/lib/udev/iphone-set-info consumes 99% CPU

[Bruce Edge]

Here's the htop output showing iphone-set-info spinning at 99% CPU.

622 root 18 -2 50496 3340 2168 R 99.0 0.1 22h17:18 /lib/udev/iphone-set-info

This happens a lot and whenever I feel my machine is sluggish I check for this and kill it.

0 #> uname -a

Linux ice 2.6.32-24-generic #38-Ubuntu SMP Mon Jul 5 09:20:59 UTC 2010 x86_64 GNU/Linux

0 #> cat /etc/lsb-release

DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=10.04
DISTRIB_CODENAME=lucid
DISTRIB_DESCRIPTION="Ubuntu 10.04.1 LTS"

ii libgpod-common 0.7.93-0ubuntu1 common files for libgpod
ii libgpod4 0.7.93-0ubuntu1 library to read and write songs and artwork

[Ivan Kharlamov (the-paper-men)]

It consumes all CPU time of all 3 of my Athlon II cores! Since I run few servers on my machine I consider It a critical bug. Just connect an Iphone and you virtually kill a server.

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 4881 root 18 -2 50408 3204 2144 R 35 0.1 79:00.77 iphone-set-info
 4872 root 18 -2 50408 3204 2144 R 33 0.1 81:11.44 iphone-set-info
 4864 root 18 -2 50408 3204 2144 R 31 0.1 78:36.21 iphone-set-info
 4857 root 18 -2 50408 3196 2144 R 30 0.1 83:27.52 iphone-set-info
 4915 root 18 -2 50408 3200 2140 R 29 0.1 84:16.46 iphone-set-info
 4908 root 18 -2 50408 3232 2180 R 28 0.1 81:10.99 iphone-set-info
 4848 root 18 -2 50408 3200 2144 R 27 0.1 82:52.93 iphone-set-info
 4842 root 18 -2 50408 3196 2144 R 25 0.1 83:23.11 iphone-set-info
 4888 root 18 -2 50408 3204 2144 R 24 0.1 80:45.42 iphone-set-info
 4922 root 18 -2 50408 3192 2140 R 24 0.1 77:40.33 iphone-set-info

[Roland (roland-breedveld)]

   PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3440 root 18 -2 5992 2588 2108 R 98 0.1 40:18.93 iphone-set-info

2.6.32-24-generic #39-Ubuntu SMP

Packages:
ipod-convenience 0.10-0ubuntu3
libimobiledevice0 0.9.7-1ubuntu1
libiphone0 0.9.1-3ubuntu1
libusbmuxd1 1.0.2-1ubuntu2
usbmuxd 1.0.2-1ubuntu2

[Ivan Kharlamov (the-paper-men)]

Developers of this package ask to install the debugging packages for libplist, libimobiledevice and
libgpod, and attach gdb when this happens to get a backtrace of where it's
stuck? (gdb --pid .... as root, and then "thread apply all bt" to get the
backtrace)

[Bruce Edge (bruce-edge)]

I installed the debug packages and started a plug/unplug cycle to get it to happen and saw iphone-set-info is segfaulting:

2010-08-10 08:58:12 usb 1-7 USB disconnect, address 15
2010-08-10 08:58:12 usb 1-7 new high speed USB device using ehci_hcd and address 16
2010-08-10 08:58:12 hub 1-0 .0: unable to enumerate USB device on port 7
2010-08-10 08:58:13 usb 5-1 new full speed USB device using uhci_hcd and address 3
2010-08-10 08:58:13 usb 5-1 not running at top speed; connect to a high speed hub
2010-08-10 08:58:13 usb 5-1 configuration #1 chosen from 4 choices
2010-08-10 08:58:14 iphone-set-info[28589]: segfault at 18 ip 00007fbaf56be156 sp 00007fff5ef5ac00 error 6 in libimobiledevice.so.0.0.0[7fbaf56b8000+14000]
2010-08-10 08:58:14 usb 5-1 USB disconnect, address 3
2010-08-10 08:58:18 usb 1-7 new high speed USB device using ehci_hcd and address 17
2010-08-10 08:58:18 hub 1-0 .0: unable to enumerate USB device on port 7
2010-08-10 08:58:18 usb 1-7 new high speed USB device using ehci_hcd and address 18
2010-08-10 08:58:18 usb 1-7 configuration #1 chosen from 4 choices
2010-08-10 08:58:20 usb 1-7 USB disconnect, address 18
2010-08-10 08:58:20 usb 1-7 new high speed USB device using ehci_hcd and address 19
2010-08-10 08:58:21 hub 1-0 .0: unable to enumerate USB device on port 7
2010-08-10 08:58:21 usb 5-1 new full speed USB device using uhci_hcd and address 4
2010-08-10 08:58:21 usb 5-1 not running at top speed; connect to a high speed hub
2010-08-10 08:58:21 usb 5-1 configuration #1 chosen from 4 choices

And for now, the CPU burn is not happening.

