sys: add uuid module for central ID/addr generation #6314
Conversation
haukepetersen
added
Area: drivers
|
255 ID's ought to be enough for everybody! Seriously, would it be hard to use a larger type for last_used, and XOR all of it's bits? |
|
would be slightly more overhead in two dimensions:
|
| /** | ||
| * @defgroup sys_uuid | ||
| * @ingroup sys | ||
| * @brief Generate system wide unique IDs |
There was a problem hiding this comment.
With that name it maybe woth pointing out, that these "U"UIDs are not compatible to RFC 4122.
There was a problem hiding this comment.
hmm, other ideas how we could name the module so there is no confusion?
There was a problem hiding this comment.
Maybe not applicable due to the more generalized nature of your API, but what about devid (or dev_id, so nobody will call it like some Germans call Davids ;-))
There was a problem hiding this comment.
hm, I think a more general name would be the better fit.
There was a problem hiding this comment.
but now that I thought about it, let's do it like that: I add that comment you requested and if we find a better name, we can refactor then.
| */ | ||
|
|
||
| /** | ||
| * @defgroup sys_uuid |
| /** | ||
| * @brief Get a unique ID | ||
| * | ||
| * The resulting ID is build from the base ID XORed with an 8-bit incrementing |
There was a problem hiding this comment.
s/base ID/base ID generated with uuid_base() which is/
| * The resulting ID is build from the base ID XORed with an 8-bit incrementing | ||
| * counter value into the most significant byte. | ||
| * | ||
| * @note The resulting UUID will repeat after 256 calls. |
| * | ||
| * @param[out] buf memory location to copy the UUID into. MUST be able to | ||
| * hold at least @p len byte | ||
| * @param[in] len targeted length of the UUID in byte |
There was a problem hiding this comment.
s/targeted//
s/byte/bytes/
| * @note The resulting UUID will repeat after 256 calls. | ||
| * | ||
| * @param[out] buf memory location to copy the UUID into. MUST be able to | ||
| * hold at least @p len byte |
| * | ||
| * @param[out] buf memory location to copy the UUID into. MUST be able to | ||
| * hold at least @p len byte | ||
| * @param[in] len targeted length of the UUID in byte |
There was a problem hiding this comment.
s/targeted//
s/byte/bytes/
| * not. | ||
| * | ||
| * @param[out] buf memory location to copy the UUID into. MUST be able to | ||
| * hold at least @p len byte |
| /** | ||
| * @brief Get a custom unique ID based on a user given generator value | ||
| * | ||
| * The resulting ID is build from the base ID XORed with @p gen in the base |
| * | ||
| * @param[out] buf memory location to copy the UUID into. MUST be able to | ||
| * hold at least @p len byte | ||
| * @param[in] len targeted length of the UUID in byte |
| * @brief Get a custom unique ID based on a user given generator value | ||
| * | ||
| * The resulting ID is build from the base ID XORed with @p gen in the base | ||
| * ID's most significant bytes. |
There was a problem hiding this comment.
Is this really true? It's late and the code is somewhat hard to read ;)
haukepetersen
added
the
Community: Hack'n'ACK candidate
haukepetersen
force-pushed
the
add_uuid
branch
from
a97245c to
ef07ddd
Compare
|
rebased and (hopefully) fixed doc issues |
OlegHahm
added
the
CI: needs squashing
haukepetersen
force-pushed
the
add_uuid
branch
from
ef07ddd to
863a872
Compare
|
squashed and rebased. |
haukepetersen
added
CI: ready for build
haukepetersen
force-pushed
the
add_uuid
branch
from
863a872 to
88636b6
Compare
|
rebased. Somehow Murdock failed, but can't really tell what the problem was.. |
|
And go |
Wanted to do this for a long time: this PR centralizes the generation of unique IDs/addresses. Now all network drivers use the same code to generate their hardware addresses.
Addresses #5055 (though not all problems are solved?!):
uuid_custom()), but I did not yet think of a good value to use as custom generator param for this function...@PeterKietzmann: feel free to postpone the milestone to 2017.04...