When a tx timeout occured, et131x_tx_timeout closed and re-opened the device to fix.
As et131x_close called free_irq(), bad things ensued (see warning trace below), namely a
storm of errors and warnings.
Fixed by replacing the close() and open() calls with just the relevant functions previously
called from these.
Verified on an ET-1310 device.
Signed-off-by: Mark Einon <mark.einon@xxxxxxxxx>
----------
Aug 2 21:26:08 msilap kernel: [ 6484.816024] ------------[ cut here ]------------
Aug 2 21:26:08 msilap kernel: [ 6484.816039] WARNING: at /home/mark/Source/staging-2.6/net/sched/sch_generic.c:255 dev_watchdog+0xf0/0x150()
Aug 2 21:26:08 msilap kernel: [ 6484.816045] Hardware name: MS-1727
Aug 2 21:26:08 msilap kernel: [ 6484.816050] NETDEV WATCHDOG: eth1 (et131x): transmit queue 0 timed out
Aug 2 21:26:08 msilap kernel: [ 6484.816054] Modules linked in: et131x(C) aes_generic tcp_lp fuse nouveau ttm drm_kms_helper drm i2c_algo_bit sunrpc ipv6 cpufreq_ondemand acpi_cpufreq freq_table mperf uinput arc4 iwlagn snd_hda_codec_hdmi mac80211 sdhci_pci sdhci snd_hda_codec_realtek cfg80211 snd_hda_intel snd_hda_codec mmc_core snd_hwdep firewire_ohci firewire_core jmb38x_ms snd_seq snd_seq_device snd_pcm mxm_wmi ir_lirc_codec lirc_dev ir_sony_decoder i2c_i801 ir_jvc_decoder ir_rc6_decoder rc_rc6_mce i7core_edac edac_core snd_timer pcspkr rfkill r8169 mii i2c_core video ir_rc5_decoder memstick iTCO_wdt iTCO_vendor_support ir_nec_decoder ene_ir rc_core joydev microcode wmi snd crc_itu_t soundcore snd_page_alloc [last unloaded: et131x]
Aug 2 21:26:08 msilap kernel: [ 6484.816163] Pid: 0, comm: kworker/0:1 Tainted: G C 3.0.0-rc6+ #21
Aug 2 21:26:08 msilap kernel: [ 6484.816167] Call Trace:
Aug 2 21:26:08 msilap kernel: [ 6484.816171] <IRQ> [<ffffffff8103caf4>] warn_slowpath_common+0x85/0x9d
Aug 2 21:26:08 msilap kernel: [ 6484.816189] [<ffffffff8103cbaf>] warn_slowpath_fmt+0x46/0x48
Aug 2 21:26:08 msilap kernel: [ 6484.816196] [<ffffffff813aa6fb>] ? netif_tx_lock+0x4a/0x7b
Aug 2 21:26:08 msilap kernel: [ 6484.816203] [<ffffffff813aa86f>] dev_watchdog+0xf0/0x150
Aug 2 21:26:08 msilap kernel: [ 6484.816211] [<ffffffff810497b6>] run_timer_softirq+0x1a9/0x279
Aug 2 21:26:08 msilap kernel: [ 6484.816218] [<ffffffff813aa77f>] ? netif_tx_unlock+0x53/0x53
Aug 2 21:26:08 msilap kernel: [ 6484.816226] [<ffffffff810429d7>] __do_softirq+0xd5/0x1a4
Aug 2 21:26:08 msilap kernel: [ 6484.816234] [<ffffffff810079ed>] ? paravirt_read_tsc+0x9/0xd
Aug 2 21:26:08 msilap kernel: [ 6484.816245] [<ffffffff8144561c>] call_softirq+0x1c/0x26
Aug 2 21:26:08 msilap kernel: [ 6484.816251] [<ffffffff81003b9f>] do_softirq+0x46/0x83
Aug 2 21:26:08 msilap kernel: [ 6484.816257] [<ffffffff81042ca2>] irq_exit+0x52/0x9b
Aug 2 21:26:08 msilap kernel: [ 6484.816265] [<ffffffff81445749>] smp_apic_timer_interrupt+0x7c/0x8a
Aug 2 21:26:08 msilap kernel: [ 6484.816273] [<ffffffff814450d3>] apic_timer_interrupt+0x13/0x20
