Takashi Sakamoto wrote:
On Apr 19 2018 17:54, Clemens Ladisch wrote:
Takashi Sakamoto wrote:
183 /* Some old firmware versions do not have the following global registers: */ ... 188 #define GLOBAL_CLOCK_CAPABILITIES 0x064
But this is not proper in my case because the maximum offset on global sub-address space is 0x05c in quadlet unit.
IIRC that comment was based on the documentation. Apparently, there was a misunderstanding of "size 0x60" vs. "up to the register at 0x60".
From my curiosity, do you know which version of DICE firstly supports GLOBAL_VERSION register? Can you read it on the documentation?
I misremembered: I did not find anything about this in the documentation.
This is derived from TCAT's host driver, which unconditionally reads everything up to the clock capabilities register, and checks for a size larger than 0x64 to detect if the firmware supports clock capabilities.
The version register is checked only once (in get_subaddrs()) and does not actually affect the behaviour of the driver, so it would not hurt to simply omit that check on the old firmware.
Hm. The check is applied to the other spaces such as rx/tx, so I'd like to leave it as is.
Sorry, "that check" meant the version check for 0x01000000. Your patch already implements what I suggested.
With this patch[1], ALSA dice driver can handle the unit. In this weekend, I'll post it. Your reviewing is very helpful.
[1] https://github.com/takaswie/snd-firewire-improve/commit/1889e06b9593e16da281...
- /* Old firmware doesn't support these fields. */
- if (dice->clock_caps == 0) {
dice->sync_offset = be32_to_cpu(pointers[6]) * 4;dice->rsrv_offset = be32_to_cpu(pointers[8]) * 4;- }
Shouldn't this be clock_caps!=0?
Regards, Clemens