diff options
author | Marcin Wojtas <mw@FreeBSD.org> | 2018-05-10 09:37:54 +0000 |
---|---|---|
committer | Marcin Wojtas <mw@FreeBSD.org> | 2018-05-10 09:37:54 +0000 |
commit | 2339f28c6e791527deed6cd2a860fb5987637237 (patch) | |
tree | 72973df44de1fec725e9afa3d4324112e526e4dd /sys/dev/ena | |
parent | 43fd679efbdbf300dee851fb77eb4240ce51e014 (diff) | |
download | src-2339f28c6e791527deed6cd2a860fb5987637237.tar.gz src-2339f28c6e791527deed6cd2a860fb5987637237.zip |
Do not pass header length to the ENA controller
Header length is optional hint for the ENA device. Because It is not
guaranteed that every packet header will be in the first mbuf
segment, it is better to skip passing any information. If the header
length will be indicating invalid value (different than 0), then the
packet will be dropped.
This kind situation can appear, when the UDP packet will be fragmented
by the stack in the ip_fragment() function.
Submitted by: Michal Krawczyk <mk@semihalf.com>
Reported by: Krishna Yenduri <kyenduri@brkt.com>
Obtained from: Semihalf
Sponsored by: Amazon, Inc.
Notes
Notes:
svn path=/head/; revision=333456
Diffstat (limited to 'sys/dev/ena')
-rw-r--r-- | sys/dev/ena/ena.c | 12 |
1 files changed, 9 insertions, 3 deletions
diff --git a/sys/dev/ena/ena.c b/sys/dev/ena/ena.c index 174549120d92..53d81d100498 100644 --- a/sys/dev/ena/ena.c +++ b/sys/dev/ena/ena.c @@ -2742,7 +2742,7 @@ ena_xmit_mbuf(struct ena_ring *tx_ring, struct mbuf **mbuf) uint16_t req_id; uint16_t push_len; uint16_t ena_qid; - uint32_t len, nsegs, header_len; + uint32_t nsegs, header_len; int i, rc; int nb_hw_desc; @@ -2766,12 +2766,18 @@ ena_xmit_mbuf(struct ena_ring *tx_ring, struct mbuf **mbuf) tx_info->num_of_bufs = 0; ena_buf = tx_info->bufs; - len = (*mbuf)->m_len; ena_trace(ENA_DBG | ENA_TXPTH, "Tx: %d bytes", (*mbuf)->m_pkthdr.len); push_len = 0; - header_len = min_t(uint32_t, len, tx_ring->tx_max_header_size); + /* + * header_len is just a hint for the device. Because FreeBSD is not + * giving us information about packet header length and it is not + * guaranteed that all packet headers will be in the 1st mbuf, setting + * header_len to 0 is making the device ignore this value and resolve + * header on it's own. + */ + header_len = 0; push_hdr = NULL; rc = bus_dmamap_load_mbuf_sg(adapter->tx_buf_tag, tx_info->map, |