[Ivan Kharlamov (the-paper-men)]

To reproduce the bug I repeatedly connect and disconnect Iphone's usb cord.
Here's gdb output while :
(gdb) thread apply all bt

Thread 1 (Thread 0x7f795d8e7700 (LWP 7953)):
#0 0x00007f795d4dee72 in send () from /lib/libpthread.so.0
#1 0x00007f795b3708bb in usbmuxd_send () from /usr/lib/libusbmuxd.so.1
#2 0x00007f795ca8a01d in internal_connection_send ()
   from /usr/lib/libimobiledevice.so.0
#3 0x00007f795ca8a0fa in internal_ssl_write ()
   from /usr/lib/libimobiledevice.so.0
#4 0x00007f795bbac5da in ?? () from /usr/lib/libgnutls.so.26
#5 0x00007f795bba978d in _gnutls_send_int () from /usr/lib/libgnutls.so.26
#6 0x00007f795bbc607d in gnutls_alert_send () from /usr/lib/libgnutls.so.26
#7 0x00007f795bba9c43 in gnutls_bye () from /usr/lib/libgnutls.so.26
#8 0x00007f795ca89a59 in idevice_connection_disable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#9 0x00007f795ca8a357 in idevice_disconnect ()
   from /usr/lib/libimobiledevice.so.0
#10 0x00007f795ca8bf46 in property_list_service_client_free ()
   from /usr/lib/libimobiledevice.so.0
#11 0x00007f795ca8e2c9 in lockdownd_client_free ()
   from /usr/lib/libimobiledevice.so.0
#12 0x0000000000401582 in ?? ()
#13 0x0000000000401302 in ?? ()
#14 0x00007f795c71fc4d in __libc_start_main () from /lib/libc.so.6
#15 0x00000000004011b9 in ?? ()
#16 0x00007fff4cdec398 in ?? ()
#17 0x000000000000001c in ?? ()
#18 0x0000000000000001 in ?? ()
#19 0x00007fff4cdecd4d in ?? ()
#20 0x0000000000000000 in ?? ()

[Roland (roland-breedveld)]

disconnecting and connecting my iPhone does not make sense, it stay's consuming 1 core of my cpu

[Ivan Kharlamov (the-paper-men)]

Now I cannot reproduce the bug either... I've got somewhat damaged iPhone connector, maybe that's the cause.

[JLotspeich (jlotspeich34)]

Same problem. The process stays resident after the phone has been disconnected. It occurs after multiple connect/disconnect cycles.

Arch:
Linux 2.6.32-25-generic #44-Ubuntu SMP Fri Sep 17 20:05:27 UTC 2010 x86_64

I have an Athlon II processor if that matters.

Here is the backtrace:

(gdb) thread apply all bt

Thread 1 (Thread 0x7ff48efd3700 (LWP 29856)):
#0 0x00007ff48ebd7e72 in send () from /lib/libpthread.so.0
#1 0x00007ff48ca698bb in usbmuxd_send () from /usr/lib/libusbmuxd.so.1
#2 0x00007ff48e18301d in ?? () from /usr/lib/libimobiledevice.so.0
#3 0x00007ff48e1830fa in ?? () from /usr/lib/libimobiledevice.so.0
#4 0x00007ff48d2a55da in ?? () from /usr/lib/libgnutls.so.26
#5 0x00007ff48d2a278d in _gnutls_send_int () from /usr/lib/libgnutls.so.26
#6 0x00007ff48d2bf07d in gnutls_alert_send () from /usr/lib/libgnutls.so.26
#7 0x00007ff48d2a2c43 in gnutls_bye () from /usr/lib/libgnutls.so.26
#8 0x00007ff48e182a59 in idevice_connection_disable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#9 0x00007ff48e183357 in idevice_disconnect ()
   from /usr/lib/libimobiledevice.so.0
#10 0x00007ff48e184f46 in property_list_service_client_free ()
   from /usr/lib/libimobiledevice.so.0
#11 0x00007ff48e1872c9 in lockdownd_client_free ()
   from /usr/lib/libimobiledevice.so.0
#12 0x000000000040182f in ?? ()
---Type <return> to continue, or q <return> to quit---
#13 0x00000000004012ef in ?? ()
#14 0x00007ff48de18c4d in __libc_start_main () from /lib/libc.so.6
#15 0x00000000004011b9 in ?? ()
#16 0x00007fff47d05f58 in ?? ()
#17 0x000000000000001c in ?? ()
#18 0x0000000000000001 in ?? ()
#19 0x00007fff47d07d44 in ?? ()
#20 0x0000000000000000 in ?? ()

