TODO.AR7

   1 TODO list for AR7 / QEMU
   2
   3 Hello Qemu Developers.
   4
   5 Here at work we're implementing qemu mcast networking while using avahi-autoipd to obtain network address. The implementation of qemu networking is the following:
   6
   7 -net nic,vlan=1,macaddr=52:54:00:12:34:01,model=virtio -net socket,vlan=1,mcast=230.0.0.1:1234
   8
   9 So, while running avahi-autoipd we get his messages:
  10
  11 [root@localhost ~]# avahi-autoipd eth1
  12 Found user 'avahi-autoipd' (UID 499) and group 'avahi-autoipd' (GID 499).
  13 Successfully called chroot().
  14 Successfully dropped root privileges.
  15 Starting with address 169.254.2.176
  16 Callout BIND, address 169.254.2.176 on interface eth1
  17 Received conflicting normal ARP packet.
  18 Callout CONFLICT, address 169.254.2.176 on interface eth1
  19 Trying address 169.254.189.32
  20 Callout BIND, address 169.254.189.32 on interface eth1
  21 Received conflicting normal ARP packet.
  22
  23 After sniffing ARP packets, we've realized that we're getting ARP packets back, which screws with autoipd. I think this is due to the mcast implementation not doing RPF. Is your mcast implementation supposed to do RPF?? If it is, we've got ourselves a bug.
  24
  25 Thanks for your help,
  26
  27 Look forward to hear from you.
  28
  29
  30 --
  31 Andres E. Rodriguez Lazo (4nDr3s - RoAkSoAx)
  32 Ubuntu Counter Project - User # 17582
  33
  34
  35 2009-01-30
  36
  37 High network load crashs Qemu (tested with malta and ar7).
  38 Reverting change r6288 (from official Qemu trunk) fixes this.
  39
  40 #0  ip_reass (ip=0x1bfa800, fp=0x1874050) at /home/stefan/src/qemu/trunk/slirp/ip_input.c:403
  41 #1  0x00000000004cca01 in ip_input (m=0x1bfa7b0) at /home/stefan/src/qemu/trunk/slirp/ip_input.c:224
  42 #2  0x00000000004aaba5 in slirp_input (pkt=0x1977bd0 "RT", pkt_len=1294) at /home/stefan/src/qemu/trunk/slirp/slirp.c:679
  43 #3  0x000000000049295d in qemu_send_packet (vc1=0x177a480, buf=0x1977bd0 "RT", size=1294) at /home/stefan/src/qemu/trunk/net.c:399
  44 #4  0x000000000042bc74 in pcnet_transmit (s=0x1977810) at /home/stefan/src/qemu/trunk/hw/pcnet.c:1300
  45 #5  0x000000000042c190 in pcnet_poll_timer (opaque=0x1977810) at /home/stefan/src/qemu/trunk/hw/pcnet.c:1363
  46 #6  0x000000000042ca79 in pcnet_ioport_writew (opaque=0x1977810, addr=4146, val=0) at /home/stefan/src/qemu/trunk/hw/pcnet.c:1645
  47 #7  0x0000000040632fb5 in ?? ()
  48 #8  0x0000000000000000 in ?? ()
  49
  50 2008-12-16
  51
  52 AR7_RESET_PEREPHERIAL
  53
  54 * Linux 2.6.27.9 working, but some locking issues with acx driver:
  55
  56 [ 9424.190000] get_mask 0x00000000, set_mask 0x00000040
  57 [ 9424.200000] setting RXconfig to 2010:0FDD
  58 [ 9424.220000] get_mask 0x00000000, set_mask 0x00000000 - after update
  59 [ 9424.250000]
  60 [ 9424.250000] =======================================================
  61 [ 9424.250000] [ INFO: possible circular locking dependency detected ]
  62 [ 9424.250000] 2.6.27.9 #3
  63 [ 9424.250000] -------------------------------------------------------
  64 [ 9424.250000] ifconfig/410 is trying to acquire lock:
  65 [ 9424.250000]  (events){--..}, at: [<94044984>] flush_workqueue+0x0/0x80
  66 [ 9424.250000]
  67 [ 9424.250000] but task is already holding lock:
  68 [ 9424.250000]  (rtnl_mutex){--..}, at: [<941f907c>] devinet_ioctl+0x160/0x858
  69 [ 9424.250000]
  70 [ 9424.250000] which lock already depends on the new lock.
