From: Jonathan Cameron <Jonathan.Cameron@Huawei.com>
To: Dan Williams <dan.j.williams@intel.com>
Cc: <linux-cxl@vger.kernel.org>, Dave Jiang <dave.jiang@intel.com>
Subject: Re: [PATCH] cxl/hdm: Extend DVSEC range register emulation for region enumeration
Date: Thu, 30 Mar 2023 18:00:19 +0100 [thread overview]
Message-ID: <20230330180019.00001b7b@Huawei.com> (raw)
In-Reply-To: <168012575521.221280.14177293493678527326.stgit@dwillia2-xfh.jf.intel.com>
On Wed, 29 Mar 2023 14:35:55 -0700
Dan Williams <dan.j.williams@intel.com> wrote:
> One motivation for mapping range registers to decoder objects is
> to use those settings for region autodiscovery.
>
> The need to map a region for devices programmed to use range registers
> is especially urgent now that the kernel no longer routes "Soft
> Reserved" ranges in the memory map to device-dax by default. The CXL
> memory range loses all access mechanisms.
>
> Complete the implementation by filling out ways and granularity, marking
> the DPA reservation, and setting the endpoint-decoder state to signal
> autodiscovery.
>
> Fixes: 09d09e04d2fc ("cxl/dax: Create dax devices for CXL RAM regions")
> Tested-by: Dave Jiang <dave.jiang@intel.com>
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> ---
> drivers/cxl/core/hdm.c | 30 ++++++++++++++++++++++++------
> 1 file changed, 24 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/cxl/core/hdm.c b/drivers/cxl/core/hdm.c
> index 9884b6d4d930..5339c0719177 100644
> --- a/drivers/cxl/core/hdm.c
> +++ b/drivers/cxl/core/hdm.c
> @@ -738,20 +738,26 @@ static int cxl_decoder_reset(struct cxl_decoder *cxld)
> return 0;
> }
>
> -static int cxl_setup_hdm_decoder_from_dvsec(struct cxl_port *port,
> - struct cxl_decoder *cxld, int which,
> - struct cxl_endpoint_dvsec_info *info)
> +static int cxl_setup_hdm_decoder_from_dvsec(
> + struct cxl_port *port, struct cxl_decoder *cxld, u64 *dpa_base,
> + int which, struct cxl_endpoint_dvsec_info *info)
> {
> + struct cxl_endpoint_decoder *cxled;
> + struct range *range;
> + int rc;
> +
> if (!is_cxl_endpoint(port))
> return -EOPNOTSUPP;
>
> - if (!range_len(&info->dvsec_range[which]))
> + cxled = to_cxl_endpoint_decoder(&cxld->dev);
> + range = &info->dvsec_range[which];
> + if (!range_len(range))
You call range_len(range) quite a few times in here.
I'd be tempted to suggest a more minor diff of
u64 len;
...
len = range_len(&info->dvsec_range[which]);
then use that throughout and don't bother with
the local range variable.
range is a bit odd to keep in a local variable as
it's not immediately obvious it's an hpa range - obviously you
could call it hpa_range which would help, but feels like
the length is the important bit except in one place.
> return -ENOENT;
>
> cxld->target_type = CXL_DECODER_EXPANDER;
> cxld->commit = NULL;
> cxld->reset = NULL;
> - cxld->hpa_range = info->dvsec_range[which];
> + cxld->hpa_range = *range;
>
> /*
> * Set the emulated decoder as locked pending additional support to
> @@ -760,6 +766,17 @@ static int cxl_setup_hdm_decoder_from_dvsec(struct cxl_port *port,
> cxld->flags |= CXL_DECODER_F_ENABLE | CXL_DECODER_F_LOCK;
> port->commit_end = cxld->id;
>
> + rc = devm_cxl_dpa_reserve(cxled, *dpa_base, range_len(range), 0);
> + if (rc) {
> + dev_err(&port->dev,
> + "decoder%d.%d: Failed to reserve DPA range %#llx - %#llx\n (%d)",
> + port->id, cxld->id, *dpa_base,
> + *dpa_base + range_len(range) - 1, rc);
> + return rc;
> + }
> + *dpa_base += range_len(range);
> + cxled->state = CXL_DECODER_STATE_AUTO;
> +
> return 0;
> }
>
> @@ -779,7 +796,8 @@ static int init_hdm_decoder(struct cxl_port *port, struct cxl_decoder *cxld,
> } target_list;
>
> if (should_emulate_decoders(info))
> - return cxl_setup_hdm_decoder_from_dvsec(port, cxld, which, info);
> + return cxl_setup_hdm_decoder_from_dvsec(port, cxld, dpa_base,
> + which, info);
>
> ctrl = readl(hdm + CXL_HDM_DECODER0_CTRL_OFFSET(which));
> base = ioread64_hi_lo(hdm + CXL_HDM_DECODER0_BASE_LOW_OFFSET(which));
>
next prev parent reply other threads:[~2023-03-30 17:01 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-29 21:35 [PATCH] cxl/hdm: Extend DVSEC range register emulation for region enumeration Dan Williams
2023-03-29 16:04 ` Gregory Price
2023-03-30 4:21 ` Dan Williams
2023-03-29 17:20 ` Gregory Price
2023-05-16 6:43 ` Gregory Price
2023-03-29 17:21 ` Gregory Price
2023-03-30 6:33 ` Dan Williams
2023-03-30 4:27 ` Gregory Price
2023-04-16 4:05 ` Gregory Price
2023-04-18 5:51 ` Dan Williams
2023-04-20 0:50 ` Gregory Price
2023-03-30 0:06 ` Dave Jiang
2023-03-30 17:00 ` Jonathan Cameron [this message]
2023-03-30 18:24 ` Dan Williams
2023-04-03 23:06 ` [PATCH v2] " Dan Williams
2023-04-03 23:44 ` Dave Jiang
2023-04-04 0:08 ` Dan Williams
2023-04-04 0:16 ` Dave Jiang
2023-04-04 9:19 ` Jonathan Cameron
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=20230330180019.00001b7b@Huawei.com \
--to=jonathan.cameron@huawei.com \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=linux-cxl@vger.kernel.org \
/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