[PabloAB (pabloab777)]

I'm not sure if it has something to do, but in my case (iPod Touch 2G) both iphone-set-info and gvfs-afc-volume-monitor starts consuming all CPU. Killing iphone-set-info makes gvfs-afc-volume-monitor on the top of the list all time. The problem with gvfs-afc-volume-monitor is BUG #543597.
Disconnecting and connecting again doesn't change anything.
I should restart because no longer detect my iPod, but it may have something to do with a fast disconnection/connection of the device.

[Brijam (brian-opensourcery)]

Same issue here when I disconnect my iPhone 4.

[Io) (io-e-tic)]

Hello,

aswell for me with iPhone 3GS White, burning CPU once connected. If I unplug/replug, another process starts...

[Sebastian Berthold (sleif)]

same problem on current natty

[Tomaszy (wojownik-swiatla)]

Same here! iPhone 3GS :(
Sometimes happens sometimes doesn't. But as in others' cases: the CPU usage persists after unmounting/unplugging.
iphone-set-info - process responsible

[cparg (cparg)]

strace output:
[...]
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
sendto(4, "\26\3\0\0@\215w\351\271\233\216\311\226\"\207\35z\303\237\201\v\7\370\267\22323\21\17\203K\3"..., 69, 0, NULL, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
sendto(4, "\26\3\0\0@\215w\351\271\233\216\311\226\"\207\35z\303\237\201\v\7\370\267\22323\21\17\203K\3"..., 69, 0, NULL, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
sendto(4, "\26\3\0\0@\215w\351\271\233\216\311\226\"\207\35z\303\237\201\v\7\370\267\22323\21\17\203K\3"..., 69, 0, NULL, 0 <unfinished ...>

[...]

[Emidio Planamente (eplanamente)]

Same problem on Ubuntu 10.04.2.

[dmiller309 (dmiller309)]

I seem to have the same problem. My computer had been upgraded to natty, which was released two days ago. My computer was running really slow, and when I went to log out, it complained that upowerd was still running. Both upowerd and iphone-set-info are running at about 96% CPU. iphone-set-info is using about 3.3MiB RAM. If it's significant, I've never synced my iPhone with any computer ever, as when I tried to initialize the iPhone (3GS) with Rhythmbox or Banshee it failed. I only plug the phone in now to charge it.

[Stephen Rasku (ubuntu-srasku)]

GDB Stack trace. There's a number of calls that have only the address (no name). Did I miss installing something? This is for Ubuntu 10.04.2 LTS :

(gdb) where
#0 0x00007f1ca7703e72 in send () from /lib/libpthread.so.0
#1 0x00007f1ca55958bb in usbmuxd_send () from /usr/lib/libusbmuxd.so.1
#2 0x00007f1ca6caf01d in internal_connection_send ()
   from /usr/lib/libimobiledevice.so.0
#3 0x00007f1ca6caf0fa in internal_ssl_write ()
   from /usr/lib/libimobiledevice.so.0
#4 0x00007f1ca5dd15da in ?? () from /usr/lib/libgnutls.so.26
#5 0x00007f1ca5dce78d in _gnutls_send_int () from /usr/lib/libgnutls.so.26
#6 0x00007f1ca5dd11ab in ?? () from /usr/lib/libgnutls.so.26
#7 0x00007f1ca5dd50a4 in ?? () from /usr/lib/libgnutls.so.26
#8 0x00007f1ca5dd53c1 in ?? () from /usr/lib/libgnutls.so.26
#9 0x00007f1ca5dd55d5 in ?? () from /usr/lib/libgnutls.so.26
#10 0x00007f1ca5dd6c67 in gnutls_handshake () from /usr/lib/libgnutls.so.26
#11 0x00007f1ca6caed13 in idevice_connection_enable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#12 0x00007f1ca6cb0ab5 in property_list_service_enable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#13 0x00007f1ca6cb3215 in lockdownd_start_session ()
   from /usr/lib/libimobiledevice.so.0
#14 0x00007f1ca6cb397b in lockdownd_client_new_with_handshake ()
   from /usr/lib/libimobiledevice.so.0
#15 0x00000000004016ec in ?? ()
#16 0x00000000004012ef in ?? ()
---Type <return> to continue, or q <return> to quit---
#17 0x00007f1ca6944c4d in __libc_start_main () from /lib/libc.so.6
#18 0x00000000004011b9 in ?? ()
#19 0x00007fff47081388 in ?? ()
#20 0x000000000000001c in ?? ()
#21 0x0000000000000001 in ?? ()
#22 0x00007fff47082d25 in ?? ()
#23 0x0000000000000000 in ?? ()

Let me know if I can do anything else for you.

