Bug 81551

Summary:	ATAPI-CD-ROM-drive dead after resume from suspend/s2disk - Jmicron ctrlr vs async suspend
Product:	Power Management	Reporter:	Rainer Fiebig (jrf)
Component:	Hibernation/Suspend	Assignee:	Zhang Rui (rui.zhang)
Status:	CLOSED CODE_FIX
Severity:	normal	CC:	aaron.lu, acassis, chuansheng.liu, lenb, mister.freeman, rjw, rui.zhang, stern
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	3.15	Subsystem:
Regression:	Yes	Bisected commit-id:
Attachments:	Bug_81551_Kernel_3.15.9_Log Bug_81551_Kernel_3.14.17_Log Bug_81551_Kernel_3.15.9_IRQs 2014-08-25_dmesg_patch_2_60ms 2014-08-25_dmesg_patch_2_70ms 2014-08-25_dmesg_patch_2_20ms_pm_async_0 2014-08-25_dmesg_patch_2_70ms_complete 2014-08-25_dmesg_patch_2_20ms_pm_async_1 2014-08-27_dmesg_patch_3a 2014-08-27_dmesg_patch_3a_2 Bug 81551_contents_sys_devices 2014-08-27_dmesg_patch_3a_both_async_disabled Disabling the async for 363/361 SATA/PATA controllers dmesg output after a standby dmesg output with kernel 3.17rc06 after standby, bug is still here the final patch for JMB JMicron family IDE/SATA controler disable_async_suspend_jmicron by using /drivers/pci/quirks.c file patch fix jmicron controller dependency Patch for upstream

Description Rainer Fiebig 2014-08-02 13:18:42 UTC

Good day!

After resume from suspend (I use "save to both") my PC's ATAPI-CD-ROM-drive doesn't respond any more.

I get this message "syslogd: "kernel:[2784.783099] Disabling IRQ #17". This is the IRQ that the modularized adapter-driver (JMicron) uses.

This happens with an unmodified distribution kernel (openSuse "stable", 3.15.7) and with all self-built kernels since 3.15.(0).

This issue does not exist with 3.14 and earlier (same hardware/same kernel-configuration).

Kind regards

Jay

Comment 1 Aaron Lu 2014-08-04 02:48:08 UTC

Possible to do a git bisect from v3.14 to v3.15?

Comment 2 Rainer Fiebig 2014-08-04 12:11:02 UTC

(In reply to Aaron Lu from comment #1)
> Possible to do a git bisect from v3.14 to v3.15?


I'm afraid a git-bisect is beyond my current abilities and it would certainly take me some time to figure out how to do it.


Perhaps I should have mentioned that the PC also has an SCSI-CD-ROM-drive which seems to share IRQ #17 with the ATAPI-drive. But this SCSI-drive is absolutely OK after resume. Only the IDE/ATAPI-drive doesn't work.

I don't want to lead you in a wrong direction but when I encountered this problem my first thought was that it might have something to do with this new async-feature:

http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=200421a80f6e0a9e39d698944cc35cba103eb6ce

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=3c31b52f96f7b559d950b16113c0f68c72a1985e

But I may be completely wrong and apologize in advance to the developers!


Jay

Comment 3 Rainer Fiebig 2014-08-08 14:41:25 UTC

Maybe this excerpt from the kernel-log (3.15.8) is of help.

After system-start:

...
systemd-udevd[335]	starting version 195
BIOS EDD facility v0.16 2004-Jun-25, 2 devices found
pata_jmicron 0000	2:00.1: enabling device (0000 -> 0001)
scsi8 	pata_jmicron
scsi9 	pata_jmicron
ata9	PATA max UDMA/100 cmd 0xcc00 ctl 0xc880 bmdma 0xc400 irq 17
ata10	PATA max UDMA/100 cmd 0xc800 ctl 0xc480 bmdma 0xc408 irq 17
i801_smbus 0000	0:1f.3: SMBus using PCI Interrupt
ACPI Warning	SystemIO range 0x0000000000000828-0x000000000000082f conflicts with OpRegion 0x0000000000000800-0x000000000000084f (\PMRG) (20140214/utaddress-258)
ACPI	If an ACPI driver is available for this device, you should use it instead of the native driver
lpc_ich	Resource conflict(s) found affecting gpio_ich
input	PC Speaker as /devices/platform/pcspkr/input/input4
sd 0	:0:0: Attached scsi generic sg0 type 0
sd 1	:0:0: Attached scsi generic sg1 type 0
...

After resume:

...
CPU3 is up
ACPI	Waking up from system sleep state S3
ehci-pci 0000	0:1d.0: System wakeup disabled by ACPI
pata_jmicron 0000	2:00.1: Refused to change power state, currently in D3
ehci-pci 0000	0:1a.0: System wakeup disabled by ACPI
PM	noirq resume of devices complete after 61.868 msecs
PM	early resume of devices complete after 0.221 msecs
...

Jay

Comment 4 Aaron Lu 2014-08-11 02:48:19 UTC

Is building kernel possible for you?

Comment 5 Rainer Fiebig 2014-08-11 07:36:53 UTC

(In reply to Aaron Lu from comment #4)
> Is building kernel possible for you?

Yes.

Comment 6 Rainer Fiebig 2014-08-18 22:31:50 UTC

You live and learn - as they say.

I did a "git bisect" between 3.14 and 3.15.

This was the outcome after 13 steps:

jr@linux-tez8:~/Downloads/linux-stable> git bisect bad
3d2699bc179a10eee7d2aa1db50f822be01636f7 is the first bad commit
commit 3d2699bc179a10eee7d2aa1db50f822be01636f7
Author: Liu, Chuansheng <chuansheng.liu@intel.com>
Date:   Tue Feb 18 10:28:44 2014 +0800

    PM / sleep: Two flags for async suspend_noirq and suspend_late
    
    The patch is a helper adding two new flags for implementing
    async threads for suspend_noirq and suspend_late.
    
    Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com>
    Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

:040000 040000 8e646f3aca4f23e3f695107f17cc53c830b71be2 e1a01fb5b148342a4251b6374a9ad656067654d4 M      drivers
:040000 040000 a28cd1453c85ef1533dda9ad76ce0a340d3371e8 0d579b7c315c284b272f1c0d1c3916391e983676 M      include
jr@linux-tez8:~/Downloads/linux-stable> 


Looks like a good starting point. But please consider the fact that this was my first "git bisect" - so I may have made a mistake.

Jay

Comment 7 Aaron Lu 2014-08-19 01:33:06 UTC

Good to see this, thanks.

I think you can verify by building the said commit and then the commit immediately before it. If the first kernel failed and the 2nd kernel works, then it is very likely to be the culprit.

Comment 8 Rainer Fiebig 2014-08-19 12:38:08 UTC

It was late yesterday. To be on the safe side I did another "git bisect" today but narrowed the search to "drivers/base/power" (see "git bisect log" below). Obviously I made a mistake yesterday (probably once bootet the wrong kernel) because today I got this result which I checked twice:


jr@linux-tez8:~/Downloads/linux-stable> git bisect good
76569faa62c46382e080c3e190c66e19515aae1c is the first bad commit
commit 76569faa62c46382e080c3e190c66e19515aae1c
Author: Liu, Chuansheng <chuansheng.liu@intel.com>
Date:   Tue Feb 18 10:28:45 2014 +0800

    PM / sleep: Asynchronous threads for resume_noirq
    
    In analogy with commits 5af84b82701a and 97df8c12995, using
    asynchronous threads can improve the overall resume_noirq time
    significantly.
    
    One typical case is:
    In resume_noirq phase and for the PCI devices, the function
    pci_pm_resume_noirq() will be called, and there is one d3_delay
    (10ms) at least.
    
    With the way of asynchronous threads, we just need wait d3_delay
    time once in parallel for each calling, which saves much time to
    resume quickly.
    
    Signed-off-by: Chuansheng Liu <chuansheng.liu@intel.com>
    Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

:040000 040000 e1a01fb5b148342a4251b6374a9ad656067654d4 ab0a5adf507c133d47dd410f8af3a00177f45cee M      drivers
jr@linux-tez8:~/Downloads/linux-stable> 


I'm pretty confident that this is the real culprit.

Jay



---------------------------

jr@linux-tez8:~/Downloads/linux-stable> git bisect log
git bisect start '--' 'drivers/base/power'
# good: [455c6fdbd219161bd09b1165f11699d6d73de11c] Linux 3.14
git bisect good 455c6fdbd219161bd09b1165f11699d6d73de11c
# bad: [1860e379875dfe7271c649058aeddffe5afd9d0d] Linux 3.15
git bisect bad 1860e379875dfe7271c649058aeddffe5afd9d0d
# good: [165f5fd04aa898fa7dbe4fbe2819cf80453d3bb9] Merge branches 'pm-qos', 'pm-domains' and 'pm-drivers'
git bisect good 165f5fd04aa898fa7dbe4fbe2819cf80453d3bb9
# bad: [36cc86e8ec1a0ab309052bc9b25e48a31288b8a5] Merge branches 'pm-runtime' and 'pm-sleep'
git bisect bad 36cc86e8ec1a0ab309052bc9b25e48a31288b8a5
# bad: [9e5e7910df824ba02aedd2b5d2ca556426ea6d0b] PM / sleep: Asynchronous threads for resume_early
git bisect bad 9e5e7910df824ba02aedd2b5d2ca556426ea6d0b
# bad: [76569faa62c46382e080c3e190c66e19515aae1c] PM / sleep: Asynchronous threads for resume_noirq
git bisect bad 76569faa62c46382e080c3e190c66e19515aae1c                                                                              
# good: [3d2699bc179a10eee7d2aa1db50f822be01636f7] PM / sleep: Two flags for async suspend_noirq and suspend_late                    
git bisect good 3d2699bc179a10eee7d2aa1db50f822be01636f7                                                                             
jr@linux-tez8:~/Downloads/linux-stable>  

---------------------------

Comment 9 chuansheng.liu 2014-08-20 02:13:16 UTC

Hi Jay,

Could we get one more detailed log/dmesg for this issue?
You can enable the printk log more detailed level, and disable console suspend
with the below command:

echo "N" > /sys/module/printk/parameters/console_suspend

Thanks.

From the log "kernel:[2784.783099] Disabling IRQ #17", it seems be related with shared irq thing.

Comment 10 Aaron Lu 2014-08-20 02:17:00 UTC

One confusing thing is, from #comment 2:
"
Perhaps I should have mentioned that the PC also has an SCSI-CD-ROM-drive which seems to share IRQ #17 with the ATAPI-drive. But this SCSI-drive is absolutely OK after resume. Only the IDE/ATAPI-drive doesn't work.
"
If IRQ #17 is disabled, how can the SCSI CDROM drive still work well? ...

Comment 11 chuansheng.liu 2014-08-20 02:23:38 UTC

Hi Jay,

BTW, could you share the info of command?
#cat /proc/interrupts


The log "Disabling IRQ #17" mostly means the IRQ #17 is coming out without handling for many times, then kernel irq code disabled it.

Comment 12 Rainer Fiebig 2014-08-20 12:27:17 UTC

Created attachment 147541 [details]
Bug_81551_Kernel_3.15.9_Log

Hi!

I have attached the IRQ-table and two excerpts from the kernel-log which hopefully offer enough detail: one from kernel 3.14.17 (OK), the other from 3.15.9. 

They show what happened after initiating suspend and then resume after a fresh system start. Major differences can be seen from line 98 on.

Have a good day.

Jay

Comment 13 Rainer Fiebig 2014-08-20 12:27:51 UTC

Created attachment 147551 [details]
Bug_81551_Kernel_3.14.17_Log

Comment 14 Rainer Fiebig 2014-08-20 12:28:25 UTC

Created attachment 147561 [details]
Bug_81551_Kernel_3.15.9_IRQs

Comment 15 Rainer Fiebig 2014-08-21 20:28:55 UTC

(In reply to chuansheng.liu from comment #9)
> Hi Jay,
> 
> Could we get one more detailed log/dmesg for this issue?
> You can enable the printk log more detailed level, and disable console
> suspend
> with the below command:
> 
> echo "N" > /sys/module/printk/parameters/console_suspend
> 
> Thanks.
> 
> From the log "kernel:[2784.783099] Disabling IRQ #17", it seems be related
> with shared irq thing.


Hi!

Did a little experiment (kernel 3.16 / has the same problem) which suggests that the issue is probably not the sharing of an IRQ but pata_jmicron not waking up:


To get rid of the SCSI-adapter and the sharing of IRQ #17 I de-selected all SCSI-options except for "*SCSI disk support" and compiled the kernel.

