| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <201906292107.taC5EgR4%lkp@intel.com>
Date: Sat, 29 Jun 2019 21:54:48 +0800
From: kbuild test robot <lkp@...el.com>
To: Catherine Sullivan <csully@...gle.com>
Cc: kbuild-all@...org, netdev@...r.kernel.org,
Catherine Sullivan <csully@...gle.com>,
Sagi Shahar <sagis@...gle.com>,
Jon Olson <jonolson@...gle.com>,
Willem de Bruijn <willemb@...gle.com>,
Luigi Rizzo <lrizzo@...gle.com>
Subject: Re: [PATCH net-next v2 2/4] gve: Add transmit and receive support
Hi Catherine,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net-next/master]
url: https://github.com/0day-ci/linux/commits/Catherine-Sullivan/Add-gve-driver/20190629-143743
reproduce:
# apt-get install sparse
# sparse version: v0.6.1-rc1-7-g2b96cd8-dirty
make ARCH=x86_64 allmodconfig
make C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'
If you fix the issue, kindly add following tag
Reported-by: kbuild test robot <lkp@...el.com>
sparse warnings: (new ones prefixed by >>)
drivers/net/ethernet/google/gve/gve_rx.c:11:6: sparse: sparse: symbol 'gve_rx_remove_from_block' was not declared. Should it be static?
drivers/net/ethernet/google/gve/gve_rx.c:217:16: sparse: sparse: incorrect type in argument 1 (different base types) @@ expected unsigned int val @@ got restricted __be3unsigned int val @@
drivers/net/ethernet/google/gve/gve_rx.c:217:16: sparse: expected unsigned int val
drivers/net/ethernet/google/gve/gve_rx.c:217:16: sparse: got restricted __be32 [usertype]
drivers/net/ethernet/google/gve/gve_rx.c:349:27: sparse: sparse: incorrect type in assignment (different base types) @@ expected restricted __wsum [usertype] csum @@ got restricted __wsum [usertype] csum @@
drivers/net/ethernet/google/gve/gve_rx.c:349:27: sparse: expected restricted __wsum [usertype] csum
drivers/net/ethernet/google/gve/gve_rx.c:349:27: sparse: got restricted __be16 [usertype] csum
drivers/net/ethernet/google/gve/gve_rx.c:374:19: sparse: sparse: incorrect type in assignment (different base types) @@ expected unsigned short [usertype] flags_seq @@ got resunsigned short [usertype] flags_seq @@
drivers/net/ethernet/google/gve/gve_rx.c:374:19: sparse: expected unsigned short [usertype] flags_seq
drivers/net/ethernet/google/gve/gve_rx.c:374:19: sparse: got restricted __be16 [usertype] flags_seq
>> drivers/net/ethernet/google/gve/gve_rx.c:378:17: sparse: sparse: cast to restricted __be16
>> drivers/net/ethernet/google/gve/gve_rx.c:378:17: sparse: sparse: cast to restricted __be16
>> drivers/net/ethernet/google/gve/gve_rx.c:378:17: sparse: sparse: cast to restricted __be16
>> drivers/net/ethernet/google/gve/gve_rx.c:378:17: sparse: sparse: cast to restricted __be16
vim +378 drivers/net/ethernet/google/gve/gve_rx.c
212
213 void gve_rx_write_doorbell(struct gve_priv *priv, struct gve_rx_ring *rx)
214 {
215 u32 db_idx = be32_to_cpu(rx->q_resources->db_index);
216
> 217 writel(cpu_to_be32(rx->desc.fill_cnt), &priv->db_bar2[db_idx]);
218 }
219
220 static enum pkt_hash_types gve_rss_type(__be16 pkt_flags)
221 {
222 if (likely(pkt_flags & (GVE_RXF_TCP | GVE_RXF_UDP)))
223 return PKT_HASH_TYPE_L4;
224 if (pkt_flags & (GVE_RXF_IPV4 | GVE_RXF_IPV6))
225 return PKT_HASH_TYPE_L3;
226 return PKT_HASH_TYPE_L2;
227 }
228
229 static struct sk_buff *gve_rx_copy(struct net_device *dev,
230 struct napi_struct *napi,
231 struct gve_rx_slot_page_info *page_info,
232 u16 len)
233 {
234 struct sk_buff *skb = napi_alloc_skb(napi, len);
235 void *va = page_info->page_address + GVE_RX_PAD +
236 page_info->page_offset;
237
238 if (unlikely(!skb))
239 return NULL;
240
241 __skb_put(skb, len);
242
243 skb_copy_to_linear_data(skb, va, len);
244
245 skb->protocol = eth_type_trans(skb, dev);
246 return skb;
247 }
248
249 static struct sk_buff *gve_rx_add_frags(struct net_device *dev,
