While I was looking at the implicit-fallthrough in pflash_cfi02, I noticed that the code in pflash_cfi01 was different and wrong. Specifically, the cfi01 code is attempting to do the same "fall through to treat as a normal read" trick, but somebody has moved the default case so it's no longer above the thing it wants to fall through to. Patch 2 corrects this.
Patch 1 fixes a bug which was the only way you could get into the default case in the first place: this default is a "can't happen" case which implies that the pflash_read() switch is missing handling code for a command which is implemented in pflash_write(). In fact for cfi01 there were some missing cases, so we add them, based on the spec which defines what the read should return for these commands, which is the status register value in all cases. [The patch order is merely so that by the time we get to patch 2 it is a "does not change user visible behaviour" change.] Peter Maydell (2): hw/pflash_cfi01: Make read after byte-write or erase return status hw/pflash_cfi01: Treat read in unknown command state as read hw/pflash_cfi01.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) -- 1.7.9.5