As you can see, IRQ #17 now belonged to pata_jmicron alone:

~> cat /proc/interrupts
           CPU0       CPU1       CPU2       CPU3       
  0:        119          0          0          0   IO-APIC-edge      timer
  1:       1476        186        173        177   IO-APIC-edge      i8042
  8:          0          0          0          1   IO-APIC-edge      rtc0
  9:         13          1          0          1   IO-APIC-fasteoi   acpi
 16:       4279        284        274        277   IO-APIC-fasteoi   ahci, ehci_hcd:usb1, AVM.1
 17:     199995          3          2          1   IO-APIC-fasteoi   pata_jmicron
 18:          0          0          0          0   IO-APIC-fasteoi   i801_smbus
 23:         63         15         17         16   IO-APIC-fasteoi   ehci_hcd:usb2
 40:          6          7          7          6   PCI-MSI-edge      mei_me
 41:      12529       7089       7076       7024   PCI-MSI-edge      ahci
 42:        378         98        104        107   PCI-MSI-edge      eth0
 43:       7850       9087       8685       8144   PCI-MSI-edge      i915
 44:         47         45         43         45   PCI-MSI-edge      snd_hda_intel
NMI:          0          0          0          0   Non-maskable interrupts
LOC:      86674      66333      59531      56750   Local timer interrupts
SPU:          0          0          0          0   Spurious interrupts
PMI:          0          0          0          0   Performance monitoring interrupts

snip------------------------


Even so, during resume after the first suspend, pata_jmicron refuses to change power state, IRQ #17 complains that "nobody cared" and gets disabled:

...
[  297.560221] ACPI: Waking up from system sleep state S3
[  297.571257] ehci-pci 0000:00:1d.0: System wakeup disabled by ACPI
[  297.571268] ehci-pci 0000:00:1a.0: System wakeup disabled by ACPI
[  297.571287] pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3
[  297.622560] PM: noirq resume of devices complete after 62.212 msecs
[  297.622754] PM: early resume of devices complete after 0.173 msecs
[  297.622838] mei_me 0000:00:16.0: irq 40 for MSI/MSI-X
[  297.623084] snd_hda_intel 0000:00:1b.0: irq 44 for MSI/MSI-X
[  297.623276] pcieport 0000:00:1c.5: System wakeup disabled by ACPI
[  297.623454] i8042 kbd 00:02: System wakeup disabled by ACPI
[  297.623970] serial 00:06: activated
[  297.629380] sd 0:0:0:0: [sda] Starting disk
[  297.629396] sd 1:0:0:0: [sdb] Starting disk
[  297.680709] r8169 0000:03:00.0 eth0: link down
[  297.794614] ata9.00: configured for UDMA/33
[  297.852109] i915: No ACPI video bus found
[  297.930086] ata3: SATA link down (SStatus 0 SControl 300)
[  297.935080] ata6: SATA link down (SStatus 0 SControl 300)
[  297.936091] ata5: SATA link down (SStatus 0 SControl 300)
[  297.938085] ata4: SATA link down (SStatus 0 SControl 300)
[  297.939101] ata8: SATA link down (SStatus 0 SControl 300)
[  297.939138] ata7: SATA link down (SStatus 0 SControl 300)
[  297.983093] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  297.985829] ata2.00: configured for UDMA/133
[  299.221098] irq 17: nobody cared (try booting with the "irqpoll" option)
[  299.221101] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.16.0-1.g8038aea-desktop #4
[  299.221101] Hardware name: MSI MS-7637/H55-GD65 (MS-7637)  , BIOS V1.5 09/16/2010
[  299.221103]  0000000000000000 ffff880329779c8c ffffffff81599f59 ffff880329779c00
[  299.221104]  ffffffff810a30f6 ffff880329779c00 0000000000000011 0000000000000000
[  299.221105]  ffffffff810a3636 ffffffffffd1170a 0000000000000000 0000000000000000
[  299.221106] Call Trace:
[  299.221112]  <IRQ>  [<ffffffff81599f59>] ? dump_stack+0x49/0x6a
[  299.221117]  [<ffffffff810a30f6>] ? __report_bad_irq+0x36/0xd0
[  299.221118]  [<ffffffff810a3636>] ? note_interrupt+0x266/0x2c0
[  299.221120]  [<ffffffff810a0db7>] ? handle_irq_event_percpu+0x87/0x1f0
[  299.221121]  [<ffffffff810a0f61>] ? handle_irq_event+0x41/0x70
[  299.221122]  [<ffffffff810a4195>] ? handle_fasteoi_irq+0x85/0x150
[  299.221125]  [<ffffffff81005485>] ? handle_irq+0x15/0x20
[  299.221126]  [<ffffffff81004db8>] ? do_IRQ+0x58/0x110
[  299.221128]  [<ffffffff815a076d>] ? common_interrupt+0x6d/0x6d
[  299.221132]  <EOI>  [<ffffffff81482d49>] ? cpuidle_enter_state+0x59/0xe0
[  299.221133]  [<ffffffff81482d42>] ? cpuidle_enter_state+0x52/0xe0
[  299.221135]  [<ffffffff81090401>] ? cpu_startup_entry+0x2c1/0x490
[  299.221138]  [<ffffffff8182e028>] ? start_kernel+0x454/0x45f
[  299.221139]  [<ffffffff8182d9ec>] ? set_init_arg+0x50/0x50
[  299.221141]  [<ffffffff8182d117>] ? early_idt_handlers+0x117/0x120
[  299.221142]  [<ffffffff8182d117>] ? early_idt_handlers+0x117/0x120
[  299.221143]  [<ffffffff8182d704>] ? x86_64_start_kernel+0x152/0x161
[  299.221144] handlers:
[  299.221147] [<ffffffff8140c3b0>] ata_bmdma_interrupt
[  299.221147] Disabling IRQ #17
[  299.292022] r8169 0000:03:00.0 eth0: link up
[  299.815672] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  299.828839] ata1.00: configured for UDMA/133
[  299.832885] input input1: event field not found
[  299.832887] input input1: event field not found
[  299.832888] input input1: event field not found
[  299.832889] input input1: event field not found
[  299.832891] input input1: event field not found
[  299.832892] input input1: event field not found
[  299.832893] input input1: event field not found
[  299.833015] PM: resume of devices complete after 2210.818 msecs
[  299.944185] Restarting kernel threads ... done.
[  299.944376] Restarting tasks ... done.
[  301.535566] IPv4: martian source 255.255.255.255 from xxx.xxx.x.x, on dev eth0
[  301.535574] ll header: 00000000: ff ff ff ff ff ff 00 0f b5 50 0e 88 08 00        .........P....
[  301.554978] IPv4: martian source 255.255.255.255 from xxx.xxx.x.x, on dev eth0
[  301.554998] ll header: 00000000: ff ff ff ff ff ff 00 0f b5 50 0e 88 08 00        .........P....
[  302.792918] ata9.00: qc timeout (cmd 0xa0)
[  302.792923] ata9.00: TEST_UNIT_READY failed (err_mask=0x5)
[  302.953450] ata9.00: configured for UDMA/33
[  307.951568] ata9.00: qc timeout (cmd 0xa0)
[  307.951577] ata9.00: TEST_UNIT_READY failed (err_mask=0x5)
[  307.951583] ata9.00: limiting speed to UDMA/33:PIO3
[  308.112032] ata9.00: configured for UDMA/33
[  313.110354] ata9.00: qc timeout (cmd 0xa0)
[  313.110363] ata9.00: TEST_UNIT_READY failed (err_mask=0x5)
[  313.110367] ata9.00: disabled
[  313.110396] ata9: soft resetting link
[  313.261355] ata9: EH complete

-----------------------------------------------------

Hope this helps.

Jay

Comment 16 chuansheng.liu 2014-08-22 00:04:49 UTC

Good finding, Jay.

The log "pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3" should be from pci_pm_resume_noirq() calling.

If enabling async, it is possible the pata_jmicron pci_pm_resume_noirq()'s execution is mixed with his parent or children.

So,

Could you help to give one pci tree list? And could you help to check if the device pata_jmicron is enabling async PM?

Thanks.

Comment 17 chuansheng.liu 2014-08-22 00:16:04 UTC

BTW, Jay, could you help to try the below debugging patch to see if it is the issue about d3_delay?

Thanks.

diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 4d1a5d2..fa02003 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -142,6 +142,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
                .port_ops       = &jmicron_ops,
        };
        const struct ata_port_info *ppi[] = { &info, NULL };
+       pdev->d3_delay = 200;
 
        return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
 }

Comment 18 Rainer Fiebig 2014-08-22 12:57:42 UTC

(In reply to chuansheng.liu from comment #17)
> BTW, Jay, could you help to try the below debugging patch to see if it is
> the issue about d3_delay?
> 
> Thanks.
> 
> diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
> index 4d1a5d2..fa02003 100644
> --- a/drivers/ata/pata_jmicron.c
> +++ b/drivers/ata/pata_jmicron.c
> @@ -142,6 +142,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const
> struct pci_device_id *i
>                 .port_ops       = &jmicron_ops,
>         };
>         const struct ata_port_info *ppi[] = { &info, NULL };
> +       pdev->d3_delay = 200;
>  
>         return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
>  }


Hi Liu!

Patch works! Thank you!

Details later.

Jay

Comment 19 chuansheng.liu 2014-08-22 15:20:43 UTC

Jay,

Could you help to do the below two experiment again?
1/ Disabling the pm_async to see if issue exists;
#cat /sys/power/pm_async
#echo 0 > /sys/power/pm_async

2/ Changing the d3_delay to 20 to see if it is enough;
+       pdev->d3_delay = 20;

BTW, please share your lspci -v -v -v result if possible.

Thanks.

Comment 20 Rainer Fiebig 2014-08-23 14:41:11 UTC

(In reply to chuansheng.liu from comment #19)
> Jay,
> 
> Could you help to do the below two experiment again?
> 1/ Disabling the pm_async to see if issue exists;
> #cat /sys/power/pm_async
> #echo 0 > /sys/power/pm_async
> 
> 2/ Changing the d3_delay to 20 to see if it is enough;
> +       pdev->d3_delay = 20;
> 
> BTW, please share your lspci -v -v -v result if possible.
> 
> Thanks.



Hi!

After some more testing...


ad 1)

Tested with 3.15.9

#echo 0 > /sys/power/pm_async

Works, problem gone. But only if executed BEFORE the first suspend/resume.

[  350.787603] ACPI: Waking up from system sleep state S3
[  350.820498] ehci-pci 0000:00:1a.0: System wakeup disabled by ACPI
[  350.842488] ehci-pci 0000:00:1d.0: System wakeup disabled by ACPI
[  350.936685] PM: noirq resume of devices complete after 148.973 msecs
[  350.936824] PM: early resume of devices complete after 0.118 msecs
[  351.406335] i915: No ACPI video bus found
[  351.407349] mei_me 0000:00:16.0: irq 40 for MSI/MSI-X


ad 2)

Tested with 3.16.0

20/40: bad
50/60: good - but message: "pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3"
70/80/90/100: good

good = After suspend/resume and inserting a CD the drive is automatically recognized and the CD can be accessed.
bad = The drive is not automatically recognized and the CD cannot be accessed.

For 70:
[  130.397797] ACPI: Waking up from system sleep state S3
[  130.408881] ehci-pci 0000:00:1a.0: System wakeup disabled by ACPI
[  130.408906] ehci-pci 0000:00:1d.0: System wakeup disabled by ACPI
[  130.468719] PM: noirq resume of devices complete after 70.795 msecs
[  130.468894] PM: early resume of devices complete after 0.154 msecs
[  130.468978] mei_me 0000:00:16.0: irq 40 for MSI/MSI-X

For 100:
[  196.684920] ACPI: Waking up from system sleep state S3
[  196.695931] ehci-pci 0000:00:1a.0: System wakeup disabled by ACPI
[  196.696028] ehci-pci 0000:00:1d.0: System wakeup disabled by ACPI
[  196.785888] PM: noirq resume of devices complete after 100.847 msecs
[  196.786106] PM: early resume of devices complete after 0.197 msecs
[  196.786202] mei_me 0000:00:16.0: irq 40 for MSI/MSI-X
-------------------------------------------

lspci -v -v -v for the JMicron- and the Symbios-adapter.

