Skip to contentSkip to conversation

firewire: Fix ohci free_irq() warning

PATCH Superseded Mark Einon 5 participants 21 messages Jan 28, 2013 – Feb 17, 2013

Subsystems (2)
lore
Share & tools
lore

Mark Einon

author
#0 lore raw

PATCH 1/1 firewire: Fix ohci free_irq() warning

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

[  262.299486] WARNING: at /build/buildd/linux-3.5.0/kernel/irq/manage.c:1198 __free_irq+0xa3/0x1e0()
[  262.299487] Hardware name: MS-1727
[  262.299488] Trying to free already-free IRQ 16
[  262.299488] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge stp llc joydev arc4 parport_pc ppdev coretemp kvm_intel snd_hda_codec_realtek kvm microcode snd_seq_midi snd_rawmidi psmouse snd_seq_midi_event serio_raw nvidia(PO) i7core_edac snd_seq edac_core snd_hda_intel iwlwifi snd_hda_codec jmb38x_ms snd_hwdep mac80211 snd_pcm lpc_ich memstick snd_seq_device cfg80211 snd_timer snd soundcore snd_page_alloc ir_lirc_codec lirc_dev ir_mce_kbd_decoder ir_sanyo_decoder ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder rfcomm bnep rc_rc6_mce mxm_wmi ene_ir rc_core bluetooth wmi video acpiphp mac_hid lp parport binfmt_misc hid_generic usbhid hid firewire_ohci sdhci_pci r8169 firewire_core sdhci crc_itu_t
[  262.299537] Pid: 4, comm: kworker/0:0 Tainted: P           O 3.5.0-22-generic #34-Ubuntu
[  262.299538] Call Trace:
[  262.299540]  [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[  262.299545]  [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[  262.299548]  [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[  262.299551]  [<ffffffff810df6b3>] __free_irq+0xa3/0x1e0
[  262.299554]  [<ffffffff810df844>] free_irq+0x54/0xc0
[  262.299558]  [<ffffffffa005a27e>] pci_remove+0x6e/0x210 [firewire_ohci]
[  262.299564]  [<ffffffff8135ae7f>] pci_device_remove+0x3f/0x110
[  262.299569]  [<ffffffff8141fdbc>] __device_release_driver+0x7c/0xe0
[  262.299573]  [<ffffffff8141fe4c>] device_release_driver+0x2c/0x40
[  262.299576]  [<ffffffff8141f5f1>] bus_remove_device+0xe1/0x120
[  262.299578]  [<ffffffff8141cd1a>] device_del+0x12a/0x1c0
[  262.299581]  [<ffffffff8141cdc6>] device_unregister+0x16/0x30
[  262.299583]  [<ffffffff81354784>] pci_stop_bus_device+0x94/0xa0
[  262.299588]  [<ffffffffa0091c67>] acpiphp_disable_slot+0xb7/0x1a0 [acpiphp]
[  262.299594]  [<ffffffffa0090716>] ? get_slot_status+0x46/0xc0 [acpiphp]
[  262.299599]  [<ffffffffa0091d7d>] acpiphp_check_bridge.isra.15+0x2d/0xf0 [acpiphp]
[  262.299604]  [<ffffffffa0092442>] _handle_hotplug_event_bridge+0x372/0x4d0 [acpiphp]
[  262.299608]  [<ffffffff81390f8c>] ? acpi_os_execute_deferred+0x2f/0x34
[  262.299612]  [<ffffffff8116e22d>] ? kfree+0xed/0x110
[  262.299617]  [<ffffffff8107086a>] process_one_work+0x12a/0x420
[  262.299620]  [<ffffffffa00920d0>] ? _handle_hotplug_event_func+0x1d0/0x1d0 [acpiphp]
[  262.299625]  [<ffffffff8107141e>] worker_thread+0x12e/0x2f0
[  262.299627]  [<ffffffff810712f0>] ? manage_workers.isra.26+0x200/0x200
[  262.299629]  [<ffffffff81075f13>] kthread+0x93/0xa0
[  262.299632]  [<ffffffff8168d024>] kernel_thread_helper+0x4/0x10
[  262.299636]  [<ffffffff81075e80>] ? kthread_freezable_should_stop+0x70/0x70
[  262.299639]  [<ffffffff8168d020>] ? gs_change+0x13/0x13
[  262.299641] ---[ end trace 3f830890e076911f ]---
---
 drivers/firewire/ohci.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/firewire/ohci.c +2 -1
index 45912e6..029c407 100644
@@ -3766,7 +3766,7 @@ static int pci_suspend(struct pci_dev *dev, pm_message_t state)
 	int err;
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
+	disable_irq(dev->irq);
 	pci_disable_msi(dev);
 	err = pci_save_state(dev);
 	if (err) {
@@ -3807,6 +3807,7 @@ static int pci_resume(struct pci_dev *dev)
 		return err;
 
 	ohci_resume_iso_dma(ohci);
+	enable_irq(dev->irq);
 
 	return 0;
 }
Signature · 1 line
1.7.10.4
#1 lore raw
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Because the other way around, first pci_remove(), then pci_suspend(), surely must not appen. And if it does, the bug is elsewhere but not in firewire-ohci.

On that note, is pci_suspend() -> pci_remove() actually a legal state transition that must be handled by drivers?

Another comment below.

Quoted 54 lines
[  262.299486] WARNING: at /build/buildd/linux-3.5.0/kernel/irq/manage.c:1198 __free_irq+0xa3/0x1e0()
[  262.299487] Hardware name: MS-1727
[  262.299488] Trying to free already-free IRQ 16
[  262.299488] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge stp llc joydev arc4 parport_pc ppdev coretemp kvm_intel snd_hda_codec_realtek kvm microcode snd_seq_midi snd_rawmidi psmouse snd_seq_midi_event serio_raw nvidia(PO) i7core_edac snd_seq edac_core snd_hda_intel iwlwifi snd_hda_codec jmb38x_ms snd_hwdep mac80211 snd_pcm lpc_ich memstick snd_seq_device cfg80211 snd_timer snd soundcore snd_page_alloc ir_lirc_codec lirc_dev ir_mce_kbd_decoder ir_sanyo_decoder ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder rfcomm bnep rc_rc6_mce mxm_wmi ene_ir rc_core bluetooth wmi video acpiphp mac_hid lp parport binfmt_misc hid_generic usbhid hid firewire_ohci sdhci_pci r8169 firewire_core sdhci crc_itu_t
[  262.299537] Pid: 4, comm: kworker/0:0 Tainted: P           O 3.5.0-22-generic #34-Ubuntu
[  262.299538] Call Trace:
[  262.299540]  [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[  262.299545]  [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[  262.299548]  [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[  262.299551]  [<ffffffff810df6b3>] __free_irq+0xa3/0x1e0
[  262.299554]  [<ffffffff810df844>] free_irq+0x54/0xc0
[  262.299558]  [<ffffffffa005a27e>] pci_remove+0x6e/0x210 [firewire_ohci]
[  262.299564]  [<ffffffff8135ae7f>] pci_device_remove+0x3f/0x110
[  262.299569]  [<ffffffff8141fdbc>] __device_release_driver+0x7c/0xe0
[  262.299573]  [<ffffffff8141fe4c>] device_release_driver+0x2c/0x40
[  262.299576]  [<ffffffff8141f5f1>] bus_remove_device+0xe1/0x120
[  262.299578]  [<ffffffff8141cd1a>] device_del+0x12a/0x1c0
[  262.299581]  [<ffffffff8141cdc6>] device_unregister+0x16/0x30
[  262.299583]  [<ffffffff81354784>] pci_stop_bus_device+0x94/0xa0
[  262.299588]  [<ffffffffa0091c67>] acpiphp_disable_slot+0xb7/0x1a0 [acpiphp]
[  262.299594]  [<ffffffffa0090716>] ? get_slot_status+0x46/0xc0 [acpiphp]
[  262.299599]  [<ffffffffa0091d7d>] acpiphp_check_bridge.isra.15+0x2d/0xf0 [acpiphp]
[  262.299604]  [<ffffffffa0092442>] _handle_hotplug_event_bridge+0x372/0x4d0 [acpiphp]
[  262.299608]  [<ffffffff81390f8c>] ? acpi_os_execute_deferred+0x2f/0x34
[  262.299612]  [<ffffffff8116e22d>] ? kfree+0xed/0x110
[  262.299617]  [<ffffffff8107086a>] process_one_work+0x12a/0x420
[  262.299620]  [<ffffffffa00920d0>] ? _handle_hotplug_event_func+0x1d0/0x1d0 [acpiphp]
[  262.299625]  [<ffffffff8107141e>] worker_thread+0x12e/0x2f0
[  262.299627]  [<ffffffff810712f0>] ? manage_workers.isra.26+0x200/0x200
[  262.299629]  [<ffffffff81075f13>] kthread+0x93/0xa0
[  262.299632]  [<ffffffff8168d024>] kernel_thread_helper+0x4/0x10
[  262.299636]  [<ffffffff81075e80>] ? kthread_freezable_should_stop+0x70/0x70
[  262.299639]  [<ffffffff8168d020>] ? gs_change+0x13/0x13
[  262.299641] ---[ end trace 3f830890e076911f ]---
---
 drivers/firewire/ohci.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/firewire/ohci.c +1 -1
index 45912e6..029c407 100644
@@ -3766,7 +3766,7 @@ static int pci_suspend(struct pci_dev *dev, pm_message_t state)
 	int err;
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
+	disable_irq(dev->irq);
 	pci_disable_msi(dev);
 	err = pci_save_state(dev);
 	if (err) {

drivers/firewire/ohci.c static int pci_resume(struct pci_dev *dev) · line 3813

Quoted 7 lines
drivers/firewire/ohci.c +1 -0
@@ -3807,6 +3807,7 @@ static int pci_resume(struct pci_dev *dev)
 		return err;
 
 	ohci_resume_iso_dma(ohci);
+	enable_irq(dev->irq);
 
 	return 0;
 }

firewire-ohci's pci_resume() calls request_irq() a little bit before that, in ohci_enable(). Is the sequence pci_probe/request_irq() -> pci_suspend/disable_irq() -> pci_resume/request_irq() -> pci_resume/enable_irq() legal?