250 struct napi_struct *napi,
251 struct gve_rx_slot_page_info *page_info,
252 u16 len)
253 {
254 struct sk_buff *skb = napi_get_frags(napi);
255
256 if (unlikely(!skb))
257 return NULL;
258
259 skb_add_rx_frag(skb, 0, page_info->page,
260 page_info->page_offset +
261 GVE_RX_PAD, len, PAGE_SIZE / 2);
262
263 return skb;
264 }
265
266 static void gve_rx_flip_buff(struct gve_rx_slot_page_info *page_info,
267 struct gve_rx_data_slot *data_ring)
268 {
269 u64 addr = be64_to_cpu(data_ring->qpl_offset);
270
271 page_info->page_offset ^= PAGE_SIZE / 2;
272 addr ^= PAGE_SIZE / 2;
273 data_ring->qpl_offset = cpu_to_be64(addr);
274 }
275
276 static bool gve_rx(struct gve_rx_ring *rx, struct gve_rx_desc *rx_desc,
277 netdev_features_t feat)
278 {
279 struct gve_rx_slot_page_info *page_info;
280 struct gve_priv *priv = rx->gve;
281 struct napi_struct *napi = &priv->ntfy_blocks[rx->ntfy_id].napi;
282 struct net_device *dev = priv->dev;
283 struct sk_buff *skb;
284 int pagecount;
285 u16 len;
286 u32 idx;
287
288 /* drop this packet */
289 if (unlikely(rx_desc->flags_seq & GVE_RXF_ERR))
290 return true;
291
292 len = be16_to_cpu(rx_desc->len) - GVE_RX_PAD;
293 idx = rx->data.cnt & rx->data.mask;
294 page_info = &rx->data.page_info[idx];
295
296 /* gvnic can only receive into registered segments. If the buffer
297 * can't be recycled, our only choice is to copy the data out of
298 * it so that we can return it to the device.
299 */
300
301 #if PAGE_SIZE == 4096
302 if (len <= priv->rx_copybreak) {
303 /* Just copy small packets */
304 skb = gve_rx_copy(dev, napi, page_info, len);
305 goto have_skb;
306 }
307 if (unlikely(!gve_can_recycle_pages(dev))) {
308 skb = gve_rx_copy(dev, napi, page_info, len);
309 goto have_skb;
310 }
311 pagecount = page_count(page_info->page);
312 if (pagecount == 1) {
313 /* No part of this page is used by any SKBs; we attach
314 * the page fragment to a new SKB and pass it up the
315 * stack.
316 */
317 skb = gve_rx_add_frags(dev, napi, page_info, len);
318 if (!skb)
319 return true;
320 /* Make sure the kernel stack can't release the page */
321 get_page(page_info->page);
322 /* "flip" to other packet buffer on this page */
323 gve_rx_flip_buff(page_info, &rx->data.data_ring[idx]);
324 } else if (pagecount >= 2) {
325 /* We have previously passed the other half of this
326 * page up the stack, but it has not yet been freed.
327 */
328 skb = gve_rx_copy(dev, napi, page_info, len);
329 } else {
330 WARN(pagecount < 1, "Pagecount should never be < 1");
331 return false;
332 }
333 #else
334 skb = gve_rx_copy(dev, napi, page_info, len);
335 #endif
336
337 have_skb:
338 if (!skb)
339 return true;
340
341 rx->data.cnt++;
342
343 if (likely(feat & NETIF_F_RXCSUM)) {
344 /* NIC passes up the partial sum */
345 if (rx_desc->csum)
346 skb->ip_summed = CHECKSUM_COMPLETE;
347 else
348 skb->ip_summed = CHECKSUM_NONE;
349 skb->csum = rx_desc->csum;
350 }
351
352 /* parse flags & pass relevant info up */
353 if (likely(feat & NETIF_F_RXHASH) &&
354 gve_needs_rss(rx_desc->flags_seq))
355 skb_set_hash(skb, be32_to_cpu(rx_desc->rss_hash),
356 gve_rss_type(rx_desc->flags_seq));
357
358 if (skb_is_nonlinear(skb))
359 napi_gro_frags(napi);
360 else
361 napi_gro_receive(napi, skb);
362 return true;
363 }
364
365 static bool gve_rx_work_pending(struct gve_rx_ring *rx)
366 {
367 struct gve_rx_desc *desc;
368 u16 flags_seq;
369 u32 next_idx;
370
371 next_idx = rx->desc.cnt & rx->desc.mask;
372 desc = rx->desc.desc_ring + next_idx;
373
374 flags_seq = desc->flags_seq;
375 /* Make sure we have synchronized the seq no with the device */
376 smp_rmb();
377
> 378 return (GVE_SEQNO(flags_seq) == rx->desc.seqno);
379 }
380
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
Powered by blists - more mailing lists