tree: https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git staging-testing head: 08283d30744434d8f30d386622372e8f5b03bcf2 commit: 6cc5e1c700316c11b61975af3be8ebcab1e2f8b9 [159/337] staging: octeon: remove typedef declartion for cvmx_pko_command_word0 config: mips-cavium_octeon_defconfig (attached as .config) compiler: mips64-linux-gcc (GCC) 7.4.0 reproduce: wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross git checkout 6cc5e1c700316c11b61975af3be8ebcab1e2f8b9 # save the attached .config to linux build tree GCC_VERSION=7.4.0 make.cross ARCH=mips If you fix the issue, kindly add following tag Reported-by: kbuild test robot <lkp@xxxxxxxxx> All errors (new ones prefixed by >>): In file included from drivers/staging/octeon/ethernet-tx.c:25:0: drivers/staging/octeon/octeon-ethernet.h:94:12: warning: 'union cvmx_helper_link_info' declared inside parameter list will not be visible outside of this definition or declaration union cvmx_helper_link_info li); ^~~~~~~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c: In function 'cvm_oct_xmit': >> drivers/staging/octeon/ethernet-tx.c:130:31: error: storage size of 'pko_command' isn't known union cvmx_pko_command_word0 pko_command; ^~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:130:31: warning: unused variable 'pko_command' [-Wunused-variable] drivers/staging/octeon/ethernet-tx.c: In function 'cvm_oct_xmit_pow': drivers/staging/octeon/ethernet-tx.c:562:7: error: dereferencing pointer to incomplete type 'struct cvmx_wqe' work->word0.pip.cn38xx.hw_chksum = skb->csum; ^~ drivers/staging/octeon/ethernet-tx.c:564:20: error: passing argument 1 of 'cvmx_wqe_set_port' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_wqe_set_port(work, priv->port); ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-helper.h:39:0, from drivers/staging/octeon/octeon-ethernet.h:25, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-wqe.h:610:20: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline void cvmx_wqe_set_port(cvmx_wqe_t *work, int port) ^~~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:565:19: error: passing argument 1 of 'cvmx_wqe_set_qos' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_wqe_set_qos(work, priv->port & 0x7); ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-helper.h:39:0, from drivers/staging/octeon/octeon-ethernet.h:25, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-wqe.h:650:20: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline void cvmx_wqe_set_qos(cvmx_wqe_t *work, int qos) ^~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:566:19: error: passing argument 1 of 'cvmx_wqe_set_grp' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_wqe_set_grp(work, pow_send_group); ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-helper.h:39:0, from drivers/staging/octeon/octeon-ethernet.h:25, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-wqe.h:630:20: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline void cvmx_wqe_set_grp(cvmx_wqe_t *work, int grp) ^~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:649:26: error: passing argument 1 of 'cvmx_wqe_get_qos' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_wqe_get_qos(work), cvmx_wqe_get_grp(work)); ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-helper.h:39:0, from drivers/staging/octeon/octeon-ethernet.h:25, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-wqe.h:638:19: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline int cvmx_wqe_get_qos(cvmx_wqe_t *work) ^~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:649:50: error: passing argument 1 of 'cvmx_wqe_get_grp' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_wqe_get_qos(work), cvmx_wqe_get_grp(work)); ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-helper.h:39:0, from drivers/staging/octeon/octeon-ethernet.h:25, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-wqe.h:618:19: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline int cvmx_wqe_get_grp(cvmx_wqe_t *work) ^~~~~~~~~~~~~~~~ drivers/staging/octeon/ethernet-tx.c:648:23: error: passing argument 1 of 'cvmx_pow_work_submit' from incompatible pointer type [-Werror=incompatible-pointer-types] cvmx_pow_work_submit(work, work->word1.tag, work->word1.tag_type, ^~~~ In file included from arch/mips/include/asm/octeon/cvmx-pko.h:62:0, from drivers/staging/octeon/octeon-ethernet.h:31, from drivers/staging/octeon/ethernet-tx.c:25: arch/mips/include/asm/octeon/cvmx-pow.h:1806:20: note: expected 'cvmx_wqe_t * {aka struct <anonymous> *}' but argument is of type 'struct cvmx_wqe *' static inline void cvmx_pow_work_submit(cvmx_wqe_t *wqp, uint32_t tag, ^~~~~~~~~~~~~~~~~~~~ cc1: some warnings being treated as errors vim +130 drivers/staging/octeon/ethernet-tx.c 24 > 25 #include "octeon-ethernet.h" 26 #include "ethernet-defines.h" 27 #include "ethernet-tx.h" 28 #include "ethernet-util.h" 29 30 #define CVM_OCT_SKB_CB(skb) ((u64 *)((skb)->cb)) 31 32 /* 33 * You can define GET_SKBUFF_QOS() to override how the skbuff output 34 * function determines which output queue is used. The default 35 * implementation always uses the base queue for the port. If, for 36 * example, you wanted to use the skb->priority field, define 37 * GET_SKBUFF_QOS as: #define GET_SKBUFF_QOS(skb) ((skb)->priority) 38 */ 39 #ifndef GET_SKBUFF_QOS 40 #define GET_SKBUFF_QOS(skb) 0 41 #endif 42 43 static void cvm_oct_tx_do_cleanup(unsigned long arg); 44 static DECLARE_TASKLET(cvm_oct_tx_cleanup_tasklet, cvm_oct_tx_do_cleanup, 0); 45 46 /* Maximum number of SKBs to try to free per xmit packet. */ 47 #define MAX_SKB_TO_FREE (MAX_OUT_QUEUE_DEPTH * 2) 48 49 static inline int cvm_oct_adjust_skb_to_free(int skb_to_free, int fau) 50 { 51 int undo; 52 53 undo = skb_to_free > 0 ? MAX_SKB_TO_FREE : skb_to_free + 54 MAX_SKB_TO_FREE; 55 if (undo > 0) 56 cvmx_fau_atomic_add32(fau, -undo); 57 skb_to_free = -skb_to_free > MAX_SKB_TO_FREE ? MAX_SKB_TO_FREE : 58 -skb_to_free; 59 return skb_to_free; 60 } 61 62 static void cvm_oct_kick_tx_poll_watchdog(void) 63 { 64 union cvmx_ciu_timx ciu_timx; 65 66 ciu_timx.u64 = 0; 67 ciu_timx.s.one_shot = 1; 68 ciu_timx.s.len = cvm_oct_tx_poll_interval; 69 cvmx_write_csr(CVMX_CIU_TIMX(1), ciu_timx.u64); 70 } 71 72 static void cvm_oct_free_tx_skbs(struct net_device *dev) 73 { 74 int skb_to_free; 75 int qos, queues_per_port; 76 int total_freed = 0; 77 int total_remaining = 0; 78 unsigned long flags; 79 struct octeon_ethernet *priv = netdev_priv(dev); 80 81 queues_per_port = cvmx_pko_get_num_queues(priv->port); 82 /* Drain any pending packets in the free list */ 83 for (qos = 0; qos < queues_per_port; qos++) { 84 if (skb_queue_len(&priv->tx_free_list[qos]) == 0) 85 continue; 86 skb_to_free = cvmx_fau_fetch_and_add32(priv->fau + qos * 4, 87 MAX_SKB_TO_FREE); 88 skb_to_free = cvm_oct_adjust_skb_to_free(skb_to_free, 89 priv->fau + qos * 4); 90 total_freed += skb_to_free; 91 if (skb_to_free > 0) { 92 struct sk_buff *to_free_list = NULL; 93 94 spin_lock_irqsave(&priv->tx_free_list[qos].lock, flags); 95 while (skb_to_free > 0) { 96 struct sk_buff *t; 97 98 t = __skb_dequeue(&priv->tx_free_list[qos]); 99 t->next = to_free_list; 100 to_free_list = t; 101 skb_to_free--; 102 } 103 spin_unlock_irqrestore(&priv->tx_free_list[qos].lock, 104 flags); 105 /* Do the actual freeing outside of the lock. */ 106 while (to_free_list) { 107 struct sk_buff *t = to_free_list; 108 109 to_free_list = to_free_list->next; 110 dev_kfree_skb_any(t); 111 } 112 } 113 total_remaining += skb_queue_len(&priv->tx_free_list[qos]); 