From: Stanimir Varbanov <svarbanov@mm-sol.com>
To: Rob Clark <robdclark@gmail.com>, linux-arm-msm@vger.kernel.org
Cc: Stephen Boyd <sboyd@codeaurora.org>,
Bjorn Andersson <bjorn.andersson@sonymobile.com>
Subject: Re: [PATCH] soc/qcom: add OCMEM driver
Date: Tue, 6 Oct 2015 11:35:42 +0300 [thread overview]
Message-ID: <561387DE.70501@mm-sol.com> (raw)
In-Reply-To: <1444059886-2427-1-git-send-email-robdclark@gmail.com>
Hi Rob,
<snip>
> +
> +struct ocmem_config {
> + uint8_t num_regions;
u8
> + uint32_t macro_size;
u32
> +static const struct ocmem_config ocmem_8974_config = {
> + .num_regions = 3, .macro_size = SZ_128K,
> +};
> +
> +static const struct of_device_id ocmem_dt_match[] = {
> + { .compatible = "qcom,ocmem-msm8974", .data = &ocmem_8974_config },
Just a suggestion. As we have REG_OCMEM_HW_PROFILE and
REG_OCMEM_HW_VERSION register why not distinguish by version and hw
profile instead of SoC?
> + {}
> +};
> +
> +static int ocmem_dev_probe(struct platform_device *pdev)
> +{
> + struct device *dev = &pdev->dev;
> + const struct ocmem_config *config = NULL;
> + const struct of_device_id *match;
> + struct resource *res;
> + uint32_t reg, num_banks, region_size;
> + int i, j, ret;
> +
> + /* we need scm to be available: */
> + if (!qcom_scm_is_available())
> + return -EPROBE_DEFER;
> +
> + match = of_match_device(ocmem_dt_match, dev);
> + if (match)
> + config = match->data;
> +
> + if (!config) {
> + dev_err(dev, "unknown config: %s\n", dev->of_node->name);
> + return -ENXIO;
> + }
> +
> + ocmem = devm_kzalloc(dev, sizeof(*ocmem), GFP_KERNEL);
> + if (!ocmem)
> + return -ENOMEM;
> +
> + ocmem->dev = dev;
> + ocmem->config = config;
> +
> + ocmem->core_clk = devm_clk_get(dev, "core_clk");
> + if (IS_ERR(ocmem->core_clk)) {
> + dev_err(dev, "Unable to get the core clock\n");
this dev_err will flood console in EPROBE_DEFER case, IMO should be removed.
> + return PTR_ERR(ocmem->core_clk);
> + }
> +
> + ocmem->iface_clk = devm_clk_get(dev, "iface_clk");
> + if (IS_ERR_OR_NULL(ocmem->iface_clk)) {
> + ret = PTR_ERR(ocmem->iface_clk);
> + ocmem->iface_clk = NULL;
> + /* in probe-defer case, propagate error up and try again later: */
> + if (ret == -EPROBE_DEFER)
> + goto fail;
Why error path of the iface clock is different comparing with core clk.
Also I failed to found a case when common clk framework will return NULL?
if (IS_ERR(ocmem->iface_clk)
return PTR_ERR(ocmem->iface_clk);
this should be enough.
> + }
> +
> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> + "ocmem_ctrl_physical");
> + if (!res) {
> + dev_err(&pdev->dev, "failed to get memory resource\n");
> + ret = -EINVAL;
> + goto fail;
> + }
> +
> + ocmem->mmio = devm_ioremap_nocache(&pdev->dev, res->start,
> + resource_size(res));
I think you could use following to simplify ioremap:
res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
"ocmem_ctrl_physical");
ocmem->mmio = devm_ioremap_resource(dev, res);
if (IS_ERR(ocmem->mmio))
return PTR_ERR(ocmem->mmio);
> + if (IS_ERR(ocmem->mmio)) {
> + dev_err(&pdev->dev, "failed to ioremap memory resource\n");
> + ret = -EINVAL;
> + goto fail;
> + }
> +
> + ocmem->ocmem_mem = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> + "ocmem_physical");
> + if (!ocmem->ocmem_mem) {
> + dev_err(dev, "could not get OCMEM region\n");
> + return -ENXIO;
> + }
> +
<snip>
--
regards,
Stan
next prev parent reply other threads:[~2015-10-06 8:35 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-05 15:44 [PATCH] soc/qcom: add OCMEM driver Rob Clark
2015-10-05 15:46 ` Rob Clark
2015-10-06 8:35 ` Stanimir Varbanov [this message]
2015-10-06 14:36 ` Rob Clark
2015-10-12 21:45 Rob Clark
2015-10-23 12:11 ` Stanimir Varbanov
2015-10-23 17:52 ` Bryan Huntsman
2015-10-23 19:05 ` Rob Clark
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=561387DE.70501@mm-sol.com \
--to=svarbanov@mm-sol.com \
--cc=bjorn.andersson@sonymobile.com \
--cc=linux-arm-msm@vger.kernel.org \
--cc=robdclark@gmail.com \
--cc=sboyd@codeaurora.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.