Hi Luiz, > On Tue, Nov 11, 2014 at 3:52 AM, Luiz Augusto von Dentz <luiz.dentz@xxxxxxxxx> wrote: > Hi Arman, > > On Mon, Nov 10, 2014 at 7:07 PM, Arman Uguray <armansito@xxxxxxxxxxxx> wrote: >> This patch set implements the most important server-side ATT protocol >> operations by interfacing with the new gatt-db API functions. All >> requests except "Read Multiple" and "Signed Write Command" are provided >> in this set. >> >> This patch set also doesn't address checks for encryption/authentication/ >> authorization permissions. A TODO item has been added to address this >> later. >> >> *v1: Removed offsetof math while casting service pointer to (void *) >> *v2: - Added TODO for using struct iovec >> - Removed bt_gatt_server_set_max_prep_queue_len >> - Always pass write_cb to gatt_db_attribute_write even if the opcode is >> "Write Command". >> - Added more util_debug lines in handler functions. >> >> Arman Uguray (11): >> shared/gatt-server: Implement "Read By Type" request. >> shared/gatt-server: Implement "Find Information" request. >> shared/gatt-server: Implement "Write" request and command. >> shared/gatt-server: Implement "Read" request. >> shared/gatt-server: Implement "Read Blob" request. >> shared/att: Handle incoming confirmation PDU. >> shared/gatt-server: Add functions for sending >> notifications/indications. >> shared/gatt-server: Implement "Prepare Write" request. >> shared/gatt-server: Implement "Execute Write" request. >> shared/gatt-client: Fix alignment warnings. >> TODO: Update shared/gatt-server items. >> >> TODO | 26 +- >> src/shared/att.c | 119 +++-- >> src/shared/gatt-client.c | 6 +- >> src/shared/gatt-server.c | 1221 +++++++++++++++++++++++++++++++++++++++++++--- >> src/shared/gatt-server.h | 12 + >> 5 files changed, 1257 insertions(+), 127 deletions(-) >> >> -- >> 2.1.0.rc2.206.gedb03e5 > > I went ahead and applied this set, I will probably send a patch in a > moment adding a timeout for db read/write operations, I figure 1 > seconds should be enough to start with since LE is focus on low > latency, if that doesn't cut it we should adjust the write queue > according to how many operation it can deliver before it times out. > Sounds reasonable to me. I guess, since the ATT timeout is 30 seconds, a queue length of 30 and operation timeout of 1 seem good for now, even if not optimal for all cases. We can tweak this later as necessary. > > -- > Luiz Augusto von Dentz Cheers, Arman -- To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