  71 [ 9424.250000]
  72 [ 9424.250000]
  73 [ 9424.250000] the existing dependency chain (in reverse order) is:
  74 [ 9424.250000]
  75 [ 9424.250000] -> #2 (rtnl_mutex){--..}:
  76 [ 9424.250000]        [<9405bc8c>] validate_chain+0xb84/0xee0
  77 [ 9424.250000]        [<9405c7d8>] __lock_acquire+0x7f0/0x89c
  78 [ 9424.250000]        [<9405c8f4>] lock_acquire+0x70/0xb0
  79 [ 9424.250000]        [<940078c8>] mutex_lock_nested+0x114/0x498
  80 [ 9424.250000]        [<941a81c8>] linkwatch_event+0x10/0x38
  81 [ 9424.250000]        [<94043fa0>] run_workqueue+0x1a8/0x2bc
  82 [ 9424.250000]        [<94044ff0>] worker_thread+0xac/0xcc
  83 [ 9424.250000]        [<94048f84>] kthread+0x58/0x98
  84 [ 9424.250000]        [<9400c4d4>] kernel_thread_helper+0x10/0x18
  85 [ 9424.250000]
  86 [ 9424.250000] -> #1 ((linkwatch_work).work){--..}:
  87 [ 9424.250000]        [<9405bc8c>] validate_chain+0xb84/0xee0
  88 [ 9424.250000]        [<9405c7d8>] __lock_acquire+0x7f0/0x89c
  89 [ 9424.250000]        [<9405c8f4>] lock_acquire+0x70/0xb0
  90 [ 9424.250000]        [<94043f98>] run_workqueue+0x1a0/0x2bc
  91 [ 9424.250000]        [<94044ff0>] worker_thread+0xac/0xcc
  92 [ 9424.250000]        [<94048f84>] kthread+0x58/0x98
  93 [ 9424.250000]        [<9400c4d4>] kernel_thread_helper+0x10/0x18
  94 [ 9424.250000]
  95 [ 9424.250000] -> #0 (events){--..}:
  96 [ 9424.250000]        [<9405b8e8>] validate_chain+0x7e0/0xee0
  97 [ 9424.250000]        [<9405c7d8>] __lock_acquire+0x7f0/0x89c
  98 [ 9424.250000]        [<9405c8f4>] lock_acquire+0x70/0xb0
  99 [ 9424.250000]        [<940449d8>] flush_workqueue+0x54/0x80
 100 [ 9424.250000]        [<c0046c90>] acxpci_l_get_txbuf+0x138/0x2c4 [acx]
 101 [ 9424.250000]
 102 [ 9424.250000] other info that might help us debug this:
 103 [ 9424.250000]
 104 [ 9424.250000] 1 lock held by ifconfig/410:
 105 [ 9424.250000]  #0:  (rtnl_mutex){--..}, at: [<941f907c>] devinet_ioctl+0x160/0x858
 106 [ 9424.250000]
 107 [ 9424.250000] stack backtrace:
 108 [ 9424.250000] Call Trace:
 109 [ 9424.250000] [<940111a8>] dump_stack+0x8/0x34
 110 [ 9424.250000] [<9405b0f0>] print_circular_bug_tail+0xd4/0xec
 111 [ 9424.250000] [<9405b8e8>] validate_chain+0x7e0/0xee0
 112 [ 9424.250000] [<9405c7d8>] __lock_acquire+0x7f0/0x89c
 113 [ 9424.250000] [<9405c8f4>] lock_acquire+0x70/0xb0
 114 [ 9424.250000] [<940449d8>] flush_workqueue+0x54/0x80
 115 [ 9424.250000] [<c0046c90>] acxpci_l_get_txbuf+0x138/0x2c4 [acx]
 116 [ 9424.250000]
 117 [ 9424.520000] acx_set_status(0):STOPPED
 118 [ 9424.520000] closed device
 119 root@OpenWrt:/# iwlist wlan0 scan
 120 wlan0     Interface doesn't support scanning : Resource temporarily unavailable
 121
 122 * very poor performance on cpmac:
 123   eth0 sinus-heipc84499         1,4 MBit / s
 124   eth0 sinus-heipc84499         2,1 MBit / s    linux-2.6.27.21 preemptive
 125   eth0 sinus-heipc84499         2,2 MBit / s    linux-2.6.27.21
 126   eth0 fritzbox-heipc84499      2,9 MBit / s    2009-05-10
 127   eth0 heipc84499-sinus        23,6 MBit / s
 128   eth0 heipc84499-sinus        32,3 MBit / s    linux-2.6.27.21 preemptive
 129   eth0 heipc84499-sinus        34,2 MBit / s    linux-2.6.27.21
 130   eth0 heipc84499-fritzbox     41,9 MBit / s    2009-05-10
 131   lo fritzbox-fritzbox         59,0 MBit / s    2009-05-10
 132   lo sinus-sinus               49,0 MBit / s
 133   lo sinus-sinus               51,7 MBit / s    linux-2.6.27.21 preemptive
 134   lo sinus-sinus               55,2 MBit / s    linux-2.6.27.21
 135   wlan0 fritzbox-flocke         0,6 MBit / s    2009-05-10
 136   wlan0 flocke-fritzbox        10,7 MBit / s    2009-05-10
 137   wlan0 heipc84499-flocke       3,6 MBit / s
 138   wlan0 heipc84499-flocke      14,5 MBit / s    2009-04-29
 139
 140 root@OpenWrt:/# netserver
 141 root@OpenWrt:/# netperf -H 192.168.2.2
 142 TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.2.2 (192.168.2.2) port 0 AF_INET
 143 Recv   Send    Send
 144 Socket Socket  Message  Elapsed
 145 Size   Size    Size     Time     Throughput
 146 bytes  bytes   bytes    secs.    10^6bits/sec
 147
 148    256  16384  16384    10.02       1.40
 149    256  16384  16384    10.01       2.07        2008-04-26, linux-2.6.27.21
 150  87380  16384  16384    10.04      32.29
 151  43689  16384  16384    10.01      57.00        Linux version 2.6.13.1-ar7 AVM localhost
 152    256  16384  16384    10.06       2.91
 153
 154 2008-07-14
 155
 156 * Linux 2.6.26 working.