Signature · 3 lines
Stefan Richter
-=====-===-= ---= ===--
http://arcgraph.de/sr/
author
#2 lore raw
Quoted 7 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Quoted 7 lines

Because the other way around, first pci_remove(), then pci_suspend(), surely must not appen. And if it does, the bug is elsewhere but not in firewire-ohci.

On that note, is pci_suspend() -> pci_remove() actually a legal state transition that must be handled by drivers?

It's happening on the Ubuntu 12.10 distro which performs the suspend then remove sequence. I assumed that was normal.

Quoted 2 lines

Another comment below.

<snip>

Quoted 5 lines

firewire-ohci's pci_resume() calls request_irq() a little bit before that, in ohci_enable(). Is the sequence pci_probe/request_irq() -> pci_suspend/disable_irq() -> pci_resume/request_irq() -> pci_resume/enable_irq() legal?

I missed the call to request_irq(). Perhaps the simplest way to handle the fix would be to use a flag that holds the irq state? I'm open to suggestions.

Cheers,

Mark

#3 lore raw

Added Cc: linux-pm.

Quoted 16 lines · On Jan 29 Mark Einon wrote:
Quoted 7 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Quoted 5 lines

Because the other way around, first pci_remove(), then pci_suspend(), surely must not appen. And if it does, the bug is elsewhere but not in firewire-ohci.

Quoted 5 lines
Quoted 2 lines

On that note, is pci_suspend() -> pci_remove() actually a legal state transition that must be handled by drivers?

It's happening on the Ubuntu 12.10 distro which performs the suspend then remove sequence. I assumed that was normal.

Maybe a parent device (PCI bridge or the likes) of the OHCI is requested to be removed during suspend or is being removed during resume, thereby causing a removal request to the OHCI?

Or the suspend method of firewire-ohci exited with an error return code... but then the suspend sequence would be rolled back, not the offending device be removed, right?

In any case, could you check the dmesg right before the warning which you already posted --- and right after it too --- whether there is an indication what else was going on?

Also, what are the parent devices of the OHCI according to "lspci -tv" (or what else can show PCI topology)?

Quoted 11 lines
Quoted 1 line

Another comment below.

<snip>

Quoted 5 lines

firewire-ohci's pci_resume() calls request_irq() a little bit before that, in ohci_enable(). Is the sequence pci_probe/request_irq() -> pci_suspend/disable_irq() -> pci_resume/request_irq() -> pci_resume/enable_irq() legal?

I missed the call to request_irq(). Perhaps the simplest way to handle the fix would be to use a flag that holds the irq state? I'm open to suggestions.

Not sure what extent of state tracking the PCI core or driver core already offer; I'll have to look.

Signature · 3 lines
Stefan Richter
-=====-===-= ---= ===-=
http://arcgraph.de/sr/
#4 lore raw
Quoted 13 lines · Stefan Richter

Added Cc: linux-pm.

Quoted 10 lines · On Jan 29 Mark Einon wrote:
Quoted 7 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

Alan Stern

author
#5 lore raw
Quoted 11 lines · Alan Stern
Quoted 9 lines · Stefan Richter

Added Cc: linux-pm.

Quoted 6 lines · On Jan 29 Mark Einon wrote:
Quoted 5 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

Quoted 8 lines
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

This sounds reasonable to me - I think we could probably get rid of the request_irq() call from resume, and use disable_irq()/enable_irq()?
I'll attempt a patch - but unfortunately I don't have a firewire device to test.

Cheers,

Mark

#6 lore raw
Quoted 9 lines · Mark Einon
Quoted 9 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Quoted 7 lines
Quoted 3 lines

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

This sounds reasonable to me - I think we could probably get rid of the request_irq() call from resume, and use disable_irq()/enable_irq()?

Why mess around with IRQ settings at all? Just have the suspend routine tell the controller to stop generating them.

Alan Stern

author
#7 lore raw
Quoted 15 lines · Alan Stern
Quoted 13 lines · Mark Einon
Quoted 12 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

Quoted 9 lines
Quoted 3 lines

This sounds reasonable to me - I think we could probably get rid of the request_irq() call from resume, and use disable_irq()/enable_irq()?

Why mess around with IRQ settings at all? Just have the suspend routine tell the controller to stop generating them.

Alan Stern

I looked into doing this; using context_stop() to stop the controller running.

However, removing the enable_irq() from pci_resume() involves not calling ohci_enable() (as it is also the fw_card_driver.enable function, and can't easily be modified). As this call involves a lot of register writes and I have no devices to test, I decided against it.

I'll send an updated patch for consideration that merely uses a bool to stop the irq being freed twice - crude, but it works without changing too much code.

Cheers,

Mark

#8 lore raw
Quoted 34 lines · Mark Einon
Quoted 15 lines · Alan Stern
Quoted 13 lines · Mark Einon
Quoted 12 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

Quoted 9 lines
Quoted 3 lines

This sounds reasonable to me - I think we could probably get rid of the request_irq() call from resume, and use disable_irq()/enable_irq()?

Why mess around with IRQ settings at all? Just have the suspend routine tell the controller to stop generating them.

Alan Stern

I looked into doing this; using context_stop() to stop the controller running.

However, removing the enable_irq() from pci_resume() involves not calling ohci_enable() (as it is also the fw_card_driver.enable function, and can't easily be modified). As this call involves a lot of register writes and I have no devices to test, I decided against it.

Quoted 3 lines

I'll send an updated patch for consideration that merely uses a bool to stop the irq being freed twice - crude, but it works without changing too much code.

Hi Mark,

I think what Alan means is that the suspend/resume code should just mask/unmask interrupts at the OHCI controller, via the OHCI IntEventClear/Set registers (naturally, saving the current mask and restoring it on resume).

Of course, there's a lot more to do with an OHCI controller -- as you note. Like stopping running DMA contexts :) And restarting them on resume.

I'd do it, but I'm buried to my eyeballs in tty right now -- not fun. I can _eventually_ do this as I need to address problems with the FW643 anyway at some point, but it's going to be a little while.

In the meantime, I'm a little confused: you say you can't test this code because you have no hardware; but then how'd you trip this bug?

Regards,
Peter Hurley

#9 lore raw
Quoted 35 lines · Peter Hurley
Quoted 34 lines · Mark Einon
Quoted 15 lines · Alan Stern
Quoted 13 lines · Mark Einon
Quoted 12 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

Quoted 9 lines
Quoted 3 lines

This sounds reasonable to me - I think we could probably get rid of the request_irq() call from resume, and use disable_irq()/enable_irq()?

Why mess around with IRQ settings at all? Just have the suspend routine tell the controller to stop generating them.

Alan Stern

I looked into doing this; using context_stop() to stop the controller running.

However, removing the enable_irq() from pci_resume() involves not calling ohci_enable() (as it is also the fw_card_driver.enable function, and can't easily be modified). As this call involves a lot of register writes and I have no devices to test, I decided against it.

Quoted 3 lines
Quoted 3 lines

I'll send an updated patch for consideration that merely uses a bool to stop the irq being freed twice - crude, but it works without changing too much code.

Sorry that should read...

Quoted 5 lines

Hi Mark,

I think what Alan means is that the suspend/resume code should just mask/unmask interrupts at the OHCI controller, via the OHCI IntEventClear/Set registers (naturally, saving the current mask and

     ^^^^^^^^

IntMaskClear/Set

by clearing/setting masterIntEnable

Quoted 15 lines

restoring it on resume).

Of course, there's a lot more to do with an OHCI controller -- as you note. Like stopping running DMA contexts :) And restarting them on resume.

I'd do it, but I'm buried to my eyeballs in tty right now -- not fun. I can _eventually_ do this as I need to address problems with the FW643 anyway at some point, but it's going to be a little while.

In the meantime, I'm a little confused: you say you can't test this code because you have no hardware; but then how'd you trip this bug?

Regards,
Peter Hurley

author
#10 lore raw
In reply to Peter Hurley
Quoted 7 lines · Peter Hurley

Hi Mark,

I think what Alan means is that the suspend/resume code should just mask/unmask interrupts at the OHCI controller, via the OHCI IntEventClear/Set registers (naturally, saving the current mask and restoring it on resume).

Quoted 7 lines

Of course, there's a lot more to do with an OHCI controller -- as you note. Like stopping running DMA contexts :) And restarting them on resume.

I'd do it, but I'm buried to my eyeballs in tty right now -- not fun. I can _eventually_ do this as I need to address problems with the FW643 anyway at some point, but it's going to be a little while.

Hi Peter,

Ok, understood. I can certainly attempt a patch if I get time.

Quoted 3 lines

In the meantime, I'm a little confused: you say you can't test this code because you have no hardware; but then how'd you trip this bug?

I can test the code in that I have a firewire port on my laptop, but haven't got anything to plug into the port.
I assume that any large changes I make are quite capable of breaking something there...

Cheers,

Mark

#11 lore raw
Quoted 5 lines · On Feb 01 Mark Einon wrote:
Quoted 4 lines · Peter Hurley
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · On Jan 29 Alan Stern wrote:
Quoted 3 lines

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

As far as I can tell, it happened to be done that way as a side effect of how the probe() and resume() methods share code. It has remained like this since the initial implementation:
http://git.kernel.org/linus/2aef469a35a2

Still, at this point I would like to learn whether .suspend() followed by .remove() is a valid order of sequence which drivers must support before I prepare myself to get comfortable with a refactoring of firewire-ohci's .probe()/.resume()/suspend()/remove(). Obviously, so far my assumption was that a successful .suspend() can only ever be followed by .resume().

