Hi all, Attached is the second submission for our 10GbE Adapter. Many thanks for the input; I believe we have addressed all the comments to date (I cc everybody who helped us with the comments, to ma
Leonid Grossman wrote: 1. At the moment, lspci output looks like "02:02.0 Ethernet controller: Unknown device 17d5:5831 (rev 02)"; do we need to submit a patch for drivers/pci/pci.ids? http://pciids.
Done, thanks for the pointer! I think it will be a good idea to have full support for skb frags in general, and not just for our product. Please let me know if there is a consensus (and hopefully a
A bunch of comments: - if you want to submit the driver for inclusion please submit a patch against a kernel tree, not a tarball. - please try to avoid version ifdefs by provoding the newer APIs on o
raghavendra.koushik@xxxxxxxxx wrote: Hi Jeff, 1. points 7 and 8, when initSharedMem returns error, I call freeSharedMem which should free any partially alloced memory. For #8 quite possibly, and if s
Hi Jeff, 1. points 7 and 8, when initSharedMem returns error, I call freeSharedMem which should free any partially alloced memory. 2. For point 17 and 33 We do support IPV6 checksum offload. There is
Hi Jeff, For #7, the temporary variable tmp_v_addr is assigned as follows (immediately below the code you have quoted): In the freeSharedMem() routine, we go thru' each block of each receive ring and
Hi Jeff, Since the whole nic structure is zeroed out initially, the "block_virt_addr" is initially NULL. So, if the allocation succeeds, it will hold a non-NULL value. So, in the freeSharedMem() if w
ravinandan arakali wrote: Hi Jeff, Since the whole nic structure is zeroed out initially, the "block_virt_addr" is initially NULL. So, if the allocation succeeds, it will hold a non-NULL value. So, i
some questions on few of your comments. When link goes down and I just call netif_carrier_off the upper layer still continues to send packets to the s2io_xmit routine. In order to avoid this, I stop
some questions on few of your comments. 30) do not call netif_stop_queue() and netif_wake_queue() on link events, in s2io_link. Simply call netif_carrier_{on,off}. When link goes down and I just call
Hi all, Please fund attached a source code for S2io 10GbE adapter (with some disclaimers below). Send me your comments/suggestions on the source please, and we will address the code changes (if any)
Noticed the setup loopback test seems to register for a packet type and then forget to unregister that type! Also nothing really restricts the packet type to only coming in on the expected interface;
Hi Stephen, Below are responses from our developer. Thanks for the input, Leonid The packet type gets unregistered at the end of the test in the 'reset_loopback' function through the system call 'dev
I see the point; we'll go ahead and get rid of the loopback. We actually have pretty extensive standalone diag tool that is based on diag driver. Thanks, Leonid
Leonid Grossman wrote: The loopback test is there as a part of the ethtool's diagnostic option. There are pros and cons of having the test in there I guess, anyone else has an opinion on this? Do oth
Hi all, Attached is the second submission for our 10GbE Adapter. Many thanks for the input; I believe we have addressed all the comments to date (I cc everybody who helped us with the comments, to ma