114 } 115 if (total_remaining < MAX_OUT_QUEUE_DEPTH && netif_queue_stopped(dev)) 116 netif_wake_queue(dev); 117 if (total_remaining) 118 cvm_oct_kick_tx_poll_watchdog(); 119 } 120 121 /** 122 * cvm_oct_xmit - transmit a packet 123 * @skb: Packet to send 124 * @dev: Device info structure 125 * 126 * Returns Always returns NETDEV_TX_OK 127 */ 128 int cvm_oct_xmit(struct sk_buff *skb, struct net_device *dev) 129 { > 130 union cvmx_pko_command_word0 pko_command; 131 union cvmx_buf_ptr hw_buffer; 132 u64 old_scratch; 133 u64 old_scratch2; 134 int qos; 135 int i; 136 enum {QUEUE_CORE, QUEUE_HW, QUEUE_DROP} queue_type; 137 struct octeon_ethernet *priv = netdev_priv(dev); 138 struct sk_buff *to_free_list; 139 int skb_to_free; 140 int buffers_to_free; 141 u32 total_to_clean; 142 unsigned long flags; 143 #if REUSE_SKBUFFS_WITHOUT_FREE 144 unsigned char *fpa_head; 145 #endif 146 147 /* 148 * Prefetch the private data structure. It is larger than the 149 * one cache line. 150 */ 151 prefetch(priv); 152 153 /* 154 * The check on CVMX_PKO_QUEUES_PER_PORT_* is designed to 155 * completely remove "qos" in the event neither interface 156 * supports multiple queues per port. 157 */ 158 if ((CVMX_PKO_QUEUES_PER_PORT_INTERFACE0 > 1) || 159 (CVMX_PKO_QUEUES_PER_PORT_INTERFACE1 > 1)) { 160 qos = GET_SKBUFF_QOS(skb); 161 if (qos <= 0) 162 qos = 0; 163 else if (qos >= cvmx_pko_get_num_queues(priv->port)) 164 qos = 0; 165 } else { 166 qos = 0; 167 } 168 169 if (USE_ASYNC_IOBDMA) { 170 /* Save scratch in case userspace is using it */ 171 CVMX_SYNCIOBDMA; 172 old_scratch = cvmx_scratch_read64(CVMX_SCR_SCRATCH); 173 old_scratch2 = cvmx_scratch_read64(CVMX_SCR_SCRATCH + 8); 174 175 /* 176 * Fetch and increment the number of packets to be 177 * freed. 178 */ 179 cvmx_fau_async_fetch_and_add32(CVMX_SCR_SCRATCH + 8, 180 FAU_NUM_PACKET_BUFFERS_TO_FREE, 181 0); 182 cvmx_fau_async_fetch_and_add32(CVMX_SCR_SCRATCH, 183 priv->fau + qos * 4, 184 MAX_SKB_TO_FREE); 185 } 186 187 /* 188 * We have space for 6 segment pointers, If there will be more 189 * than that, we must linearize. 190 */ 191 if (unlikely(skb_shinfo(skb)->nr_frags > 5)) { 192 if (unlikely(__skb_linearize(skb))) { 193 queue_type = QUEUE_DROP; 194 if (USE_ASYNC_IOBDMA) { 195 /* 196 * Get the number of skbuffs in use 197 * by the hardware 198 */ 199 CVMX_SYNCIOBDMA; 200 skb_to_free = 201 cvmx_scratch_read64(CVMX_SCR_SCRATCH); 202 } else { 203 /* 204 * Get the number of skbuffs in use 205 * by the hardware 206 */ 207 skb_to_free = 208 cvmx_fau_fetch_and_add32(priv->fau + 209 qos * 4, 210 MAX_SKB_TO_FREE); 211 } 212 skb_to_free = cvm_oct_adjust_skb_to_free(skb_to_free, 213 priv->fau + 214 qos * 4); 215 spin_lock_irqsave(&priv->tx_free_list[qos].lock, flags); 216 goto skip_xmit; 217 } 218 } 219 220 /* 221 * The CN3XXX series of parts has an errata (GMX-401) which 222 * causes the GMX block to hang if a collision occurs towards 223 * the end of a <68 byte packet. As a workaround for this, we 224 * pad packets to be 68 bytes whenever we are in half duplex 225 * mode. We don't handle the case of having a small packet but 226 * no room to add the padding. The kernel should always give 227 * us at least a cache line 228 */ 229 if ((skb->len < 64) && OCTEON_IS_MODEL(OCTEON_CN3XXX)) { 230 union cvmx_gmxx_prtx_cfg gmx_prt_cfg; 231 int interface = INTERFACE(priv->port); 232 int index = INDEX(priv->port); 233 234 if (interface < 2) { 235 /* We only need to pad packet in half duplex mode */ 236 gmx_prt_cfg.u64 = 237 cvmx_read_csr(CVMX_GMXX_PRTX_CFG(index, interface)); 238 if (gmx_prt_cfg.s.duplex == 0) { 239 int add_bytes = 64 - skb->len; 240 241 if ((skb_tail_pointer(skb) + add_bytes) <= 242 skb_end_pointer(skb)) 243 __skb_put_zero(skb, add_bytes); 244 } 245 } 246 } 247 