Aug 2 21:26:08 msilap kernel: [ 6484.816278] <EOI> [<ffffffff810079ed>] ? paravirt_read_tsc+0x9/0xd
Aug 2 21:26:08 msilap kernel: [ 6484.816291] [<ffffffff8124ad87>] ? intel_idle+0xd1/0xf8
Aug 2 21:26:08 msilap kernel: [ 6484.816297] [<ffffffff8124ad69>] ? intel_idle+0xb3/0xf8
Aug 2 21:26:08 msilap kernel: [ 6484.816306] [<ffffffff8136a767>] cpuidle_idle_call+0xe2/0x160
Aug 2 21:26:08 msilap kernel: [ 6484.816315] [<ffffffff81001293>] cpu_idle+0xaa/0xcc
Aug 2 21:26:08 msilap kernel: [ 6484.816323] [<ffffffff81436b4e>] start_secondary+0x248/0x24f
Aug 2 21:26:08 msilap kernel: [ 6484.816329] ---[ end trace 10ae1b2c6bae932f ]---
Aug 2 21:26:08 msilap kernel: [ 6484.816337] et131x 0000:02:00.0: Send stuck - reset. tcb->WrIndex 0, flags 0x00000000
Aug 2 21:26:08 msilap kernel: [ 6484.816344] ------------[ cut here ]------------
Aug 2 21:26:08 msilap kernel: [ 6484.816353] WARNING: at /home/mark/Source/staging-2.6/kernel/irq/manage.c:1131 __free_irq+0x58/0x192()
Aug 2 21:26:08 msilap kernel: [ 6484.816358] Hardware name: MS-1727
Aug 2 21:26:08 msilap kernel: [ 6484.816362] Trying to free IRQ 16 from IRQ context!
Aug 2 21:26:08 msilap kernel: [ 6484.816365] Modules linked in: et131x(C) aes_generic tcp_lp fuse nouveau ttm drm_kms_helper drm i2c_algo_bit sunrpc ipv6 cpufreq_ondemand acpi_cpufreq freq_table mperf uinput arc4 iwlagn snd_hda_codec_hdmi mac80211 sdhci_pci sdhci snd_hda_codec_realtek cfg80211 snd_hda_intel snd_hda_codec mmc_core snd_hwdep firewire_ohci firewire_core jmb38x_ms snd_seq snd_seq_device snd_pcm mxm_wmi ir_lirc_codec lirc_dev ir_sony_decoder i2c_i801 ir_jvc_decoder ir_rc6_decoder rc_rc6_mce i7core_edac edac_core snd_timer pcspkr rfkill r8169 mii i2c_core video ir_rc5_decoder memstick iTCO_wdt iTCO_vendor_support ir_nec_decoder ene_ir rc_core joydev microcode wmi snd crc_itu_t soundcore snd_page_alloc [last unloaded: et131x]
Aug 2 21:26:08 msilap kernel: [ 6484.816459] Pid: 0, comm: kworker/0:1 Tainted: G WC 3.0.0-rc6+ #21
Aug 2 21:26:08 msilap kernel: [ 6484.816464] Call Trace:
Aug 2 21:26:08 msilap kernel: [ 6484.816467] <IRQ> [<ffffffff8103caf4>] warn_slowpath_common+0x85/0x9d
Aug 2 21:26:08 msilap kernel: [ 6484.816480] [<ffffffff8103cbaf>] warn_slowpath_fmt+0x46/0x48
Aug 2 21:26:08 msilap kernel: [ 6484.816488] [<ffffffff81092276>] __free_irq+0x58/0x192
Aug 2 21:26:08 msilap kernel: [ 6484.816496] [<ffffffff8109240e>] free_irq+0x5e/0x77
Aug 2 21:26:08 msilap kernel: [ 6484.816506] [<ffffffffa00bf58e>] et131x_close+0x4b/0x5e [et131x]
Aug 2 21:26:08 msilap kernel: [ 6484.816516] [<ffffffffa00bf64d>] et131x_tx_timeout+0xac/0xc7 [et131x]
Aug 2 21:26:08 msilap kernel: [ 6484.816523] [<ffffffff813aa883>] dev_watchdog+0x104/0x150
Aug 2 21:26:08 msilap kernel: [ 6484.816530] [<ffffffff810497b6>] run_timer_softirq+0x1a9/0x279
