Hi On 3 September 2014 15:08, Jean-Michel Hautbois <jean-michel.hautbois@xxxxxxxxxxx> wrote: > 2014-09-03 11:09 GMT+02:00 Ulf Hansson <ulf.hansson@xxxxxxxxxx>: >> On 3 September 2014 11:02, Adrian Hunter <adrian.hunter@xxxxxxxxx> wrote: >>> On 09/03/2014 11:30 AM, Ulf Hansson wrote: >>>> On 2 September 2014 17:49, Jean-Michel Hautbois >>>> <jean-michel.hautbois@xxxxxxxxxxx> wrote: >>>>> This property is useful when we don't want to access boot partitions on eMMC >>>>> >>>>> Signed-off-by: Jean-Michel Hautbois <jean-michel.hautbois@xxxxxxxxxxx> >>>>> --- >>>>> Documentation/devicetree/bindings/mmc/mmc.txt | 1 + >>>>> drivers/mmc/host/sdhci-esdhc-imx.c | 8 ++++++++ >>>>> include/linux/platform_data/mmc-esdhc-imx.h | 1 + >>>>> 3 files changed, 10 insertions(+) >>>>> >>>>> diff --git a/Documentation/devicetree/bindings/mmc/mmc.txt b/Documentation/devicetree/bindings/mmc/mmc.txt >>>>> index 431716e..59cc854 100644 >>>>> --- a/Documentation/devicetree/bindings/mmc/mmc.txt >>>>> +++ b/Documentation/devicetree/bindings/mmc/mmc.txt >>>>> @@ -40,6 +40,7 @@ Optional properties: >>>>> - mmc-hs200-1_2v: eMMC HS200 mode(1.2V I/O) is supported >>>>> - mmc-hs400-1_8v: eMMC HS400 mode(1.8V I/O) is supported >>>>> - mmc-hs400-1_2v: eMMC HS400 mode(1.2V I/O) is supported >>>>> +- no-boot-part : when preset, tells to access boot partitions >>>>> >>>>> *NOTE* on CD and WP polarity. To use common for all SD/MMC host controllers line >>>>> polarity properties, we have to fix the meaning of the "normal" and "inverted" >>>>> diff --git a/drivers/mmc/host/sdhci-esdhc-imx.c b/drivers/mmc/host/sdhci-esdhc-imx.c >>>>> index ccec0e3..439e663 100644 >>>>> --- a/drivers/mmc/host/sdhci-esdhc-imx.c >>>>> +++ b/drivers/mmc/host/sdhci-esdhc-imx.c >>>>> @@ -942,6 +942,11 @@ sdhci_esdhc_imx_probe_dt(struct platform_device *pdev, >>>>> if (of_property_read_u32(np, "fsl,delay-line", &boarddata->delay_line)) >>>>> boarddata->delay_line = 0; >>>>> >>>>> + if (of_find_property(np, "no-boot-part", NULL)) >>>>> + boarddata->access_boot_part = false; >>>>> + else >>>>> + boarddata->access_boot_part = true; >>>>> + >>>>> return 0; >>>>> } >>>>> #else >>>>> @@ -1119,6 +1124,9 @@ static int sdhci_esdhc_imx_probe(struct platform_device *pdev) >>>>> host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V; >>>>> } >>>>> >>>>> + if (!boarddata->access_boot_part) >>>>> + host->mmc->caps2 |= MMC_CAP2_BOOTPART_NOACC; >>>>> + >>>> >>>> Hmm, I don't think MMC_CAP2_BOOTPART_NOACC should have a DT binding. >>>> Does it describe the hardware in some form? >>>> >>>> Actually I would like to question why MMC_CAP2_BOOTPART_NOACC exists >>>> at all. If there are cards that don't supports the BOOT area, >>>> shouldn't we have a card quirk for it instead of a host cap? Maybe >>>> Adrian Hunter, how originally wrote the patch for adding >>>> MMC_CAP2_BOOTPART_NOACC, could help me understand the reasons behind >>>> it!? >>> >>> It was added because platform firmware was able to prevent access to the >>> boot partitions (for security I think), so attempts to access them would >>> fail messily. It was not related to any specific card. >> >> Adrian, appreciate your clarification. After all it seems like adding >> a DT binding for it should be appropriate. >> >> Kind regards >> Uffe > > Thanks Adrian :). > Well, there is boot partitions and rpmb partition, and maybe should we > have a binding to prevent access to both of them ? > Something else came to my mind, when you want to boot on eMMC, do you > need to write u-boot in boot partitions or is it written at the > logical adress 0 which is what fdisk uses as start ? > Because, if this is not usuable but just scanned I can't see why we > bother doing it... ? Hi Jean-Michel, I am not sure adding a DT binding for non access to rpmb would be needed. At least until we heard of a similar case as Adrian describes but for rpmb. BTW, I just posted a patch which disabled partition scan of the boot area, what to you think about that? http://marc.info/?l=linux-mmc&m=140973496402028&w=2 Finally I am also wondering whether we could and thus should, handle these situations entirely without using a host cap. In principle what we need is a more sophisticated error handling when the switch errors occurs, while trying to switch to the boot area/rpmb partitions. Could you maybe investigate that option, before we decide to add a new DT binding? Kind regards Uffe -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