02:00.1 IDE interface: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 03) (prog-if 85 [Master SecO PriO])
        Subsystem: Micro-Star International Co., Ltd. Device 7637
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin B routed to IRQ 17
        Region 0: I/O ports at cc00 [size=8]
        Region 1: I/O ports at c880 [size=4]
        Region 2: I/O ports at c800 [size=8]
        Region 3: I/O ports at c480 [size=4]
        Region 4: I/O ports at c400 [size=16]
        Capabilities: [68] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Kernel driver in use: pata_jmicron

04:01.0 SCSI storage controller: LSI Logic / Symbios Logic 53c810 (rev 12)
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (2000ns min, 16000ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 17
        Region 0: I/O ports at e000 [size=256]
        Region 1: Memory at fbfff800 (32-bit, non-prefetchable) [size=256]
        Kernel driver in use: sym53c8xx        
-------------------------------------------
        
Have a good day!

Jay

p.s.: At the end of line #6 ("struct...") of your patch something is missing. And you owe me an aspirin now ;)

Comment 21 chuansheng.liu 2014-08-25 01:08:54 UTC

Thanks your details, Jay!

Here 70ms is one good point, but we can not explain why disabling async can let the issue gone;

Could you help again?
1/ lspci -v -t

2/ Try again with delay=60,70ms and without async with below debugging patch?
We need to understand if this device is depended on other devices powering on,
but the depended devices are not its parent, then ....

diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
index b67d9ae..fe52629 100644
--- a/drivers/base/power/main.c
+++ b/drivers/base/power/main.c
@@ -479,6 +479,8 @@ static int device_resume_noirq(struct device *dev, pm_message_t state, bool asyn
        TRACE_DEVICE(dev);
        TRACE_RESUME(0);
 
+       dev_info(&dev->dev, "async = %d\n", async);
+
        if (dev->power.syscore || dev->power.direct_complete)
                goto Out;

Comment 22 Rainer Fiebig 2014-08-25 06:56:02 UTC

(In reply to chuansheng.liu from comment #21)
> Thanks your details, Jay!
> 
> Here 70ms is one good point, but we can not explain why disabling async can
> let the issue gone;
> 
> Could you help again?
> 1/ lspci -v -t
> 
> 2/ Try again with delay=60,70ms and without async with below debugging patch?
> We need to understand if this device is depended on other devices powering
> on,
> but the depended devices are not its parent, then ....
> 
> diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
> index b67d9ae..fe52629 100644
> --- a/drivers/base/power/main.c
> +++ b/drivers/base/power/main.c
> @@ -479,6 +479,8 @@ static int device_resume_noirq(struct device *dev,
> pm_message_t state, bool asyn
>         TRACE_DEVICE(dev);
>         TRACE_RESUME(0);
>  
> +       dev_info(&dev->dev, "async = %d\n", async);
> +
>         if (dev->power.syscore || dev->power.direct_complete)
>                 goto Out;


OK

ad 1) lspci -v -t

-+-[0000:ff]-+-00.0  Intel Corporation Core Processor QuickPath Architecture Generic Non-core Registers
 |           +-00.1  Intel Corporation Core Processor QuickPath Architecture System Address Decoder
 |           +-02.0  Intel Corporation Core Processor QPI Link 0
 |           +-02.1  Intel Corporation Core Processor QPI Physical 0
 |           +-02.2  Intel Corporation Core Processor Reserved
 |           \-02.3  Intel Corporation Core Processor Reserved
 \-[0000:00]-+-00.0  Intel Corporation Core Processor DRAM Controller
             +-02.0  Intel Corporation Core Processor Integrated Graphics Controller
             +-16.0  Intel Corporation 5 Series/3400 Series Chipset HECI Controller
             +-1a.0  Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller
             +-1b.0  Intel Corporation 5 Series/3400 Series Chipset High Definition Audio
             +-1c.0-[01]--
             +-1c.4-[02]--+-00.0  JMicron Technology Corp. JMB363 SATA/IDE Controller
             |            \-00.1  JMicron Technology Corp. JMB363 SATA/IDE Controller
             +-1c.5-[03]----00.0  Realtek Semiconductor Co., Ltd. RTL8111/8168 PCI Express Gigabit Ethernet controller
             +-1d.0  Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller
             +-1e.0-[04]--+-00.0  AVM GmbH A1 ISDN [Fritz]
             |            \-01.0  LSI Logic / Symbios Logic 53c810
             +-1f.0  Intel Corporation 5 Series Chipset LPC Interface Controller
             +-1f.2  Intel Corporation 5 Series/3400 Series Chipset 6 port SATA AHCI Controller
             \-1f.3  Intel Corporation 5 Series/3400 Series Chipset SMBus Controller


ad 2) Tomorrow (you time).

Jay

Comment 23 chuansheng.liu 2014-08-25 07:11:47 UTC

(In reply to chuansheng.liu from comment #21)

> 2/ Try again with delay=60,70ms and without async with below debugging patch?

>  
> +       dev_info(&dev->dev, "async = %d\n", async);
> +
>         if (dev->power.syscore || dev->power.direct_complete)
>                 goto Out;

Thanks Jay.

It should be:
diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
index b67d9ae..0408ecd 100644
--- a/drivers/base/power/main.c
+++ b/drivers/base/power/main.c
@@ -479,6 +479,8 @@ static int device_resume_noirq(struct device *dev, pm_message_t state, bool asyn
        TRACE_DEVICE(dev);
        TRACE_RESUME(0);
 
+       dev_info(dev, "async = %d\n", async);
+
        if (dev->power.syscore || dev->power.direct_complete)
                goto Out;

Comment 24 chuansheng.liu 2014-08-25 07:24:17 UTC

Updated the debugging info patch again:

diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
index b67d9ae..8c8e6dd 100644
--- a/drivers/base/power/main.c
+++ b/drivers/base/power/main.c
@@ -479,6 +479,8 @@ static int device_resume_noirq(struct device *dev, pm_message_t state, bool asyn
        TRACE_DEVICE(dev);
        TRACE_RESUME(0);
 
+       dev_info(dev, "async = %d, start\n", async);
+
        if (dev->power.syscore || dev->power.direct_complete)
                goto Out;
 
@@ -512,6 +514,7 @@ static int device_resume_noirq(struct device *dev, pm_message_t state, bool asyn
  Out:
        complete_all(&dev->power.completion);
        TRACE_RESUME(error);
+       dev_info(dev, "error = %d, end\n", error);
        return error;
 }

Comment 25 Rainer Fiebig 2014-08-25 20:29:42 UTC

Created attachment 148161 [details]
2014-08-25_dmesg_patch_2_60ms

Comment 26 Rainer Fiebig 2014-08-25 20:30:22 UTC

Created attachment 148171 [details]
2014-08-25_dmesg_patch_2_70ms

Comment 27 Rainer Fiebig 2014-08-25 20:31:15 UTC

Created attachment 148181 [details]
2014-08-25_dmesg_patch_2_20ms_pm_async_0

Comment 28 Rainer Fiebig 2014-08-25 20:33:14 UTC

Hi!

Please find attached the dmesg-output for 

pdev->d3_delay = 60;
pdev->d3_delay = 70;
pdev->d3_delay = 20 with pm_async = 0;

I couldn't find anything with "Aha!"-potential for me. But perhaps with the eyes of the expert... Good luck!

The functional outcome was the same as already reported.


Jay.

Comment 29 chuansheng.liu 2014-08-25 23:09:19 UTC

Hi Jay,

My fault, the log you caught are good.
Could you catch one BAD case? Such as pdev->d3_delay = 20 with pm_async = 1;

Thanks.

Comment 30 chuansheng.liu 2014-08-25 23:36:39 UTC

BTW, in the 60ms and 70ms log, I didn't find the start for device pata_jmicron:

[  222.158394] pata_jmicron 0000:02:00.1: error = 0, end

Just ending is there.

Comment 31 chuansheng.liu 2014-08-25 23:52:55 UTC

Ah.

Jay,

From the 60ms and 70ms log, it seems device "pata_jmicron 0000:02:00.1" depends on 
the device "ahci 0000:02:00.0", but pata_jmicron is async mode, and ahci is non-async mode, need your complete log to confirm.

But from the device tree, we didn't find relationship for both devices.

             +-1c.4-[02]--+-00.0  JMicron Technology Corp. JMB363 SATA/IDE Controller
             |            \-00.1  JMicron Technology Corp. JMB363 SATA/IDE Controller

Comment 32 chuansheng.liu 2014-08-26 00:32:22 UTC

(In reply to chuansheng.liu from comment #31)
> Ah.
> 
> Jay,
> 
> From the 60ms and 70ms log, it seems device "pata_jmicron 0000:02:00.1"
> depends on 
> the device "ahci 0000:02:00.0", but pata_jmicron is async mode, and ahci is
> non-async mode, need your complete log to confirm.
> 
> But from the device tree, we didn't find relationship for both devices.
> 
>              +-1c.4-[02]--+-00.0  JMicron Technology Corp. JMB363 SATA/IDE
> Controller
>              |            \-00.1  JMicron Technology Corp. JMB363 SATA/IDE
> Controller

Jay,
So could you try the below patch? Thanks. Seems it is just JMB363 related issue,
that two controllers have the dependency.
Here let's disable the async method for it.

diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 4d1a5d2..52f1353 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -142,7 +142,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
                .port_ops       = &jmicron_ops,
        };
        const struct ata_port_info *ppi[] = { &info, NULL };
-
+       device_disable_async_suspend(&pdev->dev);
        return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
 }

Comment 33 chuansheng.liu 2014-08-26 00:39:34 UTC

Or like this:
diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 4d1a5d2..c98c7c7 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -132,6 +132,7 @@ static struct ata_port_operations jmicron_ops = {
 
 static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *id)
 {
+       int error;
        static const struct ata_port_info info = {
                .flags  = ATA_FLAG_SLAVE_POSS,
 
@@ -143,7 +144,11 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
        };
        const struct ata_port_info *ppi[] = { &info, NULL };
 
-       return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
+       error = ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
+       if (!error)
+               device_disable_async_suspend(&pdev->dev);
+
+       return error;
 }

Comment 34 Rainer Fiebig 2014-08-26 17:38:49 UTC

(In reply to chuansheng.liu from comment #30)
> BTW, in the 60ms and 70ms log, I didn't find the start for device
> pata_jmicron:
> 
> [  222.158394] pata_jmicron 0000:02:00.1: error = 0, end
> 
> Just ending is there.

OK - that was a bit sparse. Sorry. See new attachement.

Jay

Comment 35 Rainer Fiebig 2014-08-26 17:40:47 UTC

Created attachment 148381 [details]
2014-08-25_dmesg_patch_2_70ms_complete

Comment 36 Rainer Fiebig 2014-08-26 17:42:53 UTC

(In reply to chuansheng.liu from comment #29)
> Hi Jay,
> 
> My fault, the log you caught are good.
> Could you catch one BAD case? Such as pdev->d3_delay = 20 with pm_async = 1;
> 
> Thanks.

No problem. See new attachement. But not much new info.

Comment 37 Rainer Fiebig 2014-08-26 17:43:19 UTC

Created attachment 148391 [details]
2014-08-25_dmesg_patch_2_20ms_pm_async_1

Comment 38 Rainer Fiebig 2014-08-26 17:46:34 UTC

(In reply to chuansheng.liu from comment #33)
> Or like this:
> diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
> index 4d1a5d2..c98c7c7 100644
> --- a/drivers/ata/pata_jmicron.c
> +++ b/drivers/ata/pata_jmicron.c
> @@ -132,6 +132,7 @@ static struct ata_port_operations jmicron_ops = {
>  
>  static int jmicron_init_one (struct pci_dev *pdev, const struct
> pci_device_id *id)
>  {
> +       int error;
>         static const struct ata_port_info info = {
>                 .flags  = ATA_FLAG_SLAVE_POSS,
>  
> @@ -143,7 +144,11 @@ static int jmicron_init_one (struct pci_dev *pdev,
> const struct pci_device_id *i
>         };
>         const struct ata_port_info *ppi[] = { &info, NULL };
>  
> -       return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
> +       error = ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
> +       if (!error)
> +               device_disable_async_suspend(&pdev->dev);
> +
> +       return error;
>  }


Hi!