Quoted 18 lines
Quoted 5 lines

I think what Alan means is that the suspend/resume code should just mask/unmask interrupts at the OHCI controller, via the OHCI IntEventClear/Set registers (naturally, saving the current mask and restoring it on resume).

Quoted 7 lines

Of course, there's a lot more to do with an OHCI controller -- as you note. Like stopping running DMA contexts :) And restarting them on resume.

I'd do it, but I'm buried to my eyeballs in tty right now -- not fun. I can _eventually_ do this as I need to address problems with the FW643 anyway at some point, but it's going to be a little while.

Hi Peter,

Ok, understood. I can certainly attempt a patch if I get time.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

In the meantime, I'm a little confused: you say you can't test this code because you have no hardware; but then how'd you trip this bug?

I can test the code in that I have a firewire port on my laptop, but haven't got anything to plug into the port.
I assume that any large changes I make are quite capable of breaking something there...

This is a valid assumption. Some years ago I caused a regression in stable kernel branches in exactly this way myself.

Signature · 3 lines
Stefan Richter
-=====-===-= --=- ---=-
http://arcgraph.de/sr/
#12 lore raw
Quoted 11 lines · Stefan Richter
Quoted 5 lines · On Feb 01 Mark Einon wrote:
Quoted 4 lines · Peter Hurley
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · On Jan 29 Alan Stern wrote:
Quoted 3 lines

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

As far as I can tell, it happened to be done that way as a side effect of how the probe() and resume() methods share code. It has remained like this since the initial implementation:
http://git.kernel.org/linus/2aef469a35a2

At one point, quite a few years ago, Linus complained about drivers the release IRQs during suspend only to reacquire them during resume. A little refactoring should be able to separate out resource acquisition/release (done only during probe and remove) from activation and shutdown (also done during resume and suspend).

Quoted 6 lines

Still, at this point I would like to learn whether .suspend() followed by .remove() is a valid order of sequence which drivers must support before I prepare myself to get comfortable with a refactoring of firewire-ohci's .probe()/.resume()/suspend()/remove(). Obviously, so far my assumption was that a successful .suspend() can only ever be followed by .resume().

It depends on the subsystem. Some subsystems do have suspend -> remove transitions and others don't. In general, it's a good idea for drivers to be prepared for removal while the system is asleep. Presumably any hot-unpluggable bus (which includes most of the important buses these days) would have to support it.

Alan Stern

#13 lore raw
Quoted 27 lines · On Feb 02 Alan Stern wrote:
Quoted 11 lines · Stefan Richter
Quoted 5 lines · On Feb 01 Mark Einon wrote:
Quoted 4 lines · Peter Hurley
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · On Jan 29 Alan Stern wrote:
Quoted 3 lines

Why does the pci_suspend routine call free_irq() at all? As far as I know, it's not supposed to do that. Won't the device continue to use the same IRQ after it is resumed?

As far as I can tell, it happened to be done that way as a side effect of how the probe() and resume() methods share code. It has remained like this since the initial implementation:
http://git.kernel.org/linus/2aef469a35a2

At one point, quite a few years ago, Linus complained about drivers the release IRQs during suspend only to reacquire them during resume. A little refactoring should be able to separate out resource acquisition/release (done only during probe and remove) from activation and shutdown (also done during resume and suspend).

Quoted 6 lines

Still, at this point I would like to learn whether .suspend() followed by .remove() is a valid order of sequence which drivers must support before I prepare myself to get comfortable with a refactoring of firewire-ohci's .probe()/.resume()/suspend()/remove(). Obviously, so far my assumption was that a successful .suspend() can only ever be followed by .resume().

Quoted 5 lines

It depends on the subsystem. Some subsystems do have suspend -> remove transitions and others don't. In general, it's a good idea for drivers to be prepared for removal while the system is asleep. Presumably any hot-unpluggable bus (which includes most of the important buses these days) would have to support it.

OK, thank you. In this case we are of course dealing with the pci subsystem (and with PCI/ CardBus/ PCI Express/ ExpressCard attached hardware). Maybe I should have addressed my question to linux-pci instead of linux-pm; however, if this is the general expectation, then I too prefer firewire-ohci to be able to handle it even if the pci subsystem wouldn't require it presently (which now sounds unlikely).

Signature · 3 lines
Stefan Richter
-=====-===-= --=- ---=-
http://arcgraph.de/sr/
author
#14 lore raw
In reply to Stefan Richter
Quoted 13 lines · Stefan Richter

Added Cc: linux-pm.

Quoted 10 lines · On Jan 29 Mark Einon wrote:
Quoted 7 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Correction above : s/resume/remove

Quoted 16 lines
Quoted 11 lines
Quoted 8 lines

Because the other way around, first pci_remove(), then pci_suspend(), surely must not appen. And if it does, the bug is elsewhere but not in firewire-ohci.

On that note, is pci_suspend() -> pci_remove() actually a legal state transition that must be handled by drivers?

It's happening on the Ubuntu 12.10 distro which performs the suspend then remove sequence. I assumed that was normal.

Maybe a parent device (PCI bridge or the likes) of the OHCI is requested to be removed during suspend or is being removed during resume, thereby causing a removal request to the OHCI?

Quoted 7 lines

Or the suspend method of firewire-ohci exited with an error return code... but then the suspend sequence would be rolled back, not the offending device be removed, right?

In any case, could you check the dmesg right before the warning which you already posted --- and right after it too --- whether there is an indication what else was going on?

It looks like the remove call actually happens on the wakeup sequence
- although I've not yet found the time to investigate further. The lines preceding the crash are:

Jan 30 23:22:53 msilap kernel: [ 7682.288190] ACPI: Low-level resume complete
Jan 30 23:22:53 msilap kernel: [ 7682.288228] PM: Restoring platform NVS memory
Jan 30 23:22:53 msilap kernel: [ 7682.288644] Enabling non-boot CPUs ...
Jan 30 23:22:53 msilap kernel: [ 7682.288713] smpboot: Booting Node 0
Processor 1 APIC 0x2
Jan 30 23:22:53 msilap kernel: [ 7682.301963] hpet: hpet3 irq 41 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.301978] CPU1 is up
Jan 30 23:22:53 msilap kernel: [ 7682.302040] smpboot: Booting Node 0
Processor 2 APIC 0x4
Jan 30 23:22:53 msilap kernel: [ 7682.315394] hpet: hpet4 irq 42 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.315408] CPU2 is up
Jan 30 23:22:53 msilap kernel: [ 7682.315470] smpboot: Booting Node 0
Processor 3 APIC 0x6
Jan 30 23:22:53 msilap kernel: [ 7682.328815] hpet: hpet5 irq 43 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.328828] CPU3 is up
Jan 30 23:22:53 msilap kernel: [ 7682.328882] smpboot: Booting Node 0
Processor 4 APIC 0x1
Jan 30 23:22:53 msilap kernel: [ 7682.342256] hpet: hpet6 irq 44 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.342269] CPU4 is up
Jan 30 23:22:53 msilap kernel: [ 7682.342327] smpboot: Booting Node 0
Processor 5 APIC 0x3
Jan 30 23:22:53 msilap kernel: [ 7682.355737] CPU5 is up
Jan 30 23:22:53 msilap kernel: [ 7682.355797] smpboot: Booting Node 0
Processor 6 APIC 0x5
Jan 30 23:22:53 msilap kernel: [ 7682.369213] CPU6 is up
Jan 30 23:22:53 msilap kernel: [ 7682.369272] smpboot: Booting Node 0
Processor 7 APIC 0x7
Jan 30 23:22:53 msilap kernel: [ 7682.382720] CPU7 is up
Jan 30 23:22:53 msilap kernel: [ 7682.388801] ACPI: Waking up from
system sleep state S3
Jan 30 23:22:53 msilap kernel: [ 7682.403997] ------------[ cut here
]------------
Jan 30 23:22:53 msilap kernel: [ 7682.404007] WARNING: at

/home/mark/Source/linux/kernel/irq/manage.c:1248
__free_irq+0xa3/0x1e0()
<snip>

then after the first crash trace:

Jan 30 23:22:53 msilap kernel: [ 7682.404181] ---[ end trace
1a4f0a37b1022aad ]---
Jan 30 23:22:53 msilap kernel: [ 7682.404211] firewire_ohci
0000:07:00.0: removed fw-ohci device
Jan 30 23:22:53 msilap kernel: [ 7682.404327] ------------[ cut here
]------------
Jan 30 23:22:53 msilap kernel: [ 7682.404330] WARNING: at
/home/mark/Source/linux/kernel/irq/manage.c:1248
__free_irq+0xa3/0x1e0()
Jan 30 23:22:53 msilap kernel: [ 7682.404330] Hardware name: MS-1727
Jan 30 23:22:53 msilap kernel: [ 7682.404331] Trying to free already-free IRQ 16
Jan 30 23:22:53 msilap kernel: [ 7682.404363] Modules linked in:

ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter bnep rfcomm ip_tables bluetooth x_tables bridge stp llc parport_pc ppdev binfmt_misc arc4 nouveau iwldvm mac80211 snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec iwlwifi coretemp kvm_intel snd_hwdep kvm snd_pcm ttm cfg80211 snd_seq_midi snd_rawmidi drm_kms_helper snd_seq_midi_event drm snd_seq snd_timer snd_seq_device joydev snd microcode i7core_edac i2c_algo_bit psmouse soundcore edac_core jmb38x_ms msi_wmi serio_raw video memstick mxm_wmi lpc_ich sparse_keymap mac_hid wmi snd_page_alloc acpiphp lp parport r8169 firewire_ohci(O) sdhci_pci firewire_core(O) sdhci crc_itu_t

