| commit | c147b33ebd48e5737de96215472b8809394e12e6 | |
| author | Andrew Baumann <Andrew.Baumann@microsoft.com> | |
| Sat, 14 Nov 2015 00:20:43 +0000 (13 16:20 -0800) | ||
| committer | Andrew Baumann <Andrew.Baumann@microsoft.com> | |
| Sat, 14 Nov 2015 00:20:43 +0000 (13 16:20 -0800) | ||
| tree | 40b0431629b914546c3327de1501264c6234f45c | treesnapshot (tar.gz zip) |
| parent | e53cb616880ad05f35d06da6896648f659632934 | commitdiff |
sd: replace UEFI-specific kludges with correct implementations
* UEFI has a bug where it assumes that if its first ACMD41 succeeds,
the card is in ready state. This contravenes the spec, but real
cards presumably always take at least some time to initialise, so I
changed the code to model a small "delay".
* Replaced broken implementation of CMD23 (set block count) with one
that follows the spec. This shouldn't be necessary, since the
command is listed as optional, but UEFI assumes its existence.
* UEFI has a bug where it assumes that if its first ACMD41 succeeds,
the card is in ready state. This contravenes the spec, but real
cards presumably always take at least some time to initialise, so I
changed the code to model a small "delay".
* Replaced broken implementation of CMD23 (set block count) with one
that follows the spec. This shouldn't be necessary, since the
command is listed as optional, but UEFI assumes its existence.
| hw/sd/sd.c | diffblobblamehistory |