On 10/02/2015 06:55 AM, Kevin Wolf wrote:
> Am 28.09.2015 um 19:38 hat John Snow geschrieben:
>> the 16bit ide data register is LE by definition.
>>
>> Signed-off-by: John Snow <js...@redhat.com>
>> ---
>> tests/ide-test.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/ide-test.c b/tests/ide-test.c
>> index 5594738..b6e9e1a 100644
>> --- a/tests/ide-test.c
>> +++ b/tests/ide-test.c
>> @@ -633,7 +633,7 @@ static void send_scsi_cdb_read10(uint64_t lba, int
>> nblocks)
>>
>> /* Send Packet */
>> for (i = 0; i < sizeof(Read10CDB)/2; i++) {
>> - outw(IDE_BASE + reg_data, ((uint16_t *)&pkt)[i]);
>> + outw(IDE_BASE + reg_data, cpu_to_le16(((uint16_t *)&pkt)[i]));
>> }
>> }
>>
>> @@ -733,7 +733,7 @@ static void cdrom_pio_impl(int nblocks)
>> size_t offset = i * (limit / 2);
>> size_t rem = (rxsize / 2) - offset;
>> for (j = 0; j < MIN((limit / 2), rem); j++) {
>> - rx[offset + j] = inw(IDE_BASE + reg_data);
>> + rx[offset + j] = le16_to_cpu(inw(IDE_BASE + reg_data));
>> }
>> ide_wait_intr(IDE_PRIMARY_IRQ);
>> }
>
> Why doesn't the access in test_identify() need a fix?
>
> Kevin
>
The strings are stored as BE16 chunks but transmitted via an LE16
register, so if we want to decode the original string, we can apply
*either* an LE16 or a BE16 swap to obtain the byte-ordered string.
Yes, that took me a minute to figure out.
We *could* apply a le16_to_cpu filter to re-obtain the raw original
data, but then we'd just have to run it back through be16_to_cpu to get
the string out anyway.
In this test, we just cheat and run be16_to_cpu, which works.
--js