Tried this patch (named it "3b") first, but could not apply it (kernel 3.16.0):

"patch: **** malformed patch at line 7: static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *id)"

So I edited the pata_jmicron.c directly, according to your patch and without "pdev->d3_delay = 70;"

However, after suspend/resume:

[  177.948595] pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3
...
[  179.414978] irq 17: nobody cared (try booting with the "irqpoll" option)
[  179.414981] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G           O  3.16.0-1.g8038aea-desktop #23
[  179.414982] Hardware name: MSI MS-7637/H55-GD65 (MS-7637)  , BIOS V1.5 09/16/2010
[  179.414984]  0000000000000000 ffff880329777c8c ffffffff81599de9 ffff880329777c00
[  179.414985]  ffffffff810a30f6 ffff880329777c00 0000000000000011 0000000000000000
[  179.414987]  ffffffff810a3636 ffffffffffd1170a 0000000000000000 0000000000000000
[  179.414987] Call Trace:
[  179.414996]  <IRQ>  [<ffffffff81599de9>] ? dump_stack+0x49/0x6a
snip
[  179.415039] [<ffffffff8140c3d0>] ata_bmdma_interrupt
[  179.415039] Disabling IRQ #17


Then tried the first patch ("3a"; worked):

Now with this patch and without "pdev->d3_delay = 70;":

[  179.770127] pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3
...
[  179.820960] PM: noirq resume of devices complete after 62.340 msecs
[  179.821175] PM: early resume of devices complete after 0.195 msecs
[  179.821254] mei_me 0000:00:16.0: irq 40 for MSI/MSI-X
[  179.822215] pcieport 0000:00:1c.5: System wakeup disabled by ACPI
[  179.822303] snd_hda_intel 0000:00:1b.0: irq 44 for MSI/MSI-X

And after inserting the CD it wasn't recognized and could not be accessed.

And after a second suspend/resume:

[ 3457.132686] irq 17: nobody cared (try booting with the "irqpoll" option)
[ 3457.132689] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G           O  3.16.0-1.g8038aea-desktop #23
[ 3457.132690] Hardware name: MSI MS-7637/H55-GD65 (MS-7637)  , BIOS V1.5 09/16/2010
[ 3457.132692]  0000000000000000 ffff880329779c8c ffffffff81599de9 ffff880329779c00
[ 3457.132693]  ffffffff810a30f6 ffff880329779c00 0000000000000011 0000000000000000
[ 3457.132694]  ffffffff810a3636 ffffffffffd1170a 0000000000000000 0000000000000000
snip
[ 3457.132748] [<ffffffff8140c3d0>] ata_bmdma_interrupt
[ 3457.132749] Disabling IRQ #17


To me it seems that the JMicron-adapter simply does not react to "device_disable_async_suspend(&pdev->dev);" but only to "pm_async = 0". Apparently the async-method cannot be switched off for this adapter alone.

After going back to "pdev->d3_delay = 70;" everything is fine again.

Jay

Comment 39 chuansheng.liu 2014-08-26 23:49:55 UTC

(In reply to Jay from comment #35)
> Created attachment 148381 [details]
> 2014-08-25_dmesg_patch_2_70ms_complete

Still missed some info here? It is due to your log buffer is limited.
Could you help to enlarge it to try?

CONFIG_LOG_BUF_SHIFT

Comment 40 chuansheng.liu 2014-08-26 23:59:54 UTC

(In reply to Jay from comment #38)
> (In reply to chuansheng.liu from comment #33)
 
> To me it seems that the JMicron-adapter simply does not react to
> "device_disable_async_suspend(&pdev->dev);" but only to "pm_async = 0".
> Apparently the async-method cannot be switched off for this adapter alone.

Thanks Jay.

Could you check the below async files to see something?
#cat /sys/devices/..../0000:02:00.0/power/async
#cat /sys/devices/..../0000:02:00.1/power/async

If you can not find the "async" file, please try to enable CONFIG_PM_ADVANCED_DEBUG at first.

And if these both files are enabled, please try to disable them to see if the issue still existed.

Comment 41 Rainer Fiebig 2014-08-27 09:02:53 UTC

Hi!

Thanks for your messages.

I'm afraid the kernel-log-buffer is already set to its max: CONFIG_LOG_BUF_SHIFT = 17 = 128kb.

Even after enabling PM_ADVANCED_DEBUG I could not find 
#cat /sys/devices/..../0000:02:00.0/power/async
#cat /sys/devices/..../0000:02:00.1/power/async

What I found is this:
cat /sys/devices/breakpoint/power/async
disabled
cat /sys/devices/cpu/power/async
disabled
cat /sys/devices/software/power/async
disabled
cat /sys/devices/tracepoint/power/async
disabled
cat /sys/devices/uncore/power/async
disabled

Here is an excerpt from the demsg -k after the first suspend/resume after a fresh system-start (see new attachment):
Is it OK that the CD-ROM-drive (ata9) is starting in async-mode? Or is this unrelated to device_disable_async_suspend?

...
[  879.261744]  ata9: async = 1, start
[  879.261758]  ata10: async = 1, start
[  879.261772] scsi host8: async = 1, start
[  879.261786] scsi_host host8: async = 1, start
[  879.261800] scsi host9: async = 1, start
[  879.261813] scsi_host host9: async = 1, start
[  879.261827] scsi target8:0:0: async = 1, start
[  879.261839] sr 8:0:0:0: async = 1, start
[  879.261852] scsi_device 8:0:0:0: async = 1, start
[  879.271446] pata_jmicron 0000:02:00.1: Refused to change power state, currently in D3
...
[  880.509826] Disabling IRQ #17
...

Jay

Comment 42 Rainer Fiebig 2014-08-27 09:04:38 UTC

Created attachment 148431 [details]
2014-08-27_dmesg_patch_3a

Comment 43 chuansheng.liu 2014-08-27 09:17:57 UTC

(In reply to Jay from comment #41)
> Hi!
> 
> Thanks for your messages.
> 
> I'm afraid the kernel-log-buffer is already set to its max:
> CONFIG_LOG_BUF_SHIFT = 17 = 128kb.
Could it be 18 or 19?

> 
> Even after enabling PM_ADVANCED_DEBUG I could not find 
Hmm. Are you sure it is enabled?
You should enable PM_DEBUG also since PM_ADVANCED_DEBUG depends on it.

> #cat /sys/devices/..../0000:02:00.0/power/async
> #cat /sys/devices/..../0000:02:00.1/power/async
>

Comment 44 Rainer Fiebig 2014-08-27 10:35:46 UTC

(In reply to chuansheng.liu from comment #43)
> (In reply to Jay from comment #41)
> > Hi!
> > 
> > Thanks for your messages.
> > 
> > I'm afraid the kernel-log-buffer is already set to its max:
> > CONFIG_LOG_BUF_SHIFT = 17 = 128kb.
> Could it be 18 or 19?


Now set to 19, see new attachement.


> > 
> > Even after enabling PM_ADVANCED_DEBUG I could not find 
> Hmm. Are you sure it is enabled?
> You should enable PM_DEBUG also since PM_ADVANCED_DEBUG depends on it.

Both were enabled.

> 
> > #cat /sys/devices/..../0000:02:00.0/power/async
> > #cat /sys/devices/..../0000:02:00.1/power/async

I cannot find these paths in /sys/devices/ on my system. Where shall I look for the files? Please take a look at the attachment (.jepg).

Jay

Comment 45 Rainer Fiebig 2014-08-27 10:36:41 UTC

Created attachment 148441 [details]
2014-08-27_dmesg_patch_3a_2

Comment 46 Rainer Fiebig 2014-08-27 10:38:02 UTC

Created attachment 148451 [details]
Bug 81551_contents_sys_devices

Comment 47 chuansheng.liu 2014-08-27 12:36:12 UTC

(In reply to Jay from comment #44)
> (In reply to chuansheng.liu from comment #43)
> Now set to 19, see new attachement.
Now we can see the device:
[  116.654313] pata_jmicron 0000:02:00.1: async = 0, start

But still missed some logs, could you set BUF to 21?

> > > #cat /sys/devices/..../0000:02:00.0/power/async
> > > #cat /sys/devices/..../0000:02:00.1/power/async
> 
> I cannot find these paths in /sys/devices/ on my system. Where shall I look
> for the files? Please take a look at the attachment (.jepg).

The file should in the directory /sys/device/pci0000:00/0000:00:1c.4/, there should be two async files in deeper sub-directory.

Comment 48 Rainer Fiebig 2014-08-27 17:58:30 UTC

Thanks for your assistance.

One step closer to a solution now, I think.

> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
enabled
> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.1/power/async
disabled

# echo disabled > /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
disabled

Success. After suspend/resume the CD was automatically recognized and accessible. Same after a second suspend/resume.

No warnings in the dmesg-output (see attachment). The noirq-resume-time of 72.394 ms is interesting as it is approximately the d3_delay-value where the "Refused to change power state,..."-warning disappeared.

...
[  322.163865] PM: noirq resume of devices complete after 72.394 msecs
[  322.164148] PM: early resume of devices complete after 0.262 msecs
...

See below for a lspci-v-v-v - synopsis of the JMicron-controller.

Jay




02:00.0 SATA controller: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 03) (prog-if 01 [AHCI 1.0])
        Subsystem: Micro-Star International Co., Ltd. Device 7637
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 16
        Region 5: Memory at fbdfe000 (32-bit, non-prefetchable) [size=8K]
        Capabilities: [68] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot+,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [50] Express (v1) Legacy Endpoint, MSI 01
                DevCap: MaxPayload 128 bytes, PhantFunc 0, Latency L0s <64ns, L1 <1us
                        ExtTag- AttnBtn- AttnInd- PwrInd- RBE- FLReset-
                DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
                        RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop-
                        MaxPayload 128 bytes, MaxReadReq 512 bytes
                DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr- TransPend-
                LnkCap: Port #1, Speed 2.5GT/s, Width x1, ASPM L0s, Latency L0 unlimited, L1 unlimited
                        ClockPM- Surprise- LLActRep- BwNot-
                LnkCtl: ASPM Disabled; RCB 64 bytes Disabled- Retrain- CommClk-
                        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
                LnkSta: Speed 2.5GT/s, Width x1, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
        Kernel driver in use: ahci

02:00.1 IDE interface: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 03) (prog-if 85 [Master SecO PriO])
        Subsystem: Micro-Star International Co., Ltd. Device 7637
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin B routed to IRQ 17
        Region 0: I/O ports at cc00 [size=8]
        Region 1: I/O ports at c880 [size=4]
        Region 2: I/O ports at c800 [size=8]
        Region 3: I/O ports at c480 [size=4]
        Region 4: I/O ports at c400 [size=16]
        Capabilities: [68] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Kernel driver in use: pata_jmicron

Comment 49 Rainer Fiebig 2014-08-27 17:59:32 UTC

Created attachment 148521 [details]
2014-08-27_dmesg_patch_3a_both_async_disabled

Comment 50 chuansheng.liu 2014-08-29 01:14:44 UTC

How about the test result with below command?

#echo disabled > /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
#echo enabled > /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.1/power/async

Thanks.

Comment 51 Rainer Fiebig 2014-08-29 07:35:27 UTC

(In reply to chuansheng.liu from comment #50)
> How about the test result with below command?
> 
> #echo disabled >
> /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
> #echo enabled > /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.1/power/async
> 
> Thanks.

Hi!

> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
disabled
> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.1/power/async
enabled

Doesn't work: "[  617.147932] Disabling IRQ #17"

Jay

Comment 52 chuansheng.liu 2014-08-29 14:43:48 UTC

Thanks Jay.

That proved that the two controller 02:00.0 and 02:00.1 have the powering on dependency, although they don't have child-parent relationship.

Could you try this patch?
Before you do the suspend test, you can check if the two devices async have been
disabled as the patch expected.

Thanks.

Comment 53 chuansheng.liu 2014-08-29 14:44:56 UTC

Created attachment 148751 [details]
Disabling the async for 363/361 SATA/PATA controllers

Comment 54 Rainer Fiebig 2014-08-30 22:26:31 UTC

Hi!

Received the e-mails somewhat delayed, probably due to problems at my provider.

Well - good news and bad news.

The good news first: Applied your latest patch to kernels 3.16.0 and 3.15.9 and it worked for both. Tested it thoroughly, I think.

As I see it, Bug 81551 is solved. 

Thank you!


Now for the bad news.

While testing, another problem surfaced that at first I thought was perhaps connected to the patch - but it is not: 

After two successful suspends (and checking the CD-ROM-drive) the third s2disk-process (save2both) was aborted after only a part of the image had been written to disk. The system then returned to the desktop. This happened although only a small amount of RAM was used (ca. 7%) and although there is ample space in the swap-partition. 

The CD-ROM-drive was always unaffected by this, it still worked after the aborted suspend.

The reason why I'm sure that this is a different issue is because I not only saw it with 3.16 and 3.15.9, but later also with 3.14.17. Where I had not seen it before. But until now I surely never made 3 suspends in a row like in the tests.

It was reproducible and except for one instance always happened after the third suspend. 

Any ideas? A new bug-report?