248 /* Build the PKO command */ 249 pko_command.u64 = 0; 250 #ifdef __LITTLE_ENDIAN 251 pko_command.s.le = 1; 252 #endif 253 pko_command.s.n2 = 1; /* Don't pollute L2 with the outgoing packet */ 254 pko_command.s.segs = 1; 255 pko_command.s.total_bytes = skb->len; 256 pko_command.s.size0 = CVMX_FAU_OP_SIZE_32; 257 pko_command.s.subone0 = 1; 258 259 pko_command.s.dontfree = 1; 260 261 /* Build the PKO buffer pointer */ 262 hw_buffer.u64 = 0; 263 if (skb_shinfo(skb)->nr_frags == 0) { 264 hw_buffer.s.addr = XKPHYS_TO_PHYS((uintptr_t)skb->data); 265 hw_buffer.s.pool = 0; 266 hw_buffer.s.size = skb->len; 267 } else { 268 hw_buffer.s.addr = XKPHYS_TO_PHYS((uintptr_t)skb->data); 269 hw_buffer.s.pool = 0; 270 hw_buffer.s.size = skb_headlen(skb); 271 CVM_OCT_SKB_CB(skb)[0] = hw_buffer.u64; 272 for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) { 273 skb_frag_t *fs = skb_shinfo(skb)->frags + i; 274 275 hw_buffer.s.addr = 276 XKPHYS_TO_PHYS((uintptr_t)skb_frag_address(fs)); 277 hw_buffer.s.size = skb_frag_size(fs); 278 CVM_OCT_SKB_CB(skb)[i + 1] = hw_buffer.u64; 279 } 280 hw_buffer.s.addr = 281 XKPHYS_TO_PHYS((uintptr_t)CVM_OCT_SKB_CB(skb)); 282 hw_buffer.s.size = skb_shinfo(skb)->nr_frags + 1; 283 pko_command.s.segs = skb_shinfo(skb)->nr_frags + 1; 284 pko_command.s.gather = 1; 285 goto dont_put_skbuff_in_hw; 286 } 287 288 /* 289 * See if we can put this skb in the FPA pool. Any strange 290 * behavior from the Linux networking stack will most likely 291 * be caused by a bug in the following code. If some field is 292 * in use by the network stack and gets carried over when a 293 * buffer is reused, bad things may happen. If in doubt and 294 * you dont need the absolute best performance, disable the 295 * define REUSE_SKBUFFS_WITHOUT_FREE. The reuse of buffers has 296 * shown a 25% increase in performance under some loads. 297 */ 298 #if REUSE_SKBUFFS_WITHOUT_FREE 299 fpa_head = skb->head + 256 - ((unsigned long)skb->head & 0x7f); 300 if (unlikely(skb->data < fpa_head)) { 301 /* TX buffer beginning can't meet FPA alignment constraints */ 302 goto dont_put_skbuff_in_hw; 303 } 304 if (unlikely 305 ((skb_end_pointer(skb) - fpa_head) < CVMX_FPA_PACKET_POOL_SIZE)) { 306 /* TX buffer isn't large enough for the FPA */ 307 goto dont_put_skbuff_in_hw; 308 } 309 if (unlikely(skb_shared(skb))) { 310 /* TX buffer sharing data with someone else */ 311 goto dont_put_skbuff_in_hw; 312 } 313 if (unlikely(skb_cloned(skb))) { 314 /* TX buffer has been cloned */ 315 goto dont_put_skbuff_in_hw; 316 } 317 if (unlikely(skb_header_cloned(skb))) { 318 /* TX buffer header has been cloned */ 319 goto dont_put_skbuff_in_hw; 320 } 321 if (unlikely(skb->destructor)) { 322 /* TX buffer has a destructor */ 323 goto dont_put_skbuff_in_hw; 324 } 325 if (unlikely(skb_shinfo(skb)->nr_frags)) { 326 /* TX buffer has fragments */ 327 goto dont_put_skbuff_in_hw; 328 } 329 if (unlikely 330 (skb->truesize != 331 sizeof(*skb) + skb_end_offset(skb))) { 332 /* TX buffer truesize has been changed */ 333 goto dont_put_skbuff_in_hw; 334 } 335 336 /* 337 * We can use this buffer in the FPA. We don't need the FAU 338 * update anymore 339 */ 340 pko_command.s.dontfree = 0; 341 342 hw_buffer.s.back = ((unsigned long)skb->data >> 7) - 343 ((unsigned long)fpa_head >> 7); 344 345 *(struct sk_buff **)(fpa_head - sizeof(void *)) = skb; 346 347 /* 348 * The skbuff will be reused without ever being freed. We must 349 * cleanup a bunch of core things. 350 */ 351 dst_release(skb_dst(skb)); 352 skb_dst_set(skb, NULL); 353 skb_ext_reset(skb); 354 nf_reset_ct(skb); 355 --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx Intel Corporation
Attachment:
.config.gz
Description: application/gzip
_______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
