Re: Hardware checksums on bge(4) interfaces

[buhrow%lothlorien.nfbcal.org@localhost (Brian Buhrow)]

        Hello.  Well, it seems that making an assumption had, well, the usual
result of assumptions.  Here's a patch against if_bge.c -current sources.
The previous patch was checked against 5.x, 4.x and 3.x sources, which I
believe covers a broad range of installed users.

-thanks
-Brian


--- if_bge.c.orig       2010-09-15 00:50:06.000000000 -0700
+++ if_bge.c    2010-09-15 00:53:07.000000000 -0700
@@ -1866,7 +1866,7 @@
         */
        CSR_WRITE_4(sc, BGE_MODE_CTL, BGE_DMA_SWAP_OPTIONS |
            BGE_MODECTL_MAC_ATTN_INTR | BGE_MODECTL_HOST_SEND_BDS |
-           BGE_MODECTL_TX_NO_PHDR_CSUM | BGE_MODECTL_RX_NO_PHDR_CSUM);
+           BGE_MODECTL_TX_NO_PHDR_CSUM);
 
        /*
         * BCM5701 B5 have a bug causing data corruption when using
@@ -3444,7 +3444,7 @@
                            cur_rx->bge_tcp_udp_csum;
                        m->m_pkthdr.csum_flags |=
                            (M_CSUM_TCPv4|M_CSUM_UDPv4|
-                            M_CSUM_DATA|M_CSUM_NO_PSEUDOHDR);
+                            M_CSUM_DATA);
                }
 
                /*