-
Michael Zaidman authored
A random i2c read operation in EEPROM devices is implemented as a dummy write operation, followed by a current address read operation. The dummy write operation is used to load the target byte or word address (a.k.a offset) into the offset counter, from which the subsequent read operation then reads. To support longer than one HID report size random read, the ft260 driver issues multiple pairs of i2c write offset + read data transactions of HID report size so that the EEPROM device sees many i2c random read requests from different offsets. Two issues with the current implementation: - This approach suffers from extra overhead caused by writing offset requests. - Necessity to handle offset per HID report in big-endian representation as EEPROM devices expect. The current implementation does not do it and correctly handles the reads up to 60 bytes only. This patch addresses both issues by implementing more efficient approach. It issues a single i2c read request of up to the EEPROM page size and then waits for the data to arrive in multiple HID reports. For example, to read the 256 bytes from a 24LC512 chip, which has 128 bytes page size, the old method performs six ft260_i2c_write_read transactions while the new - two only. Before: $ sudo ./i2cperf -d 2 -o 2 -s 128 -r 0-0xff 13 0x51 -S Read block via i2ctransfer by chunks ------------------------------------------------------------------- data rate(bps) efficiency(%) data size(B) total IOs IO size(B) ------------------------------------------------------------------- 40803 85 256 2 128 Kernel log of a single 128 bytes read request: [ +2.376308] ft260_i2c_write_read: read_off 0x0 left_len 128 len 60 [ +0.000002] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0 [ +0.000707] ft260_xfer_status: bus_status 0x41, clock 100 [ +0.000173] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 60 [ +0.008660] ft260_raw_event: i2c resp: rep 0xde len 60 [ +0.000156] ft260_xfer_status: bus_status 0x20, clock 100 [ +0.000001] ft260_i2c_write_read: read_off 0x3c left_len 68 len 60 [ +0.000001] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x3c [ +0.001034] ft260_xfer_status: bus_status 0x41, clock 100 [ +0.000191] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 60 [ +0.008614] ft260_raw_event: i2c resp: rep 0xde len 60 [ +0.000203] ft260_xfer_status: bus_status 0x20, clock 100 [ +0.000001] ft260_i2c_write_read: read_off 0x78 left_len 8 len 8 [ +0.000001] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x78 [ +0.000987] ft260_xfer_status: bus_status 0x41, clock 100 [ +0.000192] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 8 [ +0.002614] ft260_raw_event: i2c resp: rep 0xd1 len 8 [ +0.000200] ft260_xfer_status: bus_status 0x20, clock 100 After: $ sudo ./i2cperf -d 2 -o 2 -s 128 -r 0-0xff 13 0x51 -S Read block via i2ctransfer by chunks ------------------------------------------------------------------- data rate(bps) efficiency(%) data size(B) total IOs IO size(B) ------------------------------------------------------------------- 43990 85 256 2 128 Kernel log of a single 128 bytes read request: [ +1.464346] ft260_i2c_write_read: off 0x0 rlen 128 wlen 2 [ +0.000002] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0 [ +0.001653] ft260_xfer_status: bus_status 0x41, clock 100 [ +0.000188] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 128 rlen 60 flag 0x3 [ +0.008609] ft260_raw_event: i2c resp: rep 0xde len 60 [ +0.000157] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 68 rlen 60 flag 0x0 [ +0.008840] ft260_raw_event: i2c resp: rep 0xde len 60 [ +0.000203] ft260_xfer_status: bus_status 0x40, clock 100 [ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 8 rlen 8 flag 0x4 [ +0.002794] ft260_raw_event: i2c resp: rep 0xd1 len 8 [ +0.000201] ft260_xfer_status: bus_status 0x20, clock 100 Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com> Tested-by: Guillaume Champagne <champagne.guillaume.c@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
0acb869f