Jan 30 23:22:53 msilap kernel: [ 7682.404365] Pid: 5968, comm:
kworker/0:2 Tainted: G        W  O 3.8.0-rc5-next-20130128+ #6
Jan 30 23:22:53 msilap kernel: [ 7682.404366] Call Trace:
Jan 30 23:22:53 msilap kernel: [ 7682.404369]  [<ffffffff810589af>]
warn_slowpath_common+0x7f/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404371]  [<ffffffff81058aa6>]
warn_slowpath_fmt+0x46/0x50
Jan 30 23:22:53 msilap kernel: [ 7682.404374]  [<ffffffff81044c69>] ?
default_spin_lock_flags+0x9/0x10
Jan 30 23:22:53 msilap kernel: [ 7682.404376]  [<ffffffff810eabc3>]
__free_irq+0xa3/0x1e0
Jan 30 23:22:53 msilap kernel: [ 7682.404378]  [<ffffffff810ead54>]
free_irq+0x54/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404384]  [<ffffffffa000968a>]
sdhci_remove_host+0x6a/0x190 [sdhci]
Jan 30 23:22:53 msilap kernel: [ 7682.404389]  [<ffffffffa002f3ff>]
sdhci_pci_remove_slot+0x4f/0xc0 [sdhci_pci]
Jan 30 23:22:53 msilap kernel: [ 7682.404394]  [<ffffffffa002fa70>]
sdhci_pci_remove+0x50/0xa0 [sdhci_pci]
Jan 30 23:22:53 msilap kernel: [ 7682.404398]  [<ffffffff8137ff26>]
pci_device_remove+0x46/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404402]  [<ffffffff81456f3c>]
__device_release_driver+0x7c/0xe0
Jan 30 23:22:53 msilap kernel: [ 7682.404404]  [<ffffffff814571ac>]
device_release_driver+0x2c/0x40
Jan 30 23:22:53 msilap kernel: [ 7682.404406]  [<ffffffff814569b1>]
bus_remove_device+0xe1/0x120
Jan 30 23:22:53 msilap kernel: [ 7682.404409]  [<ffffffff8145403a>]
device_del+0x11a/0x1b0
Jan 30 23:22:53 msilap kernel: [ 7682.404411]  [<ffffffff81379d94>]
pci_stop_bus_device+0x94/0xa0
Jan 30 23:22:53 msilap kernel: [ 7682.404414]  [<ffffffff81379f36>]
pci_stop_and_remove_bus_device+0x16/0x30
Jan 30 23:22:53 msilap kernel: [ 7682.404419]  [<ffffffffa0028b36>]
acpiphp_disable_slot+0xf6/0x1a0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404424]  [<ffffffffa0027716>] ?
get_slot_status+0x46/0xc0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404428]  [<ffffffffa0028c0d>]
acpiphp_check_bridge.isra.13+0x2d/0xf0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404433]  [<ffffffffa0029247>]
_handle_hotplug_event_bridge+0x2e7/0x410 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404436]  [<ffffffff813bdc64>] ?
acpi_os_execute_deferred+0x2d/0x32
Jan 30 23:22:53 msilap kernel: [ 7682.404439]  [<ffffffff8117df25>] ?
kfree+0x105/0x140
Jan 30 23:22:53 msilap kernel: [ 7682.404441]  [<ffffffff81077249>]
process_one_work+0x159/0x4c0
Jan 30 23:22:53 msilap kernel: [ 7682.404444]  [<ffffffff8107895e>]
worker_thread+0x15e/0x4a0
Jan 30 23:22:53 msilap kernel: [ 7682.404446]  [<ffffffff81078800>] ?
manage_workers+0x2a0/0x2a0
Jan 30 23:22:53 msilap kernel: [ 7682.404448]  [<ffffffff8107df60>]
kthread+0xc0/0xd0
Jan 30 23:22:53 msilap kernel: [ 7682.404450]  [<ffffffff8107dea0>] ?
kthread_create_on_node+0x130/0x130
Jan 30 23:22:53 msilap kernel: [ 7682.404452]  [<ffffffff816d756c>]
ret_from_fork+0x7c/0xb0
Jan 30 23:22:53 msilap kernel: [ 7682.404454]  [<ffffffff8107dea0>] ?
kthread_create_on_node+0x130/0x130
Jan 30 23:22:53 msilap kernel: [ 7682.404456] ---[ end trace

1a4f0a37b1022aae ]---

Quoted 3 lines

Also, what are the parent devices of the OHCI according to "lspci -tv" (or what else can show PCI topology)?

I ran 'lshw' on the machine, which may be of more use. The relevant output is (removing parts not associated with the firewire device):

msilap
description: Notebook
product: MS-1727 (To be filled by O.E.M.)
vendor: MICRO-STAR INTERNATIONAL CO., LTD
version: REV:1.0
serial: FFFFFFFF
width: 64 bits
capabilities: smbios-2.6 dmi-2.6 vsyscall32
configuration: boot=normal chassis=notebook family=To be filled by
O.E.M. sku=To be filled by O.E.M. uuid=00020003-0004-0005-0006-000700080009
*-core
description: Motherboard
product: MS-1727
vendor: MICRO-STAR INTERNATIONAL CO., LTD
physical id: 0
version: REV:1.0
serial: BSS-0123456789
slot: To be filled by O.E.M.
*-pci:0
description: Host bridge
product: Core Processor DMI
vendor: Intel Corporation
physical id: 100
bus info: pci@0000:00:00.0
version: 11
width: 32 bits
clock: 33MHz
*-pci:4
description: PCI bridge
product: 5 Series/3400 Series Chipset PCI Express Root Port 5
vendor: Intel Corporation
physical id: 1c.4
bus info: pci@0000:00:1c.4
version: 05
width: 32 bits
clock: 33MHz
capabilities: pci pciexpress msi pm normal_decode
bus_master cap_list

             configuration: driver=pcieport
             resources: irq:16 ioport:9000(size=4096)

memory:d4600000-d59fffff ioport:d9d00000(size=2097152)
*-firewire
description: FireWire (IEEE 1394)
product: IEEE 1394 Host Controller
vendor: JMicron Technology Corp.
physical id: 0
bus info: pci@0000:07:00.0
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi ohci bus_master cap_list
configuration: driver=firewire_ohci latency=0
resources: irq:16 memory:d4607000-d46077ff
memory:d4606000-d460607f memory:d4605000-d460507f memory:d4604000-d460407f
*-generic:0
description: System peripheral
product: SD/MMC Host Controller
vendor: JMicron Technology Corp.
physical id: 0.1
bus info: pci@0000:07:00.1
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: driver=sdhci-pci latency=0
resources: irq:16 memory:d4603000-d46030ff
*-generic:1 UNCLAIMED
description: SD Host controller
product: Standard SD Host Controller
vendor: JMicron Technology Corp.
physical id: 0.2
bus info: pci@0000:07:00.2
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi cap_list
configuration: latency=0
resources: memory:d4602000-d46020ff
*-generic:2
description: System peripheral
product: MS Host Controller
vendor: JMicron Technology Corp.
physical id: 0.3
bus info: pci@0000:07:00.3
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: driver=jmb38x_ms latency=0
resources: irq:16 memory:d4601000-d46010ff
*-generic:3 UNCLAIMED
description: System peripheral
product: xD Host Controller
vendor: JMicron Technology Corp.
physical id: 0.4
bus info: pci@0000:07:00.4
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: latency=0
resources: memory:d4600000-d46000ff

Quoted 19 lines
Quoted 4 lines
Quoted 2 lines

Another comment below.

<snip>

Quoted 1 line
Quoted 8 lines
Quoted 4 lines

firewire-ohci's pci_resume() calls request_irq() a little bit before that, in ohci_enable(). Is the sequence pci_probe/request_irq() -> pci_suspend/disable_irq() -> pci_resume/request_irq() -> pci_resume/enable_irq() legal?

I missed the call to request_irq(). Perhaps the simplest way to handle the fix would be to use a flag that holds the irq state? I'm open to suggestions.

Not sure what extent of state tracking the PCI core or driver core already offer; I'll have to look.
--
Stefan Richter
-=====-===-= ---= ===-=
http://arcgraph.de/sr/

#15 lore raw
Quoted 37 lines · On Jan 30 Mark Einon wrote:
Quoted 13 lines · Stefan Richter

Added Cc: linux-pm.

Quoted 10 lines · On Jan 29 Mark Einon wrote:
Quoted 7 lines · Stefan Richter
Quoted 3 lines · On Jan 28 Mark Einon wrote:

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

You mean /first/ in pci_suspend() and /then/ in pci_remove() on the already suspended devices, right?

Yes, I did. The call sequence is suspend then resume. My bad.

Correction above : s/resume/remove

Quoted 20 lines
Quoted 11 lines
Quoted 8 lines

Because the other way around, first pci_remove(), then pci_suspend(), surely must not appen. And if it does, the bug is elsewhere but not in firewire-ohci.

On that note, is pci_suspend() -> pci_remove() actually a legal state transition that must be handled by drivers?

It's happening on the Ubuntu 12.10 distro which performs the suspend then remove sequence. I assumed that was normal.

Maybe a parent device (PCI bridge or the likes) of the OHCI is requested to be removed during suspend or is being removed during resume, thereby causing a removal request to the OHCI?

Or the suspend method of firewire-ohci exited with an error return code... but then the suspend sequence would be rolled back, not the offending device be removed, right?

Quoted 6 lines
Quoted 3 lines

In any case, could you check the dmesg right before the warning which you already posted --- and right after it too --- whether there is an indication what else was going on?

It looks like the remove call actually happens on the wakeup sequence
- although I've not yet found the time to investigate further.

OK. I, as a naive driver maintainer, expect a sequence like
[suspend request] -> suspend() -> [if ret==0, suspended state] -> [resume request] -> resume() -> ...
but what apparently happens is
[suspend request] -> suspend() -> [if ret==0, suspended state] -> [resume request] -> remove() -> ...