Aug 2 21:26:08 msilap kernel: [ 6484.816537] [<ffffffff813aa77f>] ? netif_tx_unlock+0x53/0x53
Aug 2 21:26:08 msilap kernel: [ 6484.816544] [<ffffffff810429d7>] __do_softirq+0xd5/0x1a4
Aug 2 21:26:08 msilap kernel: [ 6484.816552] [<ffffffff810079ed>] ? paravirt_read_tsc+0x9/0xd
Aug 2 21:26:08 msilap kernel: [ 6484.816560] [<ffffffff8144561c>] call_softirq+0x1c/0x26
Aug 2 21:26:08 msilap kernel: [ 6484.816566] [<ffffffff81003b9f>] do_softirq+0x46/0x83
Aug 2 21:26:08 msilap kernel: [ 6484.816572] [<ffffffff81042ca2>] irq_exit+0x52/0x9b
Aug 2 21:26:08 msilap kernel: [ 6484.816580] [<ffffffff81445749>] smp_apic_timer_interrupt+0x7c/0x8a
Aug 2 21:26:08 msilap kernel: [ 6484.816588] [<ffffffff814450d3>] apic_timer_interrupt+0x13/0x20
Aug 2 21:26:08 msilap kernel: [ 6484.816592] <EOI> [<ffffffff810079ed>] ? paravirt_read_tsc+0x9/0xd
Aug 2 21:26:08 msilap kernel: [ 6484.816604] [<ffffffff8124ad87>] ? intel_idle+0xd1/0xf8
Aug 2 21:26:08 msilap kernel: [ 6484.816610] [<ffffffff8124ad69>] ? intel_idle+0xb3/0xf8
Aug 2 21:26:08 msilap kernel: [ 6484.816617] [<ffffffff8136a767>] cpuidle_idle_call+0xe2/0x160
Aug 2 21:26:08 msilap kernel: [ 6484.816625] [<ffffffff81001293>] cpu_idle+0xaa/0xcc
Aug 2 21:26:08 msilap kernel: [ 6484.816632] [<ffffffff81436b4e>] start_secondary+0x248/0x24f
---
drivers/staging/et131x/et1310_phy.c | 1 -
drivers/staging/et131x/et131x_netdev.c | 23 +++++++++++++++++++++--
2 files changed, 21 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/et131x/et1310_phy.c b/drivers/staging/et131x/et1310_phy.c
index 2c6898f..c845c27 100644
--- a/drivers/staging/et131x/et1310_phy.c
+++ b/drivers/staging/et131x/et1310_phy.c
@@ -732,7 +732,6 @@ void et131x_mii_check(struct et131x_adapter *adapter,
u32 mdi_mdix;
u32 masterslave;
u32 polarity;
- unsigned long flags;
if (bmsr_ints & MI_BMSR_LINK_STATUS) {
if (bmsr & MI_BMSR_LINK_STATUS) {
diff --git a/drivers/staging/et131x/et131x_netdev.c b/drivers/staging/et131x/et131x_netdev.c
index 37e1069..70a0da3 100644
--- a/drivers/staging/et131x/et131x_netdev.c
+++ b/drivers/staging/et131x/et131x_netdev.c
@@ -478,9 +478,28 @@ void et131x_tx_timeout(struct net_device *netdev)
tcb->index,
tcb->flags);
- et131x_close(netdev);
- et131x_open(netdev);
+ adapter->net_stats.tx_errors++;
+ /* perform reset */
+ /* First thing is to stop the queue */
+ netif_stop_queue(netdev);
+
+ /* Stop the Tx and Rx DMA engines */
+ et131x_rx_dma_disable(adapter);
+ et131x_tx_dma_disable(adapter);
+
+ /* Disable device interrupts */
+ et131x_disable_interrupts(adapter);
+
+ /* Enable the Tx and Rx DMA engines (if not already enabled) */
+ et131x_rx_dma_enable(adapter);
+ et131x_tx_dma_enable(adapter);
+
+ /* Enable device interrupts */
+ et131x_enable_interrupts(adapter);
+
+ /* We're ready to move some data, so start the queue */
+ netif_start_queue(netdev);
return;
}
}
--
1.7.4.4
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