 157
 158 2008-04-01
 159
 160 * Switched to new QEMU code using TCG. Some changes in ar7.c were needed,
 161   others may still be missing, so this revision is highly experimental.
 162
 163 2008-03-16
 164
 165 * USB Treiber für TNETV115PAP
 166
 167 * Absturz CPMAC unter hoher Dauerlast
 168
 169 94139fd0:       8c450000        lw      a1,0(v0)
 170 94139fd4:       26100004        addiu   s0,s0,4
 171 94139fd8:       0d00927f        jal     940249fc <printk>
 172 94139fdc:       24842f24        addiu   a0,a0,12068
 173 94139fe0:       2a020680        slti    v0,s0,1664
 174 94139fe4:       3203000f        andi    v1,s0,0xf
 175 94139fe8:       1440ffe9        bnez    v0,94139f90 <cpmac_irq+0x4c0>
 176 94139fec:       02009021        move    s2,s0
 177 94139ff0:       3c049425        lui     a0,0x9425
 178 94139ff4:       0d00927f        jal     940249fc <printk>
 179 94139ff8:       2484d7a4        addiu   a0,a0,-10332
 180 94139ffc:       8e620060        lw      v0,96(s3)
 181 9413a000:       26650058        addiu   a1,s3,88
 182 9413a004:       38420001        xori    v0,v0,0x1
 183 9413a008:       30420001        andi    v0,v0,0x1
 184 9413a00c:       00028036        tne     zero,v0,0x200                   <= BUG_ON
 185 9413a010:       3c029429        lui     v0,0x9429
 186 9413a014:       8c427528        lw      v0,29992(v0)
 187 9413a018:       3c030001        lui     v1,0x1
 188 9413a01c:       00431024        and     v0,v0,v1
 189 9413a020:       10400009        beqz    v0,9413a048 <cpmac_irq+0x578>
 190 9413a024:       00000000        nop
 191 9413a028:       c0a20008        ll      v0,8(a1)
 192 9413a02c:       2401fffe        li      at,-2
 193 9413a030:       00411024        and     v0,v0,at
 194 9413a034:       e0a20008        sc      v0,8(a1)
 195
 196 Cpu 0
 197 $ 0   : 00000000 10008401 00000001 a8610000
 198 $ 4   : 0000001b 9435c3d8 00000001 94307b2c
 199 $ 8   : 10008400 1000001f 94a47194 0000006c
 200 $12   : 94b3f8cc 00000000 00000000 00000000
 201 $16   : 00000000 9435c000 00000000 9435c380
 202 $20   : 00000000 00010000 0000000f 94307b2c
 203 $24   : 00000000 94000b5c
 204 $28   : 94918000 94919848 0003017a 94051400
 205 Hi    : 00007bda
 206 Lo    : e1481a74
 207 epc   : 9413a00c cpmac_irq+0x53c/0x5f8     Not tainted
 208 ra    : 94051400 handle_IRQ_event+0x64/0xd4
 209 Status: 10008403    KERNEL EXL IE
 210 Cause : 30800034
 211 PrId  : 00018448 (MIPS 4KEc)
 212 Modules linked in:
 213 Process dd (pid: 362, threadinfo=94918000, task=94329908)
 214 Stack : 00000000 943bdd60 94d1bd80 94191ef4 94d1bd80 943bdd60 9438e900 00000000
 215         00000000 0000001b 00000010 94315560 94051400 0000006c 94e0ea8c 00000000
 216         9435c000 94191dc0 94268eb8 0000001b 0000000a 9438d81c 9405339c 943bdd60
 217         94d1a824 943bdee4 941b580c 00000010 00000013 94919ab0 94000af4 00000000
 218         0102a8c0 00000000 943bdd60 00000000 949199e0 94001b84 0102a8c0 941b6c90
 219         ...