If you had an ExpressCard or CardBus controller, then it could of course happen that the PC is suspended, then the controller card physically removed, and then a resume been requested. Question to all: Would the driver's resume() method be called on the MIA device be called, or would the driver's remove() method be called in such a case?

However, this is not what happened in your case: According to the hardware information which you provided below, you have a PCI Express controller which is apparently soldered onto the mainboard, and it is not obvious to me why the ACPI wake-up is causing this controller attempted to be removed.

Still, I am not clear on the following two questions:

- Why is this happening on your laptop at all?

- Must Linux PCI kernel drivers (or any kernel device drivers supporting PM suspend/ remove) be prepared to handle direct state transitions from [suspended] to [remove]?

Quoted 63 lines

The lines preceding the crash are:

Jan 30 23:22:53 msilap kernel: [ 7682.288190] ACPI: Low-level resume complete
Jan 30 23:22:53 msilap kernel: [ 7682.288228] PM: Restoring platform NVS memory
Jan 30 23:22:53 msilap kernel: [ 7682.288644] Enabling non-boot CPUs ...
Jan 30 23:22:53 msilap kernel: [ 7682.288713] smpboot: Booting Node 0 Processor 1 APIC 0x2
Jan 30 23:22:53 msilap kernel: [ 7682.301963] hpet: hpet3 irq 41 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.301978] CPU1 is up
Jan 30 23:22:53 msilap kernel: [ 7682.302040] smpboot: Booting Node 0 Processor 2 APIC 0x4
Jan 30 23:22:53 msilap kernel: [ 7682.315394] hpet: hpet4 irq 42 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.315408] CPU2 is up
Jan 30 23:22:53 msilap kernel: [ 7682.315470] smpboot: Booting Node 0 Processor 3 APIC 0x6
Jan 30 23:22:53 msilap kernel: [ 7682.328815] hpet: hpet5 irq 43 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.328828] CPU3 is up
Jan 30 23:22:53 msilap kernel: [ 7682.328882] smpboot: Booting Node 0 Processor 4 APIC 0x1
Jan 30 23:22:53 msilap kernel: [ 7682.342256] hpet: hpet6 irq 44 for MSI
Jan 30 23:22:53 msilap kernel: [ 7682.342269] CPU4 is up
Jan 30 23:22:53 msilap kernel: [ 7682.342327] smpboot: Booting Node 0 Processor 5 APIC 0x3
Jan 30 23:22:53 msilap kernel: [ 7682.355737] CPU5 is up
Jan 30 23:22:53 msilap kernel: [ 7682.355797] smpboot: Booting Node 0 Processor 6 APIC 0x5
Jan 30 23:22:53 msilap kernel: [ 7682.369213] CPU6 is up
Jan 30 23:22:53 msilap kernel: [ 7682.369272] smpboot: Booting Node 0 Processor 7 APIC 0x7
Jan 30 23:22:53 msilap kernel: [ 7682.382720] CPU7 is up
Jan 30 23:22:53 msilap kernel: [ 7682.388801] ACPI: Waking up from system sleep state S3
Jan 30 23:22:53 msilap kernel: [ 7682.403997] ------------[ cut here ]------------
Jan 30 23:22:53 msilap kernel: [ 7682.404007] WARNING: at /home/mark/Source/linux/kernel/irq/manage.c:1248 __free_irq+0xa3/0x1e0()

<snip>

then after the first crash trace:

Jan 30 23:22:53 msilap kernel: [ 7682.404181] ---[ end trace 1a4f0a37b1022aad ]---
Jan 30 23:22:53 msilap kernel: [ 7682.404211] firewire_ohci 0000:07:00.0: removed fw-ohci device
Jan 30 23:22:53 msilap kernel: [ 7682.404327] ------------[ cut here ]------------
Jan 30 23:22:53 msilap kernel: [ 7682.404330] WARNING: at /home/mark/Source/linux/kernel/irq/manage.c:1248 __free_irq+0xa3/0x1e0()
Jan 30 23:22:53 msilap kernel: [ 7682.404330] Hardware name: MS-1727
Jan 30 23:22:53 msilap kernel: [ 7682.404331] Trying to free already-free IRQ 16
Jan 30 23:22:53 msilap kernel: [ 7682.404363] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter bnep rfcomm ip_tables bluetooth x_tables bridge stp llc parport_pc ppdev binfmt_misc arc4 nouveau iwldvm mac80211 snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec iwlwifi coretemp kvm_intel snd_hwdep kvm
snd_pcm ttm cfg80211 snd_seq_midi snd_rawmidi drm_kms_helper snd_seq_midi_event drm snd_seq snd_timer snd_seq_device joydev snd
microcode i7core_edac i2c_algo_bit psmouse soundcore edac_core jmb38x_ms msi_wmi serio_raw video memstick mxm_wmi lpc_ich
sparse_keymap mac_hid wmi snd_page_alloc acpiphp lp parport r8169 firewire_ohci(O) sdhci_pci firewire_core(O) sdhci crc_itu_t
Jan 30 23:22:53 msilap kernel: [ 7682.404365] Pid: 5968, comm: kworker/0:2 Tainted: G        W  O 3.8.0-rc5-next-20130128+ #6
Jan 30 23:22:53 msilap kernel: [ 7682.404366] Call Trace:
Jan 30 23:22:53 msilap kernel: [ 7682.404369]  [<ffffffff810589af>] warn_slowpath_common+0x7f/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404371]  [<ffffffff81058aa6>] warn_slowpath_fmt+0x46/0x50
Jan 30 23:22:53 msilap kernel: [ 7682.404374]  [<ffffffff81044c69>] ? default_spin_lock_flags+0x9/0x10
Jan 30 23:22:53 msilap kernel: [ 7682.404376]  [<ffffffff810eabc3>] __free_irq+0xa3/0x1e0
Jan 30 23:22:53 msilap kernel: [ 7682.404378]  [<ffffffff810ead54>] free_irq+0x54/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404384]  [<ffffffffa000968a>] sdhci_remove_host+0x6a/0x190 [sdhci]
Jan 30 23:22:53 msilap kernel: [ 7682.404389]  [<ffffffffa002f3ff>] sdhci_pci_remove_slot+0x4f/0xc0 [sdhci_pci]
Jan 30 23:22:53 msilap kernel: [ 7682.404394]  [<ffffffffa002fa70>] sdhci_pci_remove+0x50/0xa0 [sdhci_pci]
Jan 30 23:22:53 msilap kernel: [ 7682.404398]  [<ffffffff8137ff26>] pci_device_remove+0x46/0xc0
Jan 30 23:22:53 msilap kernel: [ 7682.404402]  [<ffffffff81456f3c>] __device_release_driver+0x7c/0xe0
Jan 30 23:22:53 msilap kernel: [ 7682.404404]  [<ffffffff814571ac>] device_release_driver+0x2c/0x40
Jan 30 23:22:53 msilap kernel: [ 7682.404406]  [<ffffffff814569b1>] bus_remove_device+0xe1/0x120
Jan 30 23:22:53 msilap kernel: [ 7682.404409]  [<ffffffff8145403a>] device_del+0x11a/0x1b0
Jan 30 23:22:53 msilap kernel: [ 7682.404411]  [<ffffffff81379d94>] pci_stop_bus_device+0x94/0xa0
Jan 30 23:22:53 msilap kernel: [ 7682.404414]  [<ffffffff81379f36>] pci_stop_and_remove_bus_device+0x16/0x30
Jan 30 23:22:53 msilap kernel: [ 7682.404419]  [<ffffffffa0028b36>] acpiphp_disable_slot+0xf6/0x1a0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404424]  [<ffffffffa0027716>] ? get_slot_status+0x46/0xc0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404428]  [<ffffffffa0028c0d>] acpiphp_check_bridge.isra.13+0x2d/0xf0 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404433]  [<ffffffffa0029247>] _handle_hotplug_event_bridge+0x2e7/0x410 [acpiphp]
Jan 30 23:22:53 msilap kernel: [ 7682.404436]  [<ffffffff813bdc64>] ? acpi_os_execute_deferred+0x2d/0x32
Jan 30 23:22:53 msilap kernel: [ 7682.404439]  [<ffffffff8117df25>] ? kfree+0x105/0x140
Quoted 8 lines
Jan 30 23:22:53 msilap kernel: [ 7682.404441]  [<ffffffff81077249>] process_one_work+0x159/0x4c0
Jan 30 23:22:53 msilap kernel: [ 7682.404444]  [<ffffffff8107895e>] worker_thread+0x15e/0x4a0
Jan 30 23:22:53 msilap kernel: [ 7682.404446]  [<ffffffff81078800>] ? manage_workers+0x2a0/0x2a0
Jan 30 23:22:53 msilap kernel: [ 7682.404448]  [<ffffffff8107df60>] kthread+0xc0/0xd0
Jan 30 23:22:53 msilap kernel: [ 7682.404450]  [<ffffffff8107dea0>] ? kthread_create_on_node+0x130/0x130
Jan 30 23:22:53 msilap kernel: [ 7682.404452]  [<ffffffff816d756c>] ret_from_fork+0x7c/0xb0
Jan 30 23:22:53 msilap kernel: [ 7682.404454]  [<ffffffff8107dea0>] ? kthread_create_on_node+0x130/0x130
Jan 30 23:22:53 msilap kernel: [ 7682.404456] ---[ end trace 1a4f0a37b1022aae ]---

As a side note: The sdhci driver (which is used by the sdhci-pci driver here) is featuring the same issue as firewire-ohci does: It is calling free_irq() in its suspend() method, and this is in conflict with the free_irq() in its remove() method.

Quoted 37 lines
Quoted 3 lines

Also, what are the parent devices of the OHCI according to "lspci -tv" (or what else can show PCI topology)?

