From: Alison Schofield <alison.schofield@intel.com>
To: Yao Xingtao <yaoxt.fnst@fujitsu.com>
Cc: dave@stgolabs.net, jonathan.cameron@huawei.com,
dave.jiang@intel.com, vishal.l.verma@intel.com,
ira.weiny@intel.com, dan.j.williams@intel.com,
jim.harris@samsung.com, linux-cxl@vger.kernel.org
Subject: Re: [PATCH] cxl/core/region: check interleave way capability
Date: Mon, 1 Apr 2024 09:50:40 -0700 [thread overview]
Message-ID: <Zgrl4HTeOPCow/tu@aschofie-mobl2> (raw)
In-Reply-To: <20240401075635.9333-1-yaoxt.fnst@fujitsu.com>
On Mon, Apr 01, 2024 at 03:56:35AM -0400, Yao Xingtao wrote:
> I used qemu to emulaute the cxl.mem and attmpted to create a 6-way
> region, the region was created successfully, but I could not access the
> memory properly:
> $ numactl -m 2 ls
> Segmentation fault (core dumped)
>
> I found the root cause is that the logic of converting HPA to DPA for
> 3-way, 6-way and 12-way were not implimented on qemu side.
>
> But qemu has already disable the capability, so we should not create
> region in such ways.
>
> So I think we should check whether the interleave way is supported by
> the target while attaching it to region.
Nice find!
I'm wondering if this may be done where other hdm caps are parsed,
at parse_hdm_decoder_caps() and saved as a capability of the port,
as opposed to reading the register on region creation attempts.
--Alison
>
> Link: https://lore.kernel.org/qemu-devel/20240327014653.26623-1-yaoxt.fnst@fujitsu.com/T/#r
> Signed-off-by: Yao Xingtao <yaoxt.fnst@fujitsu.com>
> ---
> drivers/cxl/core/region.c | 25 +++++++++++++++++++++++++
> drivers/cxl/cxl.h | 2 ++
> 2 files changed, 27 insertions(+)
>
> diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
> index 5c186e0a39b9..dde66b7c9e3f 100644
> --- a/drivers/cxl/core/region.c
> +++ b/drivers/cxl/core/region.c
> @@ -1786,6 +1786,24 @@ static int cxl_region_sort_targets(struct cxl_region *cxlr)
> return rc;
> }
>
> +static bool check_iw_capability(struct cxl_endpoint_decoder *cxled, u8 iw)
> +{
> + struct cxl_port *port = to_cxl_port(cxled->cxld.dev.parent);
> + struct cxl_hdm *cxlhdm = dev_get_drvdata(&port->dev);
> + void __iomem *hdm = cxlhdm->regs.hdm_decoder;
> + u32 cap;
> +
> + cap = readl(hdm + CXL_HDM_DECODER_CAP_OFFSET);
> + if (!FIELD_GET(CXL_HDM_DECODER_INTERLEAVE_3_6_12_WAY, cap) &&
> + (iw == 3 || iw == 6 || iw == 12))
> + return false;
> +
> + if (!FIELD_GET(CXL_HDM_DECODER_INTERLEAVE_16_WAY, cap) && iw == 16)
> + return false;
> +
> + return true;
> +}
> +
> static int cxl_region_attach(struct cxl_region *cxlr,
> struct cxl_endpoint_decoder *cxled, int pos)
> {
> @@ -1796,6 +1814,13 @@ static int cxl_region_attach(struct cxl_region *cxlr,
> struct cxl_dport *dport;
> int rc = -ENXIO;
>
> + if (!check_iw_capability(cxled, p->interleave_ways)) {
> + dev_dbg(&cxlr->dev,
> + "%s with region interleave ways: %d is not supported\n",
> + dev_name(&cxled->cxld.dev), p->interleave_ways);
> + return -EOPNOTSUPP;
> + }
> +
> if (cxled->mode != cxlr->mode) {
> dev_dbg(&cxlr->dev, "%s region mode: %d mismatch: %d\n",
> dev_name(&cxled->cxld.dev), cxlr->mode, cxled->mode);
> diff --git a/drivers/cxl/cxl.h b/drivers/cxl/cxl.h
> index 534e25e2f0a4..da8a487ededa 100644
> --- a/drivers/cxl/cxl.h
> +++ b/drivers/cxl/cxl.h
> @@ -45,6 +45,8 @@
> #define CXL_HDM_DECODER_TARGET_COUNT_MASK GENMASK(7, 4)
> #define CXL_HDM_DECODER_INTERLEAVE_11_8 BIT(8)
> #define CXL_HDM_DECODER_INTERLEAVE_14_12 BIT(9)
> +#define CXL_HDM_DECODER_INTERLEAVE_3_6_12_WAY BIT(11)
> +#define CXL_HDM_DECODER_INTERLEAVE_16_WAY BIT(12)
> #define CXL_HDM_DECODER_CTRL_OFFSET 0x4
> #define CXL_HDM_DECODER_ENABLE BIT(1)
> #define CXL_HDM_DECODER0_BASE_LOW_OFFSET(i) (0x20 * (i) + 0x10)
> --
> 2.37.3
>
next prev parent reply other threads:[~2024-04-01 16:50 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-01 7:56 [PATCH] cxl/core/region: check interleave way capability Yao Xingtao
2024-04-01 16:50 ` Alison Schofield [this message]
2024-04-02 5:43 ` Xingtao Yao (Fujitsu)
2024-04-01 18:34 ` Dan Williams
2024-04-02 5:55 ` Xingtao Yao (Fujitsu)
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=Zgrl4HTeOPCow/tu@aschofie-mobl2 \
--to=alison.schofield@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=dave@stgolabs.net \
--cc=ira.weiny@intel.com \
--cc=jim.harris@samsung.com \
--cc=jonathan.cameron@huawei.com \
--cc=linux-cxl@vger.kernel.org \
--cc=vishal.l.verma@intel.com \
--cc=yaoxt.fnst@fujitsu.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