[Stephen Rasku (ubuntu-srasku)]

I just "finished" a bunch of function calls. It seems to be hung in 0x00007f1ca5dd15da whatever that is.

[Stephen Rasku (ubuntu-srasku)]

I added another debug library (libgnutls-dbg, I think) and I have a bit better information:

(gdb) where
#0 0x00007f6d2ae90500 in write () from /lib/libc.so.6
#1 0x00007f6d2ae2b133 in _IO_file_write () from /lib/libc.so.6
#2 0x00007f6d2ae2afdd in _IO_file_xsputn () from /lib/libc.so.6
#3 0x00007f6d2ae02981 in ?? () from /lib/libc.so.6
#4 0x00007f6d2adfd5ae in vfprintf () from /lib/libc.so.6
#5 0x00007f6d2aeb5aab in __fprintf_chk () from /lib/libc.so.6
#6 0x00007f6d29a27964 in usbmuxd_send () from /usr/lib/libusbmuxd.so.1
#7 0x00007f6d2b14101d in internal_connection_send ()
   from /usr/lib/libimobiledevice.so.0
#8 0x00007f6d2b1410fa in internal_ssl_write ()
   from /usr/lib/libimobiledevice.so.0
#9 0x00007f6d2a2635da in _gnutls_io_write_buffered (
    session=<value optimized out>, iptr=<value optimized out>,
    n=<value optimized out>) at gnutls_buffers.c:704
#10 0x00007f6d2a26078d in _gnutls_send_int (session=0xfe98f0,
    type=GNUTLS_HANDSHAKE, htype=<value optimized out>, _data=0xfe39c0,
    sizeofdata=<value optimized out>) at gnutls_record.c:439
#11 0x00007f6d2a2631ab in _gnutls_handshake_io_send_int (session=0xfe98f0,
    type=<value optimized out>, htype=<value optimized out>,
    iptr=<value optimized out>, n=40) at gnutls_buffers.c:914
#12 0x00007f6d2a2670a4 in _gnutls_send_handshake (session=0xfe98f0,
    i_data=0xfe9dd0, i_datasize=<value optimized out>,
    type=GNUTLS_HANDSHAKE_FINISHED) at gnutls_handshake.c:982
---Type <return> to continue, or q <return> to quit---
#13 0x00007f6d2a2673c1 in _gnutls_send_finished (session=0xfe98f0,
    init=-1182648512) at gnutls_handshake.c:576
#14 _gnutls_send_handshake_final (session=0xfe98f0, init=-1182648512)
    at gnutls_handshake.c:2489
#15 0x00007f6d2a2675d5 in _gnutls_handshake_common (session=0xfe98f0)
    at gnutls_handshake.c:2700
#16 0x00007f6d2a268c67 in gnutls_handshake (session=0xfe98f0)
    at gnutls_handshake.c:2297
#17 0x00007f6d2b140d13 in idevice_connection_enable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#18 0x00007f6d2b142ab5 in property_list_service_enable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#19 0x00007f6d2b145215 in lockdownd_start_session ()
   from /usr/lib/libimobiledevice.so.0
#20 0x00007f6d2b14597b in lockdownd_client_new_with_handshake ()
   from /usr/lib/libimobiledevice.so.0
#21 0x00000000004016ec in ?? ()
#22 0x00000000004012ef in ?? ()
#23 0x00007f6d2add6c4d in __libc_start_main () from /lib/libc.so.6
#24 0x00000000004011b9 in ?? ()
#25 0x00007fffb9823b38 in ?? ()
#26 0x000000000000001c in ?? ()
#27 0x0000000000000001 in ?? ()
---Type <return> to continue, or q <return> to quit---
#28 0x00007fffb9824d25 in ?? ()
#29 0x0000000000000000 in ?? ()

I finished all the functions until it hung. It appears to be hanging in _gnutls_io_write_buffered(). This never finishes:

Run till exit from #0 0x00007f6d2a2635da in _gnutls_io_write_buffered (
    session=<value optimized out>, iptr=<value optimized out>,
    n=<value optimized out>) at gnutls_buffers.c:704