Kind regards.

Jay


p.s.: After applying the patch in 3.16 and before the first suspend/resume:

> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.0/power/async
disabled
> cat /sys/devices/pci0000:00/0000:00:1c.4/0000:02:00.1/power/async
disabled

Comment 55 chuansheng.liu 2014-08-30 23:57:50 UTC

(In reply to Jay from comment #54)

> Well - good news and bad news.
> 
> The good news first: Applied your latest patch to kernels 3.16.0 and 3.15.9
> and it worked for both. Tested it thoroughly, I think.
> 
> As I see it, Bug 81551 is solved. 
> 
> Thank you!
Thanks your help indeed.

> 
> Now for the bad news.
> 
> While testing, another problem surfaced that at first I thought was perhaps
> connected to the patch - but it is not: 
> 
> The reason why I'm sure that this is a different issue is because I not only
> saw it with 3.16 and 3.15.9, but later also with 3.14.17. Where I had not
> seen it before. But until now I surely never made 3 suspends in a row like
> in the tests.
So even with the original 3.14.17, the issue still existed? not connected to
the patch? Thanks.

> Any ideas? A new bug-report?
If it happened also at original 3.14.17, then it should be one new BUG.
Please help to upload some logs with one new BUG, and the async series patchset just existed from 3.15.

It is better to do some bisection for it also.

Comment 56 Rafael J. Wysocki 2014-09-01 12:32:17 UTC

(In reply to Jay from comment #54)

[cut]

> Any ideas? A new bug-report?

Yes, a new report, please.

Comment 57 Barto 2014-09-22 14:32:25 UTC

I still have the bug despite the patch :

https://bugzilla.kernel.org/attachment.cgi?id=148751

because probably the patch seems incomplete, I have a Jmicron JMB363/JMB368 SATA/IDE controler, and an IDE hardisk is attached on JMB368 ( IDE controler ),

you should add a new condition in your "if statement" in order to target also "PCI_DEVICE_ID_JMICRON_JMB368" and not only JMB363/JMB368 :

+	if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
+		(pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
+		pdev->device == PCI_DEVICE_ID_JMICRON_JMB361 ||
+               pdev->device == PCI_DEVICE_ID_JMICRON_JMB368 ))
+		device_disable_async_suspend(&pdev->dev);

what do you think ?

Comment 58 Barto 2014-09-22 14:40:30 UTC

lspci -v :

03:00.0 SATA controller: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 10) (prog-if 01 [AHCI 1.0])
        Subsystem: JMicron Technology Corp. JMB363 SATA/IDE Controller
        Flags: bus master, fast devsel, latency 0, IRQ 17
        Memory at e7000000 (32-bit, non-prefetchable) [size=512]
        [virtual] Expansion ROM at e6000000 [disabled] [size=64K]
        Capabilities: [8c] Power Management version 3
        Capabilities: [50] Express Legacy Endpoint, MSI 00
        Kernel driver in use: ahci
        Kernel modules: ahci

03:00.1 IDE interface: JMicron Technology Corp. JMB368 IDE controller (rev 10) (prog-if 85 [Master SecO PriO])
        Subsystem: JMicron Technology Corp. Device 1368
        Flags: bus master, fast devsel, latency 0, IRQ 18
        I/O ports at b000 [size=8]
        I/O ports at b100 [size=4]
        I/O ports at b200 [size=8]
        I/O ports at b300 [size=4]
        I/O ports at b400 [size=16]
        Capabilities: [8c] Power Management version 3
        Capabilities: [50] Express Legacy Endpoint, MSI 00
        Kernel driver in use: pata_jmicron
        Kernel modules: pata_jmicron, pata_acpi, ata_generic

we need to be sure that the patch will also target "JMB368 IDE controller" and not only JMB361/JMB363

Comment 59 Barto 2014-09-22 16:40:12 UTC

well I tried to add "pdev->device == PCI_DEVICE_ID_JMICRON_JMB368" in the "if statement" in the patch :

+	if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
+		(pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
+		pdev->device == PCI_DEVICE_ID_JMICRON_JMB361 ||
+               pdev->device == PCI_DEVICE_ID_JMICRON_JMB368 ))
+		device_disable_async_suspend(&pdev->dev);


but I still get the bug, my JMB363/368 SATA/IDE controler refuses to work after a standby :

[   96.893714] pata_jmicron 0000:03:00.1: Refused to change power state, currently in D3

the only solution is to do a "echo 0 > /sys/power/pm_async" before doing a standby

Comment 60 Barto 2014-09-22 16:41:48 UTC

Created attachment 151321 [details]
dmesg output after a standby

here is the dmesg output after a standby,

we can see this line :

[   96.893714] pata_jmicron 0000:03:00.1: Refused to change power state, currently in D3

and my IDE harddisk connected to Jmicron JMB363/368 doesn't work after a standby,

the kernel used was 3.16.3 with this patch :

https://bugzilla.kernel.org/attachment.cgi?id=148751

Comment 61 Barto 2014-09-22 18:30:18 UTC

Created attachment 151391 [details]
dmesg output with kernel 3.17rc06 after standby, bug is still here

I tried with kernel 3.17rc6 and it's the same problem : "SATA/IDE controler Jmicron JMB363/368" doesn't work after a standby,

here is the dmesg output after the standby,

I still have these messages error :

[  137.363555] pata_jmicron 0000:03:00.1: Refused to change power state, currently in D3

and 

[  304.379996] ata1: lost interrupt (Status 0x58)
[  304.395183] ata1: drained 65536 bytes to clear DRQ
[  304.507727] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[  304.507734] ata1.01: failed command: READ DMA
[  304.507738] ata1.01: cmd c8/00:08:97:a9:40/00:00:00:00:00/f2 tag 0 dma 4096 in
         res 40/00:00:00:4f:c2/00:00:00:00:00/10 Emask 0x20 (host bus error)
[  304.507739] ata1.01: status: { DRDY }

Comment 62 chuansheng.liu 2014-09-23 01:53:11 UTC

Hello Barto,

Please try this patch based on the latest code base, thanks.

diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
index a0cc0ed..ae1cd4d 100644
--- a/drivers/ata/ahci.c
+++ b/drivers/ata/ahci.c
@@ -1346,9 +1346,7 @@ static int ahci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        /* acquire resources */
diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 47e418b..2150d3d 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -150,9 +150,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);

Comment 63 Barto 2014-09-23 03:10:02 UTC

Hello Chuansheng,

I have an error message if I want to apply your patch :

patching file drivers/ata/ahci.c
Hunk #1 FAILED at 1346.
1 out of 1 hunk FAILED -- saving rejects to file drivers/ata/ahci.c.rej
patching file drivers/ata/pata_jmicron.c
Hunk #1 FAILED at 150.
1 out of 1 hunk FAILED -- saving rejects to file drivers/ata/pata_jmicron.c.rej

I tried with kernel 3.17-rc6,

can you provide a compatible patch for kernel 3.17-rc6 ?

Comment 64 chuansheng.liu 2014-09-23 03:16:47 UTC

This patch is based on 3.17-rc6, which is the same as before.

diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
index a0cc0ed..ae1cd4d 100644
--- a/drivers/ata/ahci.c
+++ b/drivers/ata/ahci.c
@@ -1346,9 +1346,7 @@ static int ahci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        /* acquire resources */
diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 47e418b..2150d3d 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -150,9 +150,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);

Comment 65 chuansheng.liu 2014-09-23 03:17:48 UTC

My code base HEAD is:
commit 0f33be009b89d2268e94194dc4fd01a7851b6d51
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Sun Sep 21 15:43:02 2014 -0700

    Linux 3.17-rc6

Comment 66 Barto 2014-09-23 04:46:17 UTC

Hello Chuansheng,

I tested your new patch and it's ok now,

the bug is gone, my Jmicron JMB363/368 can work after a standby,

if I understand : your patch tries to disable "async_suspend feature" for all devices who have the vendor id "PCI_VENDOR_ID_JMICRON" ?

so for all Jmicron controlers models this feature will be disabled with this patch ?

Comment 67 chuansheng.liu 2014-09-23 05:13:01 UTC

Hello Barto,

Thanks your confirm, then it is the same root cause that non parent-children relationship's Jmicron controllers have power dependency.

Here we can directly exclude all Jmicron controller for pm_async feature.
I will submit one patch to upstream to improve it.

Best Regards
Chuansheng

Comment 68 Len Brown 2014-09-23 14:30:07 UTC

*** Bug 84861 has been marked as a duplicate of this bug. ***

Comment 69 Barto 2014-11-02 18:16:12 UTC

the bug is still here with kernel 3.17.2, something seems to have canceled the effect of the patch of chuansheng,

see my bugreport :

https://bugzilla.kernel.org/show_bug.cgi?id=84861

Comment 70 Barto 2014-11-02 18:42:58 UTC

I found the explication :

in fact Chuansheng's patch has not been applied in official version kernel 3.17.1 and 3.17.2 source code,

that's why the bug is still here,

 it would be great if someone can push his patch for the next  version ( kernel 3.17.3 ),

it's not normal to leave this bug unfixed because the importance of this bug is "high"

Comment 71 Barto 2014-11-02 18:45:57 UTC

this is the patch that someone has forgot to push in git :

diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
index a0cc0ed..ae1cd4d 100644
--- a/drivers/ata/ahci.c
+++ b/drivers/ata/ahci.c
@@ -1346,9 +1346,7 @@ static int ahci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        /* acquire resources */
diff --git a/drivers/ata/pata_jmicron.c b/drivers/ata/pata_jmicron.c
index 47e418b..2150d3d 100644
--- a/drivers/ata/pata_jmicron.c
+++ b/drivers/ata/pata_jmicron.c
@@ -150,9 +150,7 @@ static int jmicron_init_one (struct pci_dev *pdev, const struct pci_device_id *i
         * powered on successfully, so here we disable the async suspend
         * method for these chips.
         */
-       if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
-               (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
-               pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+       if (pdev->vendor == PCI_VENDOR_ID_JMICRON)
                device_disable_async_suspend(&pdev->dev);
 
        return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);

Comment 72 Barto 2014-11-02 18:50:02 UTC

Created attachment 156301 [details]
the final patch for JMB JMicron family IDE/SATA controler

the final patch for JMB JMicron family IDE/SATA controler ( the author is Chuanseng )

the goal of this patch is to disable async_suspend for all JMicron sata/ide controlers, it's the only solution because there is too much models of JMicron who have problems with async_suspend feature

Comment 73 Len Brown 2014-11-03 21:01:55 UTC

*** Bug 84861 has been marked as a duplicate of this bug. ***

Comment 74 Len Brown 2014-11-03 21:04:22 UTC

On the list, there is an outstanding question about that patch,
regarding whether it should be done in PCI init instead.
Until the question is answered, it can't be called "final".

Comment 75 Barto 2014-11-03 21:49:13 UTC

I hope that this bug will not be forgotten by the developpers and a "final/better" solution will be found, like an updated IDE/SATA driver for JMicron JMBxxx,

or a complete rewrite of the async_suspend feature, for example a better mechanism with an automatic deactivation if the PCI device is not compatible with async_suspend feature, a kind of "try/catch" feature, some lines for testing the PCI device about the async_suspend, without the need of manual fixing in the source code, it should be automatic, a dynamic process by the code himself

Comment 76 Rainer Fiebig 2014-11-04 11:17:11 UTC

(In reply to Len Brown from comment #74)
> On the list, there is an outstanding question about that patch,
> regarding whether it should be done in PCI init instead.
> Until the question is answered, it can't be called "final".


Time is a factor here. More users will run into this when upgrading to kernel 3.15 or later.

openSuse 13.2 is out today with kernel 3.16, for instance. The latest Ubuntu (14.10) also uses 3.16.

Therefore, a more pragmatic approach would be to use the available patch and then take time to find a better solution.

Or - as Patton once said: "A good plan today is better than a perfect plan tomorrow."

Jay

Comment 77 Barto 2014-11-21 16:00:46 UTC

Created attachment 158351 [details]
disable_async_suspend_jmicron by using /drivers/pci/quirks.c file

Chuanseng's patch who solves the bug by using the /drivers/pci/quirks.c file,

it's time to push his patch to kernel 3.18rc6

Comment 78 Len Brown 2014-11-25 01:35:26 UTC

Chuanseng,
please put patchwork URL here, and set this to RESOLVED.
when it ships upstream, then we will mark it CLOSED.

Comment 79 Zhang Rui 2015-01-26 03:05:47 UTC

Chuansheng, what is the status of the patch?