 220 Call Trace:
 221 [<9413a00c>] cpmac_irq+0x53c/0x5f8
 222 [<94051400>] handle_IRQ_event+0x64/0xd4
 223 [<9405339c>] handle_level_irq+0xa4/0x118
 224 [<94000af4>] plat_irq_dispatch+0x98/0x100
 225 [<94001b84>] ret_from_irq+0x0/0x4
 226 [<9402a4c0>] __do_softirq+0x44/0x100
 227 [<9402a5d8>] do_softirq+0x5c/0x94
 228 [<9402ab54>] irq_exit+0x40/0x8c
 229 [<94001b84>] ret_from_irq+0x0/0x4
 230 [<941e7b00>] xs_send_kvec+0x6c/0xd0
 231 [<941e7c84>] xs_sendpages+0x120/0x2c8
 232 [<941ea130>] xs_udp_send_request+0x50/0x134
 233 [<941e6878>] xprt_transmit+0xc8/0x208
 234 [<941e3284>] call_transmit+0x278/0x2e4
 235 [<941ebf00>] __rpc_execute+0x94/0x4a4
 236 [<940e8264>] nfs_execute_read+0x40/0x64
 237 [<940e8500>] nfs_pagein_one+0x148/0x224
 238 [<940e5a14>] nfs_pageio_doio+0x44/0x88
 239 [<940e5b6c>] nfs_pageio_add_request+0x114/0x138
 240 [<940e8cb8>] readpage_async_filler+0x1a4/0x264
 241 [<9405fec0>] read_cache_pages+0xa0/0xf0
 242 [<940e8e94>] nfs_readpages+0x11c/0x190
 243 [<9405fca4>] __do_page_cache_readahead+0x1f4/0x2b8
 244 [<94060298>] page_cache_sync_readahead+0x28/0x34
 245 [<94055ae0>] do_generic_mapping_read+0x120/0x604
 246 [<94058244>] generic_file_aio_read+0x16c/0x1e0
 247 [<94080de0>] do_sync_read+0xc4/0x128
 248 [<94081bf4>] sys_read+0x54/0x98
 249 [<9400a390>] stack_done+0x20/0x3c
 250
 251
 252 Code: 26650058  38420001  30420001 <00028036> 3c029429  8c427528  3c030001  00431024  10400009
 253 Kernel panic - not syncing: Fatal exception in interrupt
 254
 255 * arch/mips/ar7/prom.c:205 psp_env=afb10084 Zugriff führt auf HW zum Absturz!!!
 256
 257 * cache_parity_error in Vektortabelle mit 0x8400xxxx (falsch) statt 0x9400xxxx
 258
 259 2007-10-30
 260 * cmos rtc register a is constant when emulating mips in mips.
 261
 262 2006-08-29
 263 * Wanted: network emulation for more chipsets:
 264         Intel 8255xER PCI (eepro100),
 265                 Intel 82559ER Fast Ethernet PCI Controller
 266                         ftp://download.intel.com/design/network/datashts/71468206.pdf
 267                 Intel 82559 Fast Ethernet Multifunction PCI/CardBus Controller
 268                         ftp://download.intel.com/design/network/datashts/73825902.pdf
 269
 270         Intel PRO/100 VE (? linux driver),
 271         E100B325, E100ENT (? linux driver),
 272         Realtek RTL8168/8111 PCI (r1000),
 273                 ftp://202.65.194.18/cn/nic/rtl8111brtl8168b/spec-8168b(110).pdf
 274
 275         National Semiconductor DP83815/DP83816 (natsemi)
 276
 277 2006-06-12
 278
 279 * AVM Linux kernel does not get interrupt from CP0_Count.
 280   It runs in an endless loop ("Calibrating delay loop...). Fixed.
 281
 282 * Fix handling of CP0_Count. Must not count if not told to do so.
 283
 284 * cpmac emulation for QEMU is still very incomplete.
 285
 286 * TI cpmac driver has bad performace (AVM: 30 Mbit/s from box,
 287   50 Mbit/s to box), could be close to 90 MBit/s.
 288
 289 2006-05-23
 290
 291 * serial ports in QEMU are 16450, should be 16550A.
 292   Patch exists, but decreases the performance.