Best Regards Jerry Huang > -----Original Message----- > From: Anton Vorontsov [mailto:cbouatmailru@xxxxxxxxx] > Sent: Wednesday, September 12, 2012 11:38 AM > To: Huang Changming-R66093 > Cc: Wood Scott-B07421; linuxppc-dev@xxxxxxxxxxxxxxxx; linux- > mmc@xxxxxxxxxxxxxxx > Subject: Re: [PATCH 2/3] powerpc/esdhc: add property to disable the CMD23 > > On Wed, Sep 12, 2012 at 03:19:18AM +0000, Huang Changming-R66093 wrote: > [...] > > I don't think it is the best way to do it. For the VVN2.2 or older, > > some silicon support this feature (mpc8536 and p2020), but other > > silicones don't support it (e.g. p4080, p102x). Though, the current > > p5/p4/p3 has supported this feature, can we sure the future silicon > > support it? So I think the best way is to specify it in device tree > > as 'sdhci,auto-cmd12' > > In addition to your current patches, you could just add another patch > that blacklists affected SOC revisions based on the info from PVR/SVR. > > For example, see gfar_detect_errata() in > drivers/net/ethernet/freescale/gianfar.c. > > That way you could help users that don't have the newest device trees. > Hi, Anton Thanks. But, these patches are just for latest kernel, if the user don't have the latest device tree, Then they don't have the latest kernel, and these patches can't be used for them directly, Though provide the function like gfar_detect_errata, they must modify their codes. And I don't know if the future silicon can support CMD23, if not, we must modify sdhc driver again. When we use the device tree to identify it, if the silicon does not support this feature, we just add this property into the new device tree, not need to modify SDHC driver. I think, this is the best way. ��.n��������+%������w��{.n�����{��i��)�� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