Comment 80 chuansheng.liu 2015-01-26 03:11:06 UTC

Rui,

According to experts' comments, I have implemented two version patches:
1/ WA in the Jmicron driver code;
2/ Add one quirk for the Jmicron;

But now new request is coming here from Alan:

==
> You know, this is exactly why device_pm_wait_for_dev() exists -- so 
> that asynchronous power-management operations can wait for other 
> devices even when there's no parent-child relation.
> 
> You should try to use device_pm_wait_for_dev() instead of disabling 
> async suspend/resume.
==

But I think it makes things to be difficult, need understanding here.

Comment 81 Zhang Rui 2015-01-26 03:42:13 UTC

yes, aync resume not work means we have an implicit dependency that is not met with pm_async is enabled. And Alan just want you to describe this dependency explicitly in the driver, just like Alan has done with ehci_wait_for_companions().

Comment 82 Zhang Rui 2015-02-09 05:30:02 UTC

Chuansheng, is there any update about the new patch?

Comment 83 Zhang Rui 2015-03-14 13:39:37 UTC

Ping chuansheng...

Comment 84 Barto 2015-03-14 15:26:02 UTC

(In reply to Zhang Rui from comment #83)
> Ping chuansheng...

maybe you should also contact Alan if Chuansheng can't be contacted

Comment 85 Len Brown 2015-03-24 01:23:56 UTC

Rui will try to copy what ehci_wait_for_completions() does
into the ATAPI code and come up with a test patch.

Ideally, we want to *keep* async resume for its performance,
and simply correctly manage the dependencies.

Comment 86 Alan Stern 2015-03-24 17:02:57 UTC

The routine is ehci_wait_for_companions() (in drivers/usb/core/hcd-pci.c), not ehci_wait_for_completions().

Here's a simple explanation of how you use it.  Let's say you've got two devices A and B, where A must be powered on before B.  Then at the start of B's resume_noirq callback routine, add a line that does:

    device_pm_wait_for_dev(&B->dev, &A->dev);

(where A and B are the variables pointing to the pci_dev structures).  That's all you have to do.

Comment 87 Zhang Rui 2015-03-30 07:30:51 UTC

Created attachment 172621 [details]
patch fix jmicron controller dependency

please check if this patch helps or not.

Comment 88 Alan Stern 2015-03-30 14:13:31 UTC

Note that if there is more than one JMicron controller device in the same PCI slot, your patch will make all of them wait for each other!  Do you think the patch should also test the function number (only wait for devices whose function is < the current device's function)?

Even better would be if you could tell which device was the AHCI controller and only wait for it.

Comment 89 Barto 2015-03-30 16:33:05 UTC

(In reply to Zhang Rui from comment #87)
> Created attachment 172621 [details]
> patch fix jmicron controller dependency
> 
> please check if this patch helps or not.

I tested your patch and it doesn't work,

I think it's because my Jmicron PCIe card has both a SATA port and an IDE port, 

Alan Stern has explained why your patch fails, don't forget that JMicron PCIe has often both Sata ports and one IDE port ( hybrid card )

Comment 90 Zhang Rui 2015-04-07 05:34:16 UTC

please attach the full output of lspci -vvx?

Comment 91 Zhang Rui 2015-04-07 14:00:58 UTC

(In reply to Barto from comment #89)
> (In reply to Zhang Rui from comment #87)
> > Created attachment 172621 [details]
> > patch fix jmicron controller dependency
> > 
> > please check if this patch helps or not.
> 
> I tested your patch and it doesn't work,
> 
> I think it's because my Jmicron PCIe card has both a SATA port and an IDE
> port, 
> 
Yes, I know.

> Alan Stern has explained why your patch fails, don't forget that JMicron
> PCIe has often both Sata ports and one IDE port ( hybrid card )

But, TBH, I still don't understand why the patch fails.
The patch in comment #87 makes the pata_jmicron device (IDE controller) to wait until the AHCI host controller is resumed, do I miss something?

Comment 92 Barto 2015-04-07 18:59:10 UTC

(In reply to Zhang Rui from comment #90)
> please attach the full output of lspci -vvx?

a "lspci --vvx | grep JMicron" :

lspci -vvx | grep JMicron
03:00.0 SATA controller: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 10) (prog-if 01 [AHCI 1.0])
        Subsystem: JMicron Technology Corp. JMB363 SATA/IDE Controller
03:00.1 IDE interface: JMicron Technology Corp. JMB368 IDE controller (rev 10) (prog-if 85 [Master SecO PriO])
        Subsystem: JMicron Technology Corp. Device 1368

and the full output of "lspci -vvx" :

00:00.0 Host bridge: Intel Corporation 82G33/G31/P35/P31 Express DRAM Controller (rev 10)
        Subsystem: Gigabyte Technology Co., Ltd Device 5000
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ >SERR- <PERR- INTx-
        Latency: 0
        Capabilities: <access denied>
        Kernel modules: intel_agp
00: 86 80 c0 29 06 00 90 20 10 00 00 06 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 58 14 00 50
30: 00 00 00 00 e0 00 00 00 00 00 00 00 00 00 00 00

00:01.0 PCI bridge: Intel Corporation 82G33/G31/P35/P31 Express PCI Express Root Port (rev 10) (prog-if 00 [Normal decode])
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 24
        Bus: primary=00, secondary=01, subordinate=01, sec-latency=0
        I/O behind bridge: 0000a000-0000afff
        Memory behind bridge: e4000000-e5ffffff
        Prefetchable memory behind bridge: 00000000d0000000-00000000dfffffff
        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
        BridgeCtl: Parity- SERR- NoISA- VGA+ MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
        Kernel driver in use: pcieport
        Kernel modules: shpchp
00: 86 80 c1 29 07 04 10 00 10 00 04 06 08 00 01 00
10: 00 00 00 00 00 00 00 00 00 01 01 00 a0 a0 00 20
20: 00 e4 f0 e5 01 d0 f1 df 00 00 00 00 00 00 00 00
30: 00 00 00 00 88 00 00 00 00 00 00 00 03 01 08 00

00:1b.0 Audio device: Intel Corporation NM10/ICH7 Family High Definition Audio Controller (rev 01)
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD (Realtek ALC887)
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 29
        Region 0: Memory at f0000000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: <access denied>
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel
00: 86 80 d8 27 06 04 10 00 01 00 03 04 08 00 00 00
10: 04 00 00 f0 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 58 14 02 a0
30: 00 00 00 00 50 00 00 00 00 00 00 00 03 01 00 00

00:1c.0 PCI bridge: Intel Corporation NM10/ICH7 Family PCI Express Port 1 (rev 01) (prog-if 00 [Normal decode])
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 25
        Bus: primary=00, secondary=02, subordinate=02, sec-latency=0
        I/O behind bridge: 00009000-00009fff
        Memory behind bridge: f0100000-f02fffff
        Prefetchable memory behind bridge: 00000000f0300000-00000000f04fffff
        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
        BridgeCtl: Parity- SERR- NoISA- VGA- MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
        Kernel driver in use: pcieport
        Kernel modules: shpchp
00: 86 80 d0 27 07 04 10 00 01 00 04 06 08 00 81 00
10: 00 00 00 00 00 00 00 00 00 02 02 00 90 90 00 20
20: 10 f0 20 f0 31 f0 41 f0 00 00 00 00 00 00 00 00
30: 00 00 00 00 40 00 00 00 00 00 00 00 03 01 00 00

00:1c.1 PCI bridge: Intel Corporation NM10/ICH7 Family PCI Express Port 2 (rev 01) (prog-if 00 [Normal decode])
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin B routed to IRQ 26
        Bus: primary=00, secondary=03, subordinate=03, sec-latency=0
        I/O behind bridge: 0000b000-0000bfff
        Memory behind bridge: e6000000-e7ffffff
        Prefetchable memory behind bridge: 00000000f0500000-00000000f06fffff
        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
        BridgeCtl: Parity- SERR- NoISA- VGA- MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
        Kernel driver in use: pcieport
        Kernel modules: shpchp
00: 86 80 d2 27 07 04 10 00 01 00 04 06 08 00 81 00
10: 00 00 00 00 00 00 00 00 00 03 03 00 b0 b0 00 20
20: 00 e6 f0 e7 51 f0 61 f0 00 00 00 00 00 00 00 00
30: 00 00 00 00 40 00 00 00 00 00 00 00 0a 02 00 00

00:1c.3 PCI bridge: Intel Corporation NM10/ICH7 Family PCI Express Port 4 (rev 01) (prog-if 00 [Normal decode])
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin D routed to IRQ 27
        Bus: primary=00, secondary=04, subordinate=04, sec-latency=0
        I/O behind bridge: 0000c000-0000cfff
        Memory behind bridge: e8000000-e9ffffff
        Prefetchable memory behind bridge: 00000000f0700000-00000000f08fffff
        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
        BridgeCtl: Parity- SERR- NoISA- VGA- MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
        Kernel driver in use: pcieport
        Kernel modules: shpchp
00: 86 80 d6 27 07 04 10 00 01 00 04 06 08 00 81 00
10: 00 00 00 00 00 00 00 00 00 04 04 00 c0 c0 00 20
20: 00 e8 f0 e9 71 f0 81 f0 00 00 00 00 00 00 00 00
30: 00 00 00 00 40 00 00 00 00 00 00 00 0b 04 00 00

00:1d.0 USB controller: Intel Corporation NM10/ICH7 Family USB UHCI Controller #1 (rev 01) (prog-if 00 [UHCI])
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 23
        Region 4: I/O ports at e300 [size=32]
        Kernel driver in use: uhci_hcd
        Kernel modules: uhci_hcd
00: 86 80 c8 27 05 00 80 02 01 00 03 0c 00 00 80 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 01 e3 00 00 00 00 00 00 00 00 00 00 58 14 04 50
30: 00 00 00 00 00 00 00 00 00 00 00 00 0b 01 00 00

00:1d.1 USB controller: Intel Corporation NM10/ICH7 Family USB UHCI Controller #2 (rev 01) (prog-if 00 [UHCI])
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin B routed to IRQ 19
        Region 4: I/O ports at e000 [size=32]
        Kernel driver in use: uhci_hcd
        Kernel modules: uhci_hcd
00: 86 80 c9 27 05 00 80 02 01 00 03 0c 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 01 e0 00 00 00 00 00 00 00 00 00 00 58 14 04 50
30: 00 00 00 00 00 00 00 00 00 00 00 00 0b 02 00 00

00:1d.2 USB controller: Intel Corporation NM10/ICH7 Family USB UHCI Controller #3 (rev 01) (prog-if 00 [UHCI])
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin C routed to IRQ 18
        Region 4: I/O ports at e100 [size=32]
        Kernel driver in use: uhci_hcd
        Kernel modules: uhci_hcd
00: 86 80 ca 27 05 00 80 02 01 00 03 0c 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 01 e1 00 00 00 00 00 00 00 00 00 00 58 14 04 50
30: 00 00 00 00 00 00 00 00 00 00 00 00 06 03 00 00

00:1d.3 USB controller: Intel Corporation NM10/ICH7 Family USB UHCI Controller #4 (rev 01) (prog-if 00 [UHCI])
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin D routed to IRQ 16
        Region 4: I/O ports at e200 [size=32]
        Kernel driver in use: uhci_hcd
        Kernel modules: uhci_hcd
00: 86 80 cb 27 05 00 80 02 01 00 03 0c 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 01 e2 00 00 00 00 00 00 00 00 00 00 58 14 04 50
30: 00 00 00 00 00 00 00 00 00 00 00 00 03 04 00 00

00:1d.7 USB controller: Intel Corporation NM10/ICH7 Family USB2 EHCI Controller (rev 01) (prog-if 20 [EHCI])
        Subsystem: Gigabyte Technology Co., Ltd GA-D525TUD
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 23
        Region 0: Memory at f0004000 (32-bit, non-prefetchable) [size=1K]
        Capabilities: <access denied>
        Kernel driver in use: ehci-pci
        Kernel modules: ehci_pci
00: 86 80 cc 27 06 00 90 02 01 20 03 0c 00 00 00 00
10: 00 40 00 f0 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 58 14 06 50
30: 00 00 00 00 50 00 00 00 00 00 00 00 0b 01 00 00

00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev e1) (prog-if 01 [Subtractive decode])
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Bus: primary=00, secondary=05, subordinate=05, sec-latency=32
        I/O behind bridge: 0000d000-0000dfff
        Memory behind bridge: ea000000-efffffff
        Secondary status: 66MHz- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
        BridgeCtl: Parity- SERR- NoISA- VGA- MAbort- >Reset- FastB2B-
                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
        Capabilities: <access denied>
00: 86 80 4e 24 07 00 10 00 e1 01 04 06 00 00 01 00
10: 00 00 00 00 00 00 00 00 00 05 05 20 d0 d0 80 22
20: 00 ea f0 ef f1 ff 01 00 00 00 00 00 00 00 00 00
30: 00 00 00 00 50 00 00 00 00 00 00 00 ff 00 00 00