I ran 'lshw' on the machine, which may be of more use. The relevant output is (removing parts not associated with the firewire device):

msilap

    description: Notebook
    product: MS-1727 (To be filled by O.E.M.)
    vendor: MICRO-STAR INTERNATIONAL CO., LTD
    version: REV:1.0
    serial: FFFFFFFF
    width: 64 bits
    capabilities: smbios-2.6 dmi-2.6 vsyscall32
    configuration: boot=normal chassis=notebook family=To be filled by O.E.M. sku=To be filled by O.E.M. uuid=00020003-0004-0005-0006-000700080009

*-core
description: Motherboard
product: MS-1727
vendor: MICRO-STAR INTERNATIONAL CO., LTD
physical id: 0
version: REV:1.0
serial: BSS-0123456789
slot: To be filled by O.E.M.
*-pci:0
description: Host bridge
product: Core Processor DMI
vendor: Intel Corporation
physical id: 100
bus info: pci@0000:00:00.0
version: 11
width: 32 bits
clock: 33MHz
*-pci:4
description: PCI bridge
product: 5 Series/3400 Series Chipset PCI Express Root Port 5
vendor: Intel Corporation

Quoted 8 lines

physical id: 1c.4
bus info: pci@0000:00:1c.4
version: 05
width: 32 bits
clock: 33MHz
capabilities: pci pciexpress msi pm normal_decode bus_master cap_list
configuration: driver=pcieport
resources: irq:16 ioport:9000(size=4096) memory:d4600000-d59fffff ioport:d9d00000(size=2097152)

So, is this Intel PCIe Root Complex acting up and losing contact to other endpoints at resume, or is the JMicron endpoint pretending to be nonexistent at resume, or what else is happening?

Quoted 52 lines

*-firewire
description: FireWire (IEEE 1394)
product: IEEE 1394 Host Controller
vendor: JMicron Technology Corp.
physical id: 0
bus info: pci@0000:07:00.0
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi ohci bus_master cap_list
configuration: driver=firewire_ohci latency=0
resources: irq:16 memory:d4607000-d46077ff memory:d4606000-d460607f memory:d4605000-d460507f memory:d4604000-d460407f
*-generic:0
description: System peripheral
product: SD/MMC Host Controller
vendor: JMicron Technology Corp.
physical id: 0.1
bus info: pci@0000:07:00.1
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: driver=sdhci-pci latency=0
resources: irq:16 memory:d4603000-d46030ff
*-generic:1 UNCLAIMED
description: SD Host controller
product: Standard SD Host Controller
vendor: JMicron Technology Corp.
physical id: 0.2
bus info: pci@0000:07:00.2
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi cap_list
configuration: latency=0
resources: memory:d4602000-d46020ff
*-generic:2
description: System peripheral
product: MS Host Controller
vendor: JMicron Technology Corp.
physical id: 0.3
bus info: pci@0000:07:00.3
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: driver=jmb38x_ms latency=0
resources: irq:16 memory:d4601000-d46010ff
*-generic:3 UNCLAIMED
description: System peripheral
product: xD Host Controller
vendor: JMicron Technology Corp.

Quoted 8 lines

physical id: 0.4
bus info: pci@0000:07:00.4
version: 00
width: 32 bits
clock: 33MHz
capabilities: pm pciexpress msi bus_master cap_list
configuration: latency=0
resources: memory:d4600000-d46000ff

This seems to be a JMicron JMB388, which is a PCIe combination device with OHCI-1394 1.1 link and 1394A:2000 PHY, and a 4-in-1 memory card reader for MMC/SD...SDXC/MS/xD cards.

Signature · 3 lines
Stefan Richter
-=====-===-= --=- ---=-
http://arcgraph.de/sr/
#16 lore raw
Quoted 7 lines · Stefan Richter

OK. I, as a naive driver maintainer, expect a sequence like
[suspend request] -> suspend() -> [if ret==0, suspended state] -> [resume request] -> resume() -> ...
but what apparently happens is
[suspend request] -> suspend() -> [if ret==0, suspended state] -> [resume request] -> remove() -> ...

Quoted 5 lines

If you had an ExpressCard or CardBus controller, then it could of course happen that the PC is suspended, then the controller card physically removed, and then a resume been requested. Question to all: Would the driver's resume() method be called on the MIA device be called, or would the driver's remove() method be called in such a case?

It depends on the subsystem. While processing a resume callback for a device (or for the device's parent), if the subsystem detects that the device has been removed then most likely it would skip calling the driver's resume method and would call the remove method instead.

Alan Stern

maintainer
#17 lore raw
In reply to Mark Einon
Quoted 39 lines · Mark Einon

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

[  262.299486] WARNING: at /build/buildd/linux-3.5.0/kernel/irq/manage.c:1198 __free_irq+0xa3/0x1e0()
[  262.299487] Hardware name: MS-1727
[  262.299488] Trying to free already-free IRQ 16
[  262.299488] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge stp llc joydev arc4 parport_pc ppdev coretemp kvm_intel snd_hda_codec_realtek kvm microcode snd_seq_midi snd_rawmidi psmouse snd_seq_midi_event serio_raw nvidia(PO) i7core_edac snd_seq edac_core snd_hda_intel iwlwifi snd_hda_codec jmb38x_ms snd_hwdep mac80211 snd_pcm lpc_ich memstick snd_seq_device cfg80211 snd_timer snd soundcore snd_page_alloc ir_lirc_codec lirc_dev ir_mce_kbd_decoder ir_sanyo_decoder ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder rfcomm bnep rc_rc6_mce mxm_wmi ene_ir rc_core bluetooth wmi video acpiphp mac_hid lp parport binfmt_misc hid_generic usbhid hid firewire_ohci sdhci_pci r8169 firewire_core sdhci crc_itu_t
[  262.299537] Pid: 4, comm: kworker/0:0 Tainted: P           O 3.5.0-22-generic #34-Ubuntu
[  262.299538] Call Trace:
[  262.299540]  [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[  262.299545]  [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[  262.299548]  [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[  262.299551]  [<ffffffff810df6b3>] __free_irq+0xa3/0x1e0
[  262.299554]  [<ffffffff810df844>] free_irq+0x54/0xc0
[  262.299558]  [<ffffffffa005a27e>] pci_remove+0x6e/0x210 [firewire_ohci]
[  262.299564]  [<ffffffff8135ae7f>] pci_device_remove+0x3f/0x110
[  262.299569]  [<ffffffff8141fdbc>] __device_release_driver+0x7c/0xe0
[  262.299573]  [<ffffffff8141fe4c>] device_release_driver+0x2c/0x40
[  262.299576]  [<ffffffff8141f5f1>] bus_remove_device+0xe1/0x120
[  262.299578]  [<ffffffff8141cd1a>] device_del+0x12a/0x1c0
[  262.299581]  [<ffffffff8141cdc6>] device_unregister+0x16/0x30
[  262.299583]  [<ffffffff81354784>] pci_stop_bus_device+0x94/0xa0
[  262.299588]  [<ffffffffa0091c67>] acpiphp_disable_slot+0xb7/0x1a0 [acpiphp]
[  262.299594]  [<ffffffffa0090716>] ? get_slot_status+0x46/0xc0 [acpiphp]
[  262.299599]  [<ffffffffa0091d7d>] acpiphp_check_bridge.isra.15+0x2d/0xf0 [acpiphp]
[  262.299604]  [<ffffffffa0092442>] _handle_hotplug_event_bridge+0x372/0x4d0 [acpiphp]
[  262.299608]  [<ffffffff81390f8c>] ? acpi_os_execute_deferred+0x2f/0x34
[  262.299612]  [<ffffffff8116e22d>] ? kfree+0xed/0x110
[  262.299617]  [<ffffffff8107086a>] process_one_work+0x12a/0x420
[  262.299620]  [<ffffffffa00920d0>] ? _handle_hotplug_event_func+0x1d0/0x1d0 [acpiphp]
[  262.299625]  [<ffffffff8107141e>] worker_thread+0x12e/0x2f0
[  262.299627]  [<ffffffff810712f0>] ? manage_workers.isra.26+0x200/0x200
[  262.299629]  [<ffffffff81075f13>] kthread+0x93/0xa0
[  262.299632]  [<ffffffff8168d024>] kernel_thread_helper+0x4/0x10
[  262.299636]  [<ffffffff81075e80>] ? kthread_freezable_should_stop+0x70/0x70
[  262.299639]  [<ffffffff8168d020>] ? gs_change+0x13/0x13
[  262.299641] ---[ end trace 3f830890e076911f ]---
Quoted 5 lines
---
 drivers/firewire/ohci.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

<formletter>

This is not the correct way to submit patches for inclusion in the stable kernel tree. Please read Documentation/stable_kernel_rules.txt for how to do this properly.

</formletter>

author
#18 lore raw

PATCH v2 1/1 firewire: Fix ohci free_irq() warning

In reply to Mark Einon

This patch fixes the kernel warning generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_remove() and once then in pci_suspend().

