On Tue, Feb 13, 2007 at 05:10:44PM +0900, Roman Mashak wrote: > I kept exploring the sources of rtl8139 NIC and stuck with the weird spot > in 'rtl8139_hw_start()' (same both in a 2.4.x and 2.6.x). I don't post here > very detailed code, as I assume everybody here has access to linux source > tree :) > > > ... > /* init Tx buffer DMA addresses */ > for (i = 0; i < NUM_TX_DESC; i++) > RTL_W32_F (TxAddr0 + (i * 4), tp->tx_bufs_dma + (tp->tx_buf[i] - > tp->tx_bufs)); > > RTL_W32_F is a macro exposing into 'writel(tp->tx_bufs_dma + > (tp->tx_buf[i] - tp->tx_bufs), ...)' call. By the time of this call > 'tx_buf' has been properly initialised and contains addresses of 4 TX > descriptors. > > As I understand 'tx_bufs_dma' holds the physical address of DMA memory, > while 'tx_bufs' - kernel virtual address. > > What I can't figure out is the point of this calculations: tp->tx_bufs_dma > + (tp->tx_buf[i] - tp->tx_bufs) > > Why can't we simply have 'tp->tx_bufs_dma + (i * 4)' ? Probably because you can't assume that the buffer addresses are 4 bytes apart. Erik -- They're all fools. Don't worry. Darwin may be slow, but he'll eventually get them. -- Matthew Lammers in alt.sysadmin.recovery
Attachment:
signature.asc
Description: Digital signature