00:1f.0 ISA bridge: Intel Corporation 82801GB/GR (ICH7 Family) LPC Interface Bridge (rev 01)
        Subsystem: Gigabyte Technology Co., Ltd Device 5001
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Capabilities: <access denied>
        Kernel driver in use: lpc_ich
        Kernel modules: intel_rng, lpc_ich, leds_ss4200
00: 86 80 b8 27 07 01 10 02 01 00 01 06 00 00 80 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 58 14 01 50
30: 00 00 00 00 e0 00 00 00 00 00 00 00 00 00 00 00

00:1f.1 IDE interface: Intel Corporation 82801G (ICH7 Family) IDE Controller (rev 01) (prog-if 8a [Master SecP PriP])
        Subsystem: Gigabyte Technology Co., Ltd Device b001
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 18
        Region 0: I/O ports at 01f0 [size=8]
        Region 1: I/O ports at 03f4
        Region 2: I/O ports at 0170 [size=8]
        Region 3: I/O ports at 0374
        Region 4: I/O ports at f000 [size=16]
        Kernel driver in use: ata_piix
        Kernel modules: ata_piix, pata_acpi, ata_generic
00: 86 80 df 27 05 00 80 02 01 8a 01 01 00 00 00 00
10: 01 00 00 00 01 00 00 00 01 00 00 00 01 00 00 00
20: 01 f0 00 00 00 00 00 00 00 00 00 00 58 14 01 b0
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00 00

00:1f.2 IDE interface: Intel Corporation NM10/ICH7 Family SATA Controller [IDE mode] (rev 01) (prog-if 8f [Master SecP SecO PriP PriO])
        Subsystem: Gigabyte Technology Co., Ltd Device b002
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin B routed to IRQ 19
        Region 0: I/O ports at e800 [size=8]
        Region 1: I/O ports at e900 [size=4]
        Region 2: I/O ports at ea00 [size=8]
        Region 3: I/O ports at eb00 [size=4]
        Region 4: I/O ports at ec00 [size=16]
        Capabilities: <access denied>
        Kernel driver in use: ata_piix
        Kernel modules: ata_piix, pata_acpi, ata_generic
00: 86 80 c0 27 05 00 b0 02 01 8f 01 01 00 00 00 00
10: 01 e8 00 00 01 e9 00 00 01 ea 00 00 01 eb 00 00
20: 01 ec 00 00 00 00 00 00 00 00 00 00 58 14 02 b0
30: 00 00 00 00 70 00 00 00 00 00 00 00 0b 02 00 00

00:1f.3 SMBus: Intel Corporation NM10/ICH7 Family SMBus Controller (rev 01)
        Subsystem: Gigabyte Technology Co., Ltd GA-8I945PG-RH/GA-D525TUD Mainboard
        Control: I/O+ Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Interrupt: pin B routed to IRQ 19
        Region 4: I/O ports at 0500 [size=32]
        Kernel driver in use: i801_smbus
        Kernel modules: i2c_i801
00: 86 80 da 27 01 00 80 02 01 00 05 0c 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 01 05 00 00 00 00 00 00 00 00 00 00 58 14 01 50
30: 00 00 00 00 00 00 00 00 00 00 00 00 0b 02 00 00

01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV730 PRO [Radeon HD 4650] (prog-if 00 [VGA controller])
        Subsystem: Hightech Information System Ltd. Device 2269
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 28
        Region 0: Memory at d0000000 (64-bit, prefetchable) [size=256M]
        Region 2: Memory at e5000000 (64-bit, non-prefetchable) [size=64K]
        Region 4: I/O ports at a000 [size=256]
        [virtual] Expansion ROM at e4000000 [disabled] [size=128K]
        Capabilities: <access denied>
        Kernel driver in use: radeon
        Kernel modules: radeon
00: 02 10 98 94 07 04 10 00 00 00 00 03 08 00 80 00
10: 0c 00 00 d0 00 00 00 00 04 00 00 e5 00 00 00 00
20: 01 a0 00 00 00 00 00 00 00 00 00 00 87 17 69 22
30: 00 00 00 00 50 00 00 00 00 00 00 00 03 01 00 00

01:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] RV710/730 HDMI Audio [Radeon HD 4000 series]
        Subsystem: Hightech Information System Ltd. Device aa38
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin B routed to IRQ 30
        Region 0: Memory at e5010000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: <access denied>
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel
00: 02 10 38 aa 07 04 10 00 00 00 03 04 08 00 80 00
10: 04 00 01 e5 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 87 17 38 aa
30: 00 00 00 00 50 00 00 00 00 00 00 00 0a 02 00 00

03:00.0 SATA controller: JMicron Technology Corp. JMB363 SATA/IDE Controller (rev 10) (prog-if 01 [AHCI 1.0])
        Subsystem: JMicron Technology Corp. JMB363 SATA/IDE Controller
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 17
        Region 5: Memory at e7000000 (32-bit, non-prefetchable) [size=512]
        [virtual] Expansion ROM at e6000000 [disabled] [size=64K]
        Capabilities: <access denied>
        Kernel driver in use: ahci
        Kernel modules: ahci
00: 7b 19 63 23 07 00 10 00 10 01 06 01 08 00 80 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 e7 00 00 00 00 7b 19 63 23
30: 00 00 00 00 8c 00 00 00 00 00 00 00 0a 01 00 00

03:00.1 IDE interface: JMicron Technology Corp. JMB368 IDE controller (rev 10) (prog-if 85 [Master SecO PriO])
        Subsystem: JMicron Technology Corp. Device 1368
        Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin B routed to IRQ 18
        Region 0: I/O ports at b000 [size=8]
        Region 1: I/O ports at b100 [size=4]
        Region 2: I/O ports at b200 [size=8]
        Region 3: I/O ports at b300 [size=4]
        Region 4: I/O ports at b400 [size=16]
        Capabilities: <access denied>
        Kernel driver in use: pata_jmicron
        Kernel modules: pata_jmicron, pata_acpi, ata_generic
00: 7b 19 68 03 05 00 10 00 10 85 01 01 00 00 80 00
10: 01 b0 00 00 01 b1 00 00 01 b2 00 00 01 b3 00 00
20: 01 b4 00 00 00 00 00 00 00 00 00 00 7b 19 68 13
30: 00 00 00 00 8c 00 00 00 00 00 00 00 00 02 00 00

04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 01)
        Subsystem: Gigabyte Technology Co., Ltd Motherboard
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 31
        Region 0: I/O ports at c000 [size=256]
        Region 2: Memory at e9000000 (64-bit, non-prefetchable) [size=4K]
        [virtual] Expansion ROM at e8000000 [disabled] [size=128K]
        Capabilities: <access denied>
        Kernel driver in use: r8169
        Kernel modules: r8169
00: ec 10 68 81 07 04 10 00 01 00 00 02 08 00 00 00
10: 01 c0 00 00 00 00 00 00 04 00 00 e9 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 58 14 00 e0
30: 00 00 00 00 40 00 00 00 00 00 00 00 0b 01 00 00

05:00.0 Multimedia video controller: Conexant Systems, Inc. CX23880/1/2/3 PCI Video and Audio Decoder (rev 05)
        Subsystem: LeadTek Research Inc. WinFast DTV1000-T
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (5000ns min, 13750ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 20
        Region 0: Memory at ea000000 (32-bit, non-prefetchable) [size=16M]
        Capabilities: <access denied>
        Kernel driver in use: cx8800
        Kernel modules: cx8800
00: f1 14 00 88 06 00 90 02 05 00 00 04 08 40 80 00
10: 00 00 00 ea 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 7d 10 5f 66
30: 00 00 00 00 44 00 00 00 00 00 00 00 09 01 14 37

05:00.2 Multimedia controller: Conexant Systems, Inc. CX23880/1/2/3 PCI Video and Audio Decoder [MPEG Port] (rev 05)
        Subsystem: LeadTek Research Inc. WinFast DTV1000-T
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (1500ns min, 22000ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 20
        Region 0: Memory at eb000000 (32-bit, non-prefetchable) [size=16M]
        Capabilities: <access denied>
        Kernel driver in use: cx88-mpeg driver manager
        Kernel modules: cx8802
00: f1 14 02 88 06 00 90 02 05 00 80 04 08 40 80 00
10: 00 00 00 eb 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 7d 10 5f 66
30: 00 00 00 00 4c 00 00 00 00 00 00 00 09 01 06 58

05:01.0 Multimedia audio controller: Yamaha Corporation YMF-744B [DS-1S Audio Controller] (rev 02)
        Subsystem: Yamaha Corporation DS-XG PCI Audio CODEC
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (1250ns min, 6250ns max)
        Interrupt: pin A routed to IRQ 19
        Region 0: Memory at ef000000 (32-bit, non-prefetchable) [size=32K]
        Region 1: I/O ports at d000 [size=64]
        Region 2: I/O ports at d100 [size=4]
        Capabilities: <access denied>
        Kernel driver in use: snd_ymfpci
        Kernel modules: snd_ymfpci
00: 73 10 10 00 07 00 10 02 02 00 01 04 00 40 00 00
10: 00 00 00 ef 01 d0 00 00 01 d1 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 73 10 10 00
30: 00 00 00 00 50 00 00 00 00 00 00 00 0b 01 05 19

05:02.0 Multimedia video controller: Conexant Systems, Inc. CX23880/1/2/3 PCI Video and Audio Decoder (rev 05)
        Subsystem: LeadTek Research Inc. Device 6654
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (5000ns min, 13750ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 18
        Region 0: Memory at ec000000 (32-bit, non-prefetchable) [size=16M]
        Capabilities: <access denied>
        Kernel driver in use: cx8800
        Kernel modules: cx8800
00: f1 14 00 88 06 00 90 02 05 00 00 04 08 40 80 00
10: 00 00 00 ec 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 7d 10 54 66
30: 00 00 00 00 44 00 00 00 00 00 00 00 06 01 14 37

05:02.1 Multimedia controller: Conexant Systems, Inc. CX23880/1/2/3 PCI Video and Audio Decoder [Audio Port] (rev 05)
        Subsystem: LeadTek Research Inc. Device 6654
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (1000ns min, 63750ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 18
        Region 0: Memory at ed000000 (32-bit, non-prefetchable) [size=16M]
        Capabilities: <access denied>
        Kernel driver in use: cx88_audio
        Kernel modules: cx88_alsa
00: f1 14 01 88 06 00 90 02 05 00 80 04 08 40 80 00
10: 00 00 00 ed 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 7d 10 54 66
30: 00 00 00 00 4c 00 00 00 00 00 00 00 06 01 04 ff

05:02.2 Multimedia controller: Conexant Systems, Inc. CX23880/1/2/3 PCI Video and Audio Decoder [MPEG Port] (rev 05)
        Subsystem: LeadTek Research Inc. Device 6654
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 64 (1500ns min, 22000ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 18
        Region 0: Memory at ee000000 (32-bit, non-prefetchable) [size=16M]
        Capabilities: <access denied>
        Kernel driver in use: cx88-mpeg driver manager
        Kernel modules: cx8802
00: f1 14 02 88 06 00 90 02 05 00 80 04 08 40 80 00
10: 00 00 00 ee 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 7d 10 54 66
30: 00 00 00 00 4c 00 00 00 00 00 00 00 06 01 06 58

Comment 93 Aaron Lu 2015-04-08 08:50:58 UTC

Rui,

I remembered the power on of the PCI device happened during the resume_noirq phase, so you need to do the wait there, i.e. by adding a resume_noirq callback for the pata jmicro driver and wait for its sibling to finish first.

Comment 94 Rainer Fiebig 2015-04-08 17:33:46 UTC

(In reply to Zhang Rui from comment #87)
> Created attachment 172621 [details]
> patch fix jmicron controller dependency
> 
> please check if this patch helps or not.


Your patch works here (3.18.11). 

Hardware as originally posted. Three suspends/resumes, the CD-ROM-drive was always OK.

Is it meant to be an addition to the former patch (lines 153 - 158) or a replacement?

