> On 04/09/2021 2:16 PM Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx> wrote: > Thank you for the patch! > > > From: Ulrich Hecht, Sent: Friday, April 9, 2021 5:25 PM > > > > Since the transition to hrtimers, the calculation does not involve jiffies > > any longer, which has led to confusion. State the times in ms instead. > > IIUC, the unit of rx_timeout was changed from milliseconds to microseconds > when hrtimer is used. D'oh! > So, almost all comments was not needed. > > Also, I'm wondering if the following condition is not needed or not. > This is because this was "20ms". But, perhaps, the driver will not > set rx_timeout to "20us" or small. > > if (s->rx_timeout < 20) > s->rx_timeout = 20; A more helpful version of the comment is in 3089f381fbaf5: /* * Calculate delay for 1.5 DMA buffers: see * drivers/serial/serial_core.c::uart_update_timeout(). With 10 bits * (CS8), 250Hz, 115200 baud and 64 bytes FIFO, the above function * calculates 1 jiffie for the data plus 5 jiffies for the "slop(e)." * Then below we calculate 3 jiffies (12ms) for 1.5 DMA buffers (3 FIFO * sizes), but it has been found out experimentally, that this is not * enough: the driver too often needlessly runs on a DMA timeout. 20ms * as a minimum seem to work perfectly. */ I think we still want that, but it should of course be 20000, not 20. CU Uli