[  262.299486] WARNING: at /build/buildd/linux-3.5.0/kernel/irq/manage.c:1198 __free_irq+0xa3/0x1e0()
[  262.299487] Hardware name: MS-1727
[  262.299488] Trying to free already-free IRQ 16
[  262.299488] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge stp llc joydev arc4 parport_pc ppdev coretemp kvm_intel snd_hda_codec_realtek kvm microcode snd_seq_midi snd_rawmidi psmouse snd_seq_midi_event serio_raw nvidia(PO) i7core_edac snd_seq edac_core snd_hda_intel iwlwifi snd_hda_codec jmb38x_ms snd_hwdep mac80211 snd_pcm lpc_ich memstick snd_seq_device cfg80211 snd_timer snd soundcore snd_page_alloc ir_lirc_codec lirc_dev ir_mce_kbd_decoder ir_sanyo_decoder ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder rfcomm bnep rc_rc6_mce mxm_wmi ene_ir 
 rc_core bluetooth wmi video acpiphp mac_hid lp parport binfmt_misc hid_generic usbhid hid firewire_ohci sdhci_pci r8169 firewire_core sdhci crc_itu_t
[  262.299537] Pid: 4, comm: kworker/0:0 Tainted: P           O 3.5.0-22-generic #34-Ubuntu
[  262.299538] Call Trace:
[  262.299540]  [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[  262.299545]  [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[  262.299548]  [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[  262.299551]  [<ffffffff810df6b3>] __free_irq+0xa3/0x1e0
[  262.299554]  [<ffffffff810df844>] free_irq+0x54/0xc0
[  262.299558]  [<ffffffffa005a27e>] pci_remove+0x6e/0x210 [firewire_ohci]
[  262.299564]  [<ffffffff8135ae7f>] pci_device_remove+0x3f/0x110
[  262.299569]  [<ffffffff8141fdbc>] __device_release_driver+0x7c/0xe0
[  262.299573]  [<ffffffff8141fe4c>] device_release_driver+0x2c/0x40
[  262.299576]  [<ffffffff8141f5f1>] bus_remove_device+0xe1/0x120
[  262.299578]  [<ffffffff8141cd1a>] device_del+0x12a/0x1c0
[  262.299581]  [<ffffffff8141cdc6>] device_unregister+0x16/0x30
[  262.299583]  [<ffffffff81354784>] pci_stop_bus_device+0x94/0xa0
[  262.299588]  [<ffffffffa0091c67>] acpiphp_disable_slot+0xb7/0x1a0 [acpiphp]
[  262.299594]  [<ffffffffa0090716>] ? get_slot_status+0x46/0xc0 [acpiphp]
[  262.299599]  [<ffffffffa0091d7d>] acpiphp_check_bridge.isra.15+0x2d/0xf0 [acpiphp]
[  262.299604]  [<ffffffffa0092442>] _handle_hotplug_event_bridge+0x372/0x4d0 [acpiphp]
[  262.299608]  [<ffffffff81390f8c>] ? acpi_os_execute_deferred+0x2f/0x34
[  262.299612]  [<ffffffff8116e22d>] ? kfree+0xed/0x110
[  262.299617]  [<ffffffff8107086a>] process_one_work+0x12a/0x420
[  262.299620]  [<ffffffffa00920d0>] ? _handle_hotplug_event_func+0x1d0/0x1d0 [acpiphp]
[  262.299625]  [<ffffffff8107141e>] worker_thread+0x12e/0x2f0
[  262.299627]  [<ffffffff810712f0>] ? manage_workers.isra.26+0x200/0x200
[  262.299629]  [<ffffffff81075f13>] kthread+0x93/0xa0
[  262.299632]  [<ffffffff8168d024>] kernel_thread_helper+0x4/0x10
[  262.299636]  [<ffffffff81075e80>] ? kthread_freezable_should_stop+0x70/0x70
[  262.299639]  [<ffffffff8168d020>] ? gs_change+0x13/0x13
[  262.299641] ---[ end trace 3f830890e076911f ]---
---
 drivers/firewire/ohci.c |   16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)
drivers/firewire/ohci.c +12 -4
index 45912e6..8e1c7a9 100644
@@ -196,6 +196,7 @@ struct fw_ohci {
 	bool csr_state_setclear_abdicate;
 	int n_ir;
 	int n_it;
+	bool irq_requested; /* pci dev irq request successful, not freed */
 	/*
 	 * Spinlock for accessing fw_ohci data.  Never call out of
 	 * this driver with this lock held.
@@ -2384,7 +2385,7 @@ static int ohci_enable(struct fw_card *card,
 
 	if (!(ohci->quirks & QUIRK_NO_MSI))
 		pci_enable_msi(dev);
-	if (request_irq(dev->irq, irq_handler,
+	if (!ohci->irq_requested && request_irq(dev->irq, irq_handler,
 			pci_dev_msi_enabled(dev) ? 0 : IRQF_SHARED,
 			ohci_driver_name, ohci)) {
 		dev_err(card->device, "failed to allocate interrupt %d\n",
@@ -2398,6 +2399,8 @@ static int ohci_enable(struct fw_card *card,
 			ohci->next_config_rom = NULL;
 		}
 		return -EIO;
+	} else {
+		ohci->irq_requested = true;
 	}
 
 	irqs =	OHCI1394_reqTxComplete | OHCI1394_respTxComplete |
@@ -3733,8 +3736,10 @@ static void pci_remove(struct pci_dev *dev)
 	 */
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
-
+	if (ohci->irq_requested) {
+		free_irq(dev->irq, ohci);
+		ohci->irq_requested = false;
+	}
 	if (ohci->next_config_rom && ohci->next_config_rom != ohci->config_rom)
 		dma_free_coherent(ohci->card.device, CONFIG_ROM_SIZE,
 				  ohci->next_config_rom, ohci->next_config_rom_bus);
@@ -3766,7 +3771,10 @@ static int pci_suspend(struct pci_dev *dev, pm_message_t state)
 	int err;
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
+	if (ohci->irq_requested) {
+		free_irq(dev->irq, ohci);
+		ohci->irq_requested = false;
+	}
 	pci_disable_msi(dev);
 	err = pci_save_state(dev);
 	if (err) {
Signature · 8 lines
1.7.10.4


------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_jan
author
#19 lore raw

PATCH v3 1/1 firewire: Fix ohci free_irq() warning

This patch fixes the kernel warning below, generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_suspend() and once then in pci_remove().

The fix breaks up the ohci_enable() call into four separate calls - ohci_enable_regs(), ohci_create_irq(), ohci_enable_irq() and ohci_enable_run(). The original call sequence of ohci_enable() is replaced by the four calls, but the ohci_enable() called from pci_resume() is replaced by three of the calls, missing out ohci_create_irq().

Finally, a new call, ohci_disable_irq(), replaces the free_irq() in pci_suspend() which sets the irq mask to zero.

[  262.299486] WARNING: at /build/buildd/linux-3.5.0/kernel/irq/manage.c:1198 __free_irq+0xa3/0x1e0()
[  262.299487] Hardware name: MS-1727
[  262.299488] Trying to free already-free IRQ 16
[  262.299488] Modules linked in: ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge stp llc joydev arc4 parport_pc ppdev coretemp kvm_intel snd_hda_codec_realtek kvm microcode snd_seq_midi snd_rawmidi psmouse snd_seq_midi_event serio_raw nvidia(PO) i7core_edac snd_seq edac_core snd_hda_intel iwlwifi snd_hda_codec jmb38x_ms snd_hwdep mac80211 snd_pcm lpc_ich memstick snd_seq_device cfg80211 snd_timer snd soundcore snd_page_alloc ir_lirc_codec lirc_dev ir_mce_kbd_decoder ir_sanyo_decoder ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder rfcomm bnep rc_rc6_mce mxm_wmi ene_ir 
 rc_core bluetooth wmi video acpiphp mac_hid lp parport binfmt_misc hid_generic usbhid hid firewire_ohci sdhci_pci r8169 firewire_core sdhci crc_itu_t
[  262.299537] Pid: 4, comm: kworker/0:0 Tainted: P           O 3.5.0-22-generic #34-Ubuntu
[  262.299538] Call Trace:
[  262.299540]  [<ffffffff81051c1f>] warn_slowpath_common+0x7f/0xc0
[  262.299545]  [<ffffffff81051d16>] warn_slowpath_fmt+0x46/0x50
[  262.299548]  [<ffffffff8103fa39>] ? default_spin_lock_flags+0x9/0x10
[  262.299551]  [<ffffffff810df6b3>] __free_irq+0xa3/0x1e0
[  262.299554]  [<ffffffff810df844>] free_irq+0x54/0xc0
[  262.299558]  [<ffffffffa005a27e>] pci_remove+0x6e/0x210 [firewire_ohci]
[  262.299564]  [<ffffffff8135ae7f>] pci_device_remove+0x3f/0x110
[  262.299569]  [<ffffffff8141fdbc>] __device_release_driver+0x7c/0xe0
[  262.299573]  [<ffffffff8141fe4c>] device_release_driver+0x2c/0x40
[  262.299576]  [<ffffffff8141f5f1>] bus_remove_device+0xe1/0x120
[  262.299578]  [<ffffffff8141cd1a>] device_del+0x12a/0x1c0
[  262.299581]  [<ffffffff8141cdc6>] device_unregister+0x16/0x30
[  262.299583]  [<ffffffff81354784>] pci_stop_bus_device+0x94/0xa0
[  262.299588]  [<ffffffffa0091c67>] acpiphp_disable_slot+0xb7/0x1a0 [acpiphp]
[  262.299594]  [<ffffffffa0090716>] ? get_slot_status+0x46/0xc0 [acpiphp]
[  262.299599]  [<ffffffffa0091d7d>] acpiphp_check_bridge.isra.15+0x2d/0xf0 [acpiphp]
[  262.299604]  [<ffffffffa0092442>] _handle_hotplug_event_bridge+0x372/0x4d0 [acpiphp]
[  262.299608]  [<ffffffff81390f8c>] ? acpi_os_execute_deferred+0x2f/0x34
[  262.299612]  [<ffffffff8116e22d>] ? kfree+0xed/0x110
[  262.299617]  [<ffffffff8107086a>] process_one_work+0x12a/0x420
[  262.299620]  [<ffffffffa00920d0>] ? _handle_hotplug_event_func+0x1d0/0x1d0 [acpiphp]
[  262.299625]  [<ffffffff8107141e>] worker_thread+0x12e/0x2f0
[  262.299627]  [<ffffffff810712f0>] ? manage_workers.isra.26+0x200/0x200
[  262.299629]  [<ffffffff81075f13>] kthread+0x93/0xa0
[  262.299632]  [<ffffffff8168d024>] kernel_thread_helper+0x4/0x10
[  262.299636]  [<ffffffff81075e80>] ? kthread_freezable_should_stop+0x70/0x70
[  262.299639]  [<ffffffff8168d020>] ? gs_change+0x13/0x13
[  262.299641] ---[ end trace 3f830890e076911f ]---