Comment 95 Barto 2015-04-08 17:40:35 UTC

(In reply to Jay from comment #94)

> Your patch works here (3.18.11). 
> 
> Hardware as originally posted. Three suspends/resumes, the CD-ROM-drive was
> always OK.
> 

in order to avoid confusing : do you have an hybrid Jmicron Pcie card ? ( both SATA and IDE ports on the same card )

if you have only an IDE Jmicron card then it's normal if the patch works,

the problem here ( my problem ) only occurs if the Jmicron card has both Sata and IDE ports

Comment 96 Rainer Fiebig 2015-04-08 17:50:50 UTC

(In reply to Barto from comment #95)
> (In reply to Jay from comment #94)
> 
> > Your patch works here (3.18.11). 
> > 
> > Hardware as originally posted. Three suspends/resumes, the CD-ROM-drive was
> > always OK.
> > 
> 
> in order to avoid confusing : do you have an hybrid Jmicron Pcie card ? (
> both SATA and IDE ports on the same card )
> 
> if you have only an IDE Jmicron card then it's normal if the patch works,
> 
> the problem here ( my problem ) only occurs if the Jmicron card has both
> Sata and IDE ports

See Comment 20 (lspci).

Comment 97 Barto 2015-04-08 18:06:12 UTC

(In reply to Jay from comment #96)
> 
> See Comment 20 (lspci).

I have already seen your comment, it's quite confusing because of the name "JMB363 SATA/IDE Controller (rev 03) (prog-if 85 [Master SecO PriO]",

but I assume you have only an IDE port an no sata ports on your JMB363 Jmicron card

Comment 98 Rainer Fiebig 2015-04-08 22:17:11 UTC

(In reply to Barto from comment #97)
> (In reply to Jay from comment #96)
> > 
> > See Comment 20 (lspci).
> 
> I have already seen your comment, it's quite confusing because of the name
> "JMB363 SATA/IDE Controller (rev 03) (prog-if 85 [Master SecO PriO]",
> 
> but I assume you have only an IDE port an no sata ports on your JMB363
> Jmicron card

3.18.11 would have been OK anyway - because it contains the former patch (lines 153 - 158). Therefore my question to Rui.

Now, when I say "it works" I mean that it doesn't cause any problems with 3.18.11 and the CD-ROM-drive wakes up as expected. That's all I can say - sine ira et studio.

BTW: Assumptions are one of the best ways to disaster - because more often than not they turn out to be wrong.
Like in this case: as lspci already suggests, the controller has an IDE- AND a SATA-port. They are physically there on the mainboard. Honestly!

Comment 99 Barto 2015-04-08 22:39:35 UTC

(In reply to Jay from comment #98)

> Like in this case: as lspci already suggests, the controller has an IDE- AND
> a SATA-port. They are physically there on the mainboard. Honestly!

you use an IDE CD-ROM drive for your test, but in my case there are 2 devices connected on my JMB368-363 :

- a SATA DVD burner
- an IDE hard disk

so the use case is not exactly the same, it would be interesting if you could connect also an IDE hard disk on your JMB363 and a SATA device, in order to see if the wake-up process still works after a standby

Comment 100 Barto 2015-04-09 00:09:18 UTC

(In reply to Jay from comment #94)
> 
> Is it meant to be an addition to the former patch (lines 153 - 158) or a
> replacement?

I think the last patch is a replacement, not an addition,

if you talk about the chuanseng's patch then it simply deactivates the async method for JMB363 and JMB361 :

https://bugzilla.kernel.org/attachment.cgi?id=148751

+        /* The JMicron chip 361/363 contains one SATA controller and
+         * one PATA controller,for powering on these both controllers,
+         * we must follow the sequence one by one, otherwise one of them
+         * can not be powered on successfully.
+         * So here we disabled the async suspend method for these chips.
+        */
+        if (pdev->vendor == PCI_VENDOR_ID_JMICRON &&
+                (pdev->device == PCI_DEVICE_ID_JMICRON_JMB363 ||
+                pdev->device == PCI_DEVICE_ID_JMICRON_JMB361))
+                device_disable_async_suspend(&pdev->dev);
+
 	return ata_pci_bmdma_init_one(pdev, ppi, &jmicron_sht, NULL, 0);
 }

that's why your problem seems solved, 

I think if you want to test the last patch ( comment 87 ) you should cancel the first patch,

in my case ( JMB368-363 ) I must use another patch from chuanseng who uses the file quirk.c in order to disable async for all JMicron models :

https://bugzilla.kernel.org/attachment.cgi?id=158351

but I hope developpers will find the solution, they seem close to succeed

Comment 101 Rainer Fiebig 2015-04-09 09:21:19 UTC

(In reply to Barto from comment #100)
> (In reply to Jay from comment #94)
> > 
...
> 
> I think if you want to test the last patch ( comment 87 ) you should cancel
> the first patch,
> 

Chuansheng's patch is in the sources now. 
A replacement needs to address this anyhow. 
So I'd rather wait for a new one.

...

Comment 102 Zhang Rui 2015-04-15 02:24:52 UTC

Hi, All,

After discussing with Aaron, Lu, I think the patch in comment #72 should be the right fix.
Because, on this platform, pci_resume_noirq() of IDE controller must happen after pci_resume_noirq() of AHCI controller. But as .resume_noirq() of the pata_jmicron driver is no-op, this suggests that it is the PCI common actions, aka, pci_pm_default_resume_early(), have the dependency.
To fix this, using device_pm_wait_for_dev() in pata_jmicron driver is not helpful (even in the .resume_noirq() callback) because we should wait BEFORE pci_pm_default_resume_early() being invoked, in pci_pm_resume_noirq().
Plus, as it is the PCI common actions that have the dependency, it is reasonable to fix it in PCI bus code, rather than driver code.

Alan, what do you think?

Comment 103 Barto 2015-04-15 03:19:13 UTC

(In reply to Zhang Rui from comment #102)
> Hi, All,
> 
> After discussing with Aaron, Lu, I think the patch in comment #72 should be
> the right fix.

there is also another patch by Chuanseng who uses the same logic in /drivers/pci/quirks.c file ( disabling the async_suspend method for JMicron chips ), I think you should use this patch : 


https://bugzilla.kernel.org/attachment.cgi?id=158351

diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index 90acb32..b40485f 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -29,6 +29,21 @@
#include "pci.h"
 /*
+ * For JMicron chips, we need to disable the async_suspend method, otherwise
+ * they will hit the power-on issue when doing device resume, add one quick
+ * solution to disable the async_suspend method.
+ */
+static void pci_async_suspend_fixup(struct pci_dev *pdev)
+{
+        /*
+         * disabling the async_suspend method for JMicron chips to
+         * avoid device resuming issue.
+         */
+        device_disable_async_suspend(&pdev->dev);
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_JMICRON, PCI_ANY_ID, pci_async_suspend_fixup);
+
+/*
  * Decoding should be disabled for a PCI device during BAR sizing to avoid
  * conflict. But doing so may cause problems on host bridge and perhaps other
  * key system devices. For devices that need to have mmio decoding always-on,

Comment 104 Aaron Lu 2015-04-15 04:15:56 UTC

(In reply to Barto from comment #103)
> (In reply to Zhang Rui from comment #102)
> > Hi, All,
> > 
> > After discussing with Aaron, Lu, I think the patch in comment #72 should be
> > the right fix.
> 
> there is also another patch by Chuanseng who uses the same logic in
> /drivers/pci/quirks.c file ( disabling the async_suspend method for JMicron
> chips ), I think you should use this patch : 

I prefer this too.

> 
> 
> https://bugzilla.kernel.org/attachment.cgi?id=158351
> 
> diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> index 90acb32..b40485f 100644
> --- a/drivers/pci/quirks.c
> +++ b/drivers/pci/quirks.c
> @@ -29,6 +29,21 @@
> #include "pci.h"
>  /*
> + * For JMicron chips, we need to disable the async_suspend method, otherwise
> + * they will hit the power-on issue when doing device resume, add one quick
> + * solution to disable the async_suspend method.
> + */
> +static void pci_async_suspend_fixup(struct pci_dev *pdev)
> +{
> +        /*
> +         * disabling the async_suspend method for JMicron chips to
> +         * avoid device resuming issue.
> +         */
> +        device_disable_async_suspend(&pdev->dev);
> +}
> +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_JMICRON, PCI_ANY_ID,
> pci_async_suspend_fixup);
> +
> +/*
>   * Decoding should be disabled for a PCI device during BAR sizing to avoid
>   * conflict. But doing so may cause problems on host bridge and perhaps
> other
>   * key system devices. For devices that need to have mmio decoding
> always-on,

Comment 105 Zhang Rui 2015-04-15 05:14:51 UTC

(In reply to Barto from comment #103)
> (In reply to Zhang Rui from comment #102)
> > Hi, All,
> > 
> > After discussing with Aaron, Lu, I think the patch in comment #72 should be
> > the right fix.
> 
> there is also another patch by Chuanseng who uses the same logic in
> /drivers/pci/quirks.c file ( disabling the async_suspend method for JMicron
> chips ), I think you should use this patch : 
> 
sorry, my mistake, this is the patch I wanted to refer as well, instead of the patch in comment #72.

> 
> https://bugzilla.kernel.org/attachment.cgi?id=158351
> 
> diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> index 90acb32..b40485f 100644
> --- a/drivers/pci/quirks.c
> +++ b/drivers/pci/quirks.c
> @@ -29,6 +29,21 @@
> #include "pci.h"
>  /*
> + * For JMicron chips, we need to disable the async_suspend method, otherwise
> + * they will hit the power-on issue when doing device resume, add one quick
> + * solution to disable the async_suspend method.
> + */
> +static void pci_async_suspend_fixup(struct pci_dev *pdev)
> +{
> +        /*
> +         * disabling the async_suspend method for JMicron chips to
> +         * avoid device resuming issue.
> +         */
> +        device_disable_async_suspend(&pdev->dev);
> +}
> +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_JMICRON, PCI_ANY_ID,
> pci_async_suspend_fixup);
> +
> +/*
>   * Decoding should be disabled for a PCI device during BAR sizing to avoid
>   * conflict. But doing so may cause problems on host bridge and perhaps
> other
>   * key system devices. For devices that need to have mmio decoding
> always-on,

Comment 106 Alan Stern 2015-04-15 14:31:09 UTC

Under the circumstances this seems like the best solution.  Unless someone wants to fix up the code in comment #87 and add it as a quirk to the pci_pm_default_resume_early() routine.

Comment 107 Zhang Rui 2015-04-16 02:05:35 UTC

Hi, Chuansheng,

can you please refresh the patch referred in comment #77 and send it for upstream?
Note that you'd better revert commit e6b7e41cdd8cae0591e04d9519b65470110e2d44 in your patch as well.

Comment 108 Zhang Rui 2015-07-26 13:27:49 UTC

Created attachment 183761 [details]
Patch for upstream

Hi, all,
this is the refreshed patch that I'd like to push for upstream. Please kindly test it.

Comment 109 Rainer Fiebig 2015-07-26 16:40:19 UTC

(In reply to Zhang Rui from comment #108)
> Created attachment 183761 [details]
> Patch for upstream
> 
> Hi, all,
> this is the refreshed patch that I'd like to push for upstream. Please
> kindly test it.

Works here with kernel 3.18.18 (s. below).
No problems so far after several resumes from s2both.
Thanks for your work!

Jay

_________________________

/Kernel/linux-3.18.18> patch -p1 <../0001-PCI-Disable-async-suspend-resume-for-Jmicron-chip.patch
patching file drivers/ata/ahci.c
Hunk #1 succeeded at 1378 with fuzz 2 (offset -73 lines).
patching file drivers/ata/pata_jmicron.c
patching file drivers/pci/quirks.c

Comment 110 Barto 2015-07-26 17:54:15 UTC

all is ok for me with your patch with the kernel 4.1.2 ( archlinux 64 bits ),

you can push it for upstream,

thanks for your patch


(In reply to Zhang Rui from comment #108)
> Created attachment 183761 [details]
> Patch for upstream
> 
> Hi, all,
> this is the refreshed patch that I'd like to push for upstream. Please
> kindly test it.

Comment 111 Zhang Rui 2015-07-28 01:43:55 UTC

Patch sent out.
https://patchwork.kernel.org/patch/6878961/

Mark the bug as Resovled for now, and will close it once the patch hits upstream.

Comment 112 Aaron Lu 2015-09-01 02:52:43 UTC

commit 91f15fb30c77d4a0d0d9b97e5cec647650853145
Author: Zhang Rui <rui.zhang@intel.com>
Date:   Mon Aug 24 15:27:11 2015 -0500

    PCI: Disable async suspend/resume for JMicron multi-function SATA/AHCI

Entered Linus tree as of v4.3-rc1.