| Message ID | 20201028184310.7017-3-xie.he.0141@gmail.com |
|---|---|
| State | Changes Requested |
| Delegated to: | David Miller |
| Headers | show |
| Series | net: hdlc_fr: Add support for any Ethertype | expand |
| Context | Check | Description |
|---|---|---|
| jkicinski/cover_letter | success | Link |
| jkicinski/fixes_present | success | Link |
| jkicinski/patch_count | success | Link |
| jkicinski/tree_selection | success | Clearly marked for net-next |
| jkicinski/subject_prefix | success | Link |
| jkicinski/source_inline | success | Was 0 now: 0 |
| jkicinski/verify_signedoff | success | Link |
| jkicinski/module_param | success | Was 0 now: 0 |
| jkicinski/build_32bit | fail | Errors and warnings before: 4 this patch: 4 |
| jkicinski/kdoc | success | Errors and warnings before: 0 this patch: 0 |
| jkicinski/verify_fixes | success | Link |
| jkicinski/checkpatch | fail | Link |
| jkicinski/build_allmodconfig_warn | success | Errors and warnings before: 0 this patch: 0 |
| jkicinski/header_inline | success | Link |
| jkicinski/stable | success | Stable not CCed |
On Wed, Oct 28, 2020 at 10:12 PM Xie He <xie.he.0141@gmail.com> wrote: > > The eth_type_trans function is called when we receive frames carrying > Ethernet frames. This function expects a non-NULL pointer as an argument, > and assigns it directly to skb->dev. > > However, the code handling other types of frames first assigns a pointer > to "dev", and then at the end checks whether the value is NULL, and if it > is not NULL, assigns it to skb->dev. > > The two flows are different. Mixing them in this function makes the code > messy. It's better that we convert the second flow to align with how > eth_type_trans does things. > > So this patch changes the code to: first make sure the pointer is not > NULL, then assign it directly to skb->dev. "dev" is no longer needed until > the end where we use it to update stats. No need for dev at all then?
On Thu, Oct 29, 2020 at 9:58 AM Willem de Bruijn <willemdebruijn.kernel@gmail.com> wrote: > > No need for dev at all then? Right, there's no need for "dev" at all actually. I kept "dev" just to keep the code for updating "dev->stats" simpler, because otherwise we need to write "skb->dev->stats" instead, and there are 3 lines where we need to write it.
diff --git a/drivers/net/wan/hdlc_fr.c b/drivers/net/wan/hdlc_fr.c index c774eff44534..ac65f5c435ef 100644 --- a/drivers/net/wan/hdlc_fr.c +++ b/drivers/net/wan/hdlc_fr.c @@ -880,7 +880,7 @@ static int fr_rx(struct sk_buff *skb) u8 *data = skb->data; u16 dlci; struct pvc_device *pvc; - struct net_device *dev = NULL; + struct net_device *dev; if (skb->len <= 4 || fh->ea1 || data[2] != FR_UI) goto rx_error; @@ -930,13 +930,17 @@ static int fr_rx(struct sk_buff *skb) } if (data[3] == NLPID_IP) { + if (!pvc->main) + goto rx_drop; skb_pull(skb, 4); /* Remove 4-byte header (hdr, UI, NLPID) */ - dev = pvc->main; + skb->dev = pvc->main; skb->protocol = htons(ETH_P_IP); } else if (data[3] == NLPID_IPV6) { + if (!pvc->main) + goto rx_drop; skb_pull(skb, 4); /* Remove 4-byte header (hdr, UI, NLPID) */ - dev = pvc->main; + skb->dev = pvc->main; skb->protocol = htons(ETH_P_IPV6); } else if (skb->len > 10 && data[3] == FR_PAD && @@ -950,13 +954,16 @@ static int fr_rx(struct sk_buff *skb) case ETH_P_IPX: case ETH_P_IP: /* a long variant */ case ETH_P_IPV6: - dev = pvc->main; + if (!pvc->main) + goto rx_drop; + skb->dev = pvc->main; skb->protocol = htons(pid); break; case 0x80C20007: /* bridged Ethernet frame */ - if ((dev = pvc->ether) != NULL) - skb->protocol = eth_type_trans(skb, dev); + if (!pvc->ether) + goto rx_drop; + skb->protocol = eth_type_trans(skb, pvc->ether); break; default: @@ -970,17 +977,13 @@ static int fr_rx(struct sk_buff *skb) goto rx_drop; } - if (dev) { - dev->stats.rx_packets++; /* PVC traffic */ - dev->stats.rx_bytes += skb->len; - if (pvc->state.becn) - dev->stats.rx_compressed++; - skb->dev = dev; - netif_rx(skb); - return NET_RX_SUCCESS; - } else { - goto rx_drop; - } + dev = skb->dev; + dev->stats.rx_packets++; /* PVC traffic */ + dev->stats.rx_bytes += skb->len; + if (pvc->state.becn) + dev->stats.rx_compressed++; + netif_rx(skb); + return NET_RX_SUCCESS; rx_error: frad->stats.rx_errors++; /* Mark error */
The eth_type_trans function is called when we receive frames carrying Ethernet frames. This function expects a non-NULL pointer as an argument, and assigns it directly to skb->dev. However, the code handling other types of frames first assigns a pointer to "dev", and then at the end checks whether the value is NULL, and if it is not NULL, assigns it to skb->dev. The two flows are different. Mixing them in this function makes the code messy. It's better that we convert the second flow to align with how eth_type_trans does things. So this patch changes the code to: first make sure the pointer is not NULL, then assign it directly to skb->dev. "dev" is no longer needed until the end where we use it to update stats. Cc: Krzysztof Halasa <khc@pm.waw.pl> Signed-off-by: Xie He <xie.he.0141@gmail.com> --- drivers/net/wan/hdlc_fr.c | 37 ++++++++++++++++++++----------------- 1 file changed, 20 insertions(+), 17 deletions(-)