---

Note - This patch has only been tested with a firewire controller, and not subsequently with any devices plugged into the controller.

 drivers/firewire/ohci.c |  134 ++++++++++++++++++++++++++++++++---------------
 1 file changed, 91 insertions(+), 43 deletions(-)
drivers/firewire/ohci.c +91 -43
index 45912e6..56814b8 100644
@@ -2242,12 +2242,69 @@ static int probe_tsb41ba3d(struct fw_ohci *ohci)
 	return 1;
 }
 
-static int ohci_enable(struct fw_card *card,
-		       const __be32 *config_rom, size_t length)
+static int ohci_create_irq(struct fw_card *card)
 {
 	struct fw_ohci *ohci = fw_ohci(card);
 	struct pci_dev *dev = to_pci_dev(card->device);
-	u32 lps, version, irqs;
+
+	if (!(ohci->quirks & QUIRK_NO_MSI))
+		pci_enable_msi(dev);
+	if (request_irq(dev->irq, irq_handler,
+			pci_dev_msi_enabled(dev) ? 0 : IRQF_SHARED,
+			ohci_driver_name, ohci)) {
+		dev_err(card->device, "failed to allocate interrupt %d\n",
+			dev->irq);
+		pci_disable_msi(dev);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static void ohci_disable_irq(struct fw_ohci *ohci)
+{
+	reg_write(ohci, OHCI1394_IntMaskSet, 0);
+
+	reg_write(ohci, OHCI1394_HCControlSet,
+		  ~(OHCI1394_HCControl_linkEnable &
+		  OHCI1394_HCControl_BIBimageValid));
+
+	reg_write(ohci, OHCI1394_LinkControlSet,
+		  ~(OHCI1394_LinkControl_rcvSelfID &
+		  OHCI1394_LinkControl_rcvPhyPkt));
+}
+
+static void ohci_enable_irq(struct fw_ohci *ohci)
+{
+	int irqs = OHCI1394_reqTxComplete | OHCI1394_respTxComplete |
+		   OHCI1394_RQPkt | OHCI1394_RSPkt |
+		   OHCI1394_isochTx | OHCI1394_isochRx |
+		   OHCI1394_postedWriteErr |
+		   OHCI1394_selfIDComplete |
+		   OHCI1394_regAccessFail |
+		   OHCI1394_cycleInconsistent |
+		   OHCI1394_unrecoverableError |
+		   OHCI1394_cycleTooLong |
+		   OHCI1394_masterIntEnable;
+
+	if (param_debug & OHCI_PARAM_DEBUG_BUSRESETS)
+		irqs |= OHCI1394_busReset;
+	reg_write(ohci, OHCI1394_IntMaskSet, irqs);
+
+	reg_write(ohci, OHCI1394_HCControlSet,
+		  OHCI1394_HCControl_linkEnable |
+		  OHCI1394_HCControl_BIBimageValid);
+
+	reg_write(ohci, OHCI1394_LinkControlSet,
+		  OHCI1394_LinkControl_rcvSelfID |
+		  OHCI1394_LinkControl_rcvPhyPkt);
+}
+
+static int ohci_enable_regs(struct fw_card *card,
+			    const __be32 *config_rom, size_t length)
+{
+	struct fw_ohci *ohci = fw_ohci(card);
+	u32 lps, version;
 	int i, ret;
 
 	if (software_reset(ohci)) {
@@ -2382,53 +2439,41 @@ static int ohci_enable(struct fw_card *card,
 
 	reg_write(ohci, OHCI1394_AsReqFilterHiSet, 0x80000000);
 
-	if (!(ohci->quirks & QUIRK_NO_MSI))
-		pci_enable_msi(dev);
-	if (request_irq(dev->irq, irq_handler,
-			pci_dev_msi_enabled(dev) ? 0 : IRQF_SHARED,
-			ohci_driver_name, ohci)) {
-		dev_err(card->device, "failed to allocate interrupt %d\n",
-			dev->irq);
-		pci_disable_msi(dev);
+	return 0;
+}
+
+static void ohci_enable_run(struct fw_ohci *ohci)
+{
+	ar_context_run(&ohci->ar_request_ctx);
+	ar_context_run(&ohci->ar_response_ctx);
 
+	flush_writes(ohci);
+
+	/* We are ready to go, reset bus to finish initialization. */
+	fw_schedule_bus_reset(&ohci->card, false, true);
+}
+
+static int ohci_enable(struct fw_card *card,
+		       const __be32 *config_rom, size_t length)
+{
+	struct fw_ohci *ohci = fw_ohci(card);
+	int ret;
+
+	ohci_enable_regs(card, config_rom, length);
+
+	ret = ohci_create_irq(card);
+	if (ret) {
 		if (config_rom) {
 			dma_free_coherent(ohci->card.device, CONFIG_ROM_SIZE,
 					  ohci->next_config_rom,
 					  ohci->next_config_rom_bus);
 			ohci->next_config_rom = NULL;
 		}
-		return -EIO;
+		return ret;
 	}
 
-	irqs =	OHCI1394_reqTxComplete | OHCI1394_respTxComplete |
-		OHCI1394_RQPkt | OHCI1394_RSPkt |
-		OHCI1394_isochTx | OHCI1394_isochRx |
-		OHCI1394_postedWriteErr |
-		OHCI1394_selfIDComplete |
-		OHCI1394_regAccessFail |
-		OHCI1394_cycleInconsistent |
-		OHCI1394_unrecoverableError |
-		OHCI1394_cycleTooLong |
-		OHCI1394_masterIntEnable;
-	if (param_debug & OHCI_PARAM_DEBUG_BUSRESETS)
-		irqs |= OHCI1394_busReset;
-	reg_write(ohci, OHCI1394_IntMaskSet, irqs);
-
-	reg_write(ohci, OHCI1394_HCControlSet,
-		  OHCI1394_HCControl_linkEnable |
-		  OHCI1394_HCControl_BIBimageValid);
-
-	reg_write(ohci, OHCI1394_LinkControlSet,
-		  OHCI1394_LinkControl_rcvSelfID |
-		  OHCI1394_LinkControl_rcvPhyPkt);
-
-	ar_context_run(&ohci->ar_request_ctx);
-	ar_context_run(&ohci->ar_response_ctx);
-
-	flush_writes(ohci);
-
-	/* We are ready to go, reset bus to finish initialization. */
-	fw_schedule_bus_reset(&ohci->card, false, true);
+	ohci_enable_irq(ohci);
+	ohci_enable_run(ohci);
 
 	return 0;
 }
@@ -3766,7 +3811,7 @@ static int pci_suspend(struct pci_dev *dev, pm_message_t state)
 	int err;
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
+	ohci_disable_irq(ohci);
 	pci_disable_msi(dev);
 	err = pci_save_state(dev);
 	if (err) {
@@ -3802,10 +3847,13 @@ static int pci_resume(struct pci_dev *dev)
 		reg_write(ohci, OHCI1394_GUIDHi, (u32)(ohci->card.guid >> 32));
 	}
 
-	err = ohci_enable(&ohci->card, NULL, 0);
+	err = ohci_enable_regs(&ohci->card, NULL, 0);
 	if (err)
 		return err;
 
+	ohci_enable_irq(ohci);
+	ohci_enable_run(ohci);
+
 	ohci_resume_iso_dma(ohci);
 
 	return 0;
Signature · 8 lines
1.7.10.4


------------------------------------------------------------------------------
Free Next-Gen Firewall Hardware Offer
Buy your Sophos next-gen firewall before the end March 2013 
and get the hardware for free! Learn more.
http://p.sf.net/sfu/sophos-d2d-feb
#20 lore raw
Quoted 5 lines · On Feb 05 Mark Einon wrote:

This patch fixes the kernel warning below, generated when putting an MSI MS-1727 GT740 laptop into suspend mode. The call sequence in this case calls free_irq() twice, once in pci_suspend() and once then in pci_remove().

Quoted 8 lines

The fix breaks up the ohci_enable() call into four separate calls - ohci_enable_regs(), ohci_create_irq(), ohci_enable_irq() and ohci_enable_run(). The original call sequence of ohci_enable() is replaced by the four calls, but the ohci_enable() called from pci_resume() is replaced by three of the calls, missing out ohci_create_irq().

Finally, a new call, ohci_disable_irq(), replaces the free_irq() in pci_suspend() which sets the irq mask to zero.

[...]

Hi,
I haven't forgotten about this patch; just didn't fully check it yet because I still wanted to look whether there is a more minimalistic way. Perhaps I'll post something in the next days.

Quoted 14 lines
drivers/firewire/ohci.c +1 -2
@@ -3766,7 +3811,7 @@ static int pci_suspend(struct pci_dev *dev, pm_message_t state)
 	int err;
 
 	software_reset(ohci);
-	free_irq(dev->irq, ohci);
+	ohci_disable_irq(ohci);
 	pci_disable_msi(dev);
 	err = pci_save_state(dev);
 	if (err) {
@@ -3802,10 +3847,13 @@ static int pci_resume(struct pci_dev *dev)
 		reg_write(ohci, OHCI1394_GUIDHi, (u32)(ohci->card.guid >> 32));
 	}
 
-	err = ohci_enable(&ohci->card, NULL, 0);

drivers/firewire/ohci.c static int pci_resume(struct pci_dev *dev) · line 3859

Quoted 10 lines
drivers/firewire/ohci.c +4 -0
+	err = ohci_enable_regs(&ohci->card, NULL, 0);
 	if (err)
 		return err;
 
+	ohci_enable_irq(ohci);
+	ohci_enable_run(ohci);
+
 	ohci_resume_iso_dma(ohci);
 
 	return 0;

Looks like pci_{en,dis}able_msi are now unbalanced. I suppose we better place them immediately around {request,free}_irq.

Signature · 3 lines
Stefan Richter
-=====-===-= --=- =---=
http://arcgraph.de/sr/