
21 May
2019
21 May
'19
8:42 p.m.
Hi TF,
On Tue, May 21, 2019 at 05:21:15AM +0000, Chee, Tien Fong wrote:
On Thu, 2019-05-16 at 15:54 -0500, Andreas Dannenberg wrote:
- if (mmc_dev < 0) {
pr_err("%s: Getting MMC device index failed (%d)\n",
__func__,
mmc_dev);
return mmc_dev;
- }
- ret = uclass_get_device(UCLASS_MMC, mmc_dev, &mmcdev);
- if (ret < 0) {
pr_err("%s: Getting MMC device failed (%d)\n",
__func__, ret);
return ret;
- }
- ret = mmc_initialize(NULL);
- if (ret < 0) {
pr_err("%s: Initializing MMC device failed (%d)\n",
__func__,
ret);
return ret;
- }
- mmc = mmc_get_mmc_dev(mmcdev);
- if (!mmc) {
pr_err("%s: Getting underlying MMC device failed\n",
__func__);
return -ENODEV;
- }
- ret = mmc_init(mmc);
- if (ret) {
printf("%s: mmc init failed with error: %d\n",
__func__, ret);
return ret;
- }
The probe function in fs loader already support the block init, and the block init would init the mmc.
Yes I'm sure this could have gotten simplified if I were to use the DTS to describe the boot device. But as explained earlier I was exploring a way to make it work without DTS for added flexibility so I can pick the boot device at runtime.
-- Andreas Dannenberg Texas Instruments Inc