From: alison.schofield@intel.com
To: Vishal Verma <vishal.l.verma@intel.com>
Cc: Alison Schofield <alison.schofield@intel.com>,
nvdimm@lists.linux.dev, linux-cxl@vger.kernel.org
Subject: [ndctl PATCH] cxl/test: replace a bad root decoder usage in cxl-xor-region.sh
Date: Tue, 21 Nov 2023 18:57:53 -0800 [thread overview]
Message-ID: <20231122025753.1209527-1-alison.schofield@intel.com> (raw)
From: Alison Schofield <alison.schofield@intel.com>
The 4-way XOR region as defined in this test uses a root decoder that
is created using an improperly defined CFMWS. The problem with the
CFMWS is that Host Bridges repeat in the target list like this:
{ 0, 1, 0, 1 }.
Stop using that root decoder and create a 4-way XOR region using an
x2 root decoder that supports XOR arithmetic.
The test passes prior to this patch but there is an interleave check [1]
introduced in the CXL region driver that will expose the bad interleave
this test creates via dev_dbg() messages like this:
[ ] cxl_core:cxl_region_attach:1808: cxl decoder17.0: Test cxl_calc_interleave_pos(): fail test_pos:4 cxled->pos:2
[ ] cxl_core:cxl_region_attach:1808: cxl decoder18.0: Test cxl_calc_interleave_pos(): fail test_pos:5 cxled->pos:3
Note that the CFMWS's are defined in the kernel cxl-test module, so a
kernel patch removing the bad CFMWS will also need to be merged, but
that cleanup can follow this patch. Also note that the bad CFMWS is not
used in the default cxl-test environment. It is only visible when the
cxl-test module is loaded using the param interleave_arithmetic=1. It is
a special config that provides the XOR math CFMWS's for this test.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=a3e00c964fb943934af916f48f0dd43b5110c866
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
---
Vishal - I'm hoping you will merge this in ndctl v79 even though the
exposure with the kernel doesn't happen until kernel 6.7. This way
users of cxl-test are not learning to ignore the interleave calc
warnings.
Also, hopefully I have not introduced any new shell issues, but
I know this unit test has existing warnings. Can we do a shell
cleanup in a follow-on patchset across the CXL tests?
(and not last minute for your ndctl release)
test/cxl-xor-region.sh | 14 +++++---------
1 file changed, 5 insertions(+), 9 deletions(-)
diff --git a/test/cxl-xor-region.sh b/test/cxl-xor-region.sh
index 1962327bd00a..117e7a4bba61 100644
--- a/test/cxl-xor-region.sh
+++ b/test/cxl-xor-region.sh
@@ -68,10 +68,10 @@ setup_x2()
setup_x4()
{
- # find x4 decoder
+ # find an x2 decoder
decoder=$($CXL list -b cxl_test -D -d root | jq -r ".[] |
select(.pmem_capable == true) |
- select(.nr_targets == 4) |
+ select(.nr_targets == 2) |
.decoder")
# Find a memdev for each host-bridge interleave position
@@ -79,14 +79,10 @@ setup_x4()
.targets | .[] | select(.position == 0) | .target")
port_dev1=$($CXL list -T -d $decoder | jq -r ".[] |
.targets | .[] | select(.position == 1) | .target")
- port_dev2=$($CXL list -T -d $decoder | jq -r ".[] |
- .targets | .[] | select(.position == 2) | .target")
- port_dev3=$($CXL list -T -d $decoder | jq -r ".[] |
- .targets | .[] | select(.position == 3) | .target")
mem0=$($CXL list -M -p $port_dev0 | jq -r ".[0].memdev")
- mem1=$($CXL list -M -p $port_dev1 | jq -r ".[1].memdev")
- mem2=$($CXL list -M -p $port_dev2 | jq -r ".[2].memdev")
- mem3=$($CXL list -M -p $port_dev3 | jq -r ".[3].memdev")
+ mem1=$($CXL list -M -p $port_dev1 | jq -r ".[0].memdev")
+ mem2=$($CXL list -M -p $port_dev0 | jq -r ".[1].memdev")
+ mem3=$($CXL list -M -p $port_dev1 | jq -r ".[1].memdev")
memdevs="$mem0 $mem1 $mem2 $mem3"
}
base-commit: a871e6153b11fe63780b37cdcb1eb347b296095c
--
2.37.3
next reply other threads:[~2023-11-22 2:57 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-22 2:57 alison.schofield [this message]
2023-11-22 19:38 ` [ndctl PATCH] cxl/test: replace a bad root decoder usage in cxl-xor-region.sh Alison Schofield
2023-11-22 20:13 ` Verma, Vishal L
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=20231122025753.1209527-1-alison.schofield@intel.com \
--to=alison.schofield@intel.com \
--cc=linux-cxl@vger.kernel.org \
--cc=nvdimm@lists.linux.dev \
--cc=vishal.l.verma@intel.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 read-only IMAP folder(s) and NNTP newsgroup(s).