed what you just mentioned here :) Thanks!