[Judd Rogers (juddrogers)]

I can often recreate the problem. Strace shows an endless series of:

send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49) = 49
send(4, "\26\3\0\0@\4\306\23\34\210qA\22\335\317\272\2579\246\301\253l\335\374\311V\317y\333>qb"..., 69, 0) = -1 EPIPE (Broken pipe)
write(2, "usbmuxd_send: Error -1 when send"..., 49^C <unfinished ...>
Process 15233 detached

[Judd Rogers (juddrogers)]

gdb stack trace for iphone-set-info:
#0 0x00a82422 in __kernel_vsyscall ()
#1 0x0094e3c1 in send () from /lib/tls/i686/cmov/libpthread.so.0
#2 0x00a3a374 in usbmuxd_send () from /usr/lib/libusbmuxd.so.1
#3 0x0063de53 in ?? () from /usr/lib/libimobiledevice.so.0
#4 0x0063df6d in ?? () from /usr/lib/libimobiledevice.so.0
#5 0x00f1d262 in ?? () from /usr/lib/libgnutls.so.26
#6 0x00f19e6c in _gnutls_send_int () from /usr/lib/libgnutls.so.26
#7 0x00f1cd4e in ?? () from /usr/lib/libgnutls.so.26
#8 0x00f218d0 in ?? () from /usr/lib/libgnutls.so.26
#9 0x00f21c1c in ?? () from /usr/lib/libgnutls.so.26
#10 0x00f21e78 in ?? () from /usr/lib/libgnutls.so.26
#11 0x00f2384a in gnutls_handshake () from /usr/lib/libgnutls.so.26
#12 0x0063da5e in idevice_connection_enable_ssl () from /usr/lib/libimobiledevice.so.0
#13 0x0063fb46 in property_list_service_enable_ssl ()
   from /usr/lib/libimobiledevice.so.0
#14 0x006425db in lockdownd_start_session () from /usr/lib/libimobiledevice.so.0
#15 0x00642e9a in lockdownd_client_new_with_handshake ()
   from /usr/lib/libimobiledevice.so.0
#16 0x080492e2 in ?? ()
#17 0x08048eca in ?? ()
#18 0x003c6bd6 in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6
#19 0x08048db1 in ?? ()

I can get debug libraries for iphone-set-info if anyone wants.

[Judd Rogers (juddrogers)]

By the way:
 sudo chmod -x /lib/udev/iphone-set-info /usr/lib/gvfs/gvfs-afc-volume-monitor

seems to avoid the problem.

[Geoffrey Paul (geoffro17)]

I encountered the same issue with the Katya release of Linux Mint. After trudging through the debug symbols, I was able to narrow the problem down to libimobiledevice(idevice.o). As it turns out, the internal_ssl_write function (a callback function within libimobiledevice used by libgnutls to send encrypted data) does not perform proper error checking, so instead of returning an error code after the connection closes, it returns 0 bytes sent. The calling function in libgnutls then gets stuck in an infinite loop trying to send data that never gets sent.

I have attached the patch I developed on top of libimobiledevice-1.1.0-ubuntu1. It does not resolve any underlying issues leading to the error code, but it should prevent iphone-set-info and upowerd from taking over the processor.

I hope this helps!

[Ubuntu Foundations Team Bug Bot (crichton)]

The attachment "iphone-cpu-usage.patch" of this bug report has been identified as being a patch. The ubuntu-reviewers team has been subscribed to the bug report so that they can review the patch. In the event that this is in fact not a patch you can resolve this situation by removing the tag 'patch' from the bug report and editing the attachment so that it is not flagged as a patch. Additionally, if you are member of the ubuntu-reviewers team please also unsubscribe the team from this bug report.

[This is an automated message performed by a Launchpad user owned by Brian Murray. Please contact him regarding any issues with the action taken in this bug report.]

[Ivan Kharlamov (the-paper-men)]

Geoffrey, thanks!

[Geoffrey Paul (geoffro17)]

Just wanted to let everyone know that as of 11 Jan 2012, the libimobiledevice developers have integrated my patch into their master development branch. I don't know how long this will take to filter down to the various distributions, so if you can't wait, you can obtain the source code (via git) from their repository: <git://git.sukimashita.com/libimobiledevice.git>.

[Quinta Helmer (qhelmer)]

The bug is still there in oneiric. January (comment 26) is 2 months ago; when will this get patched in the distributions?

[Ivan Kharlamov (the-paper-men)]

Quinta, consider contacting Ubuntu Bug Squad: https://wiki.ubuntu.com/BugSquad/FAQ#Where_can_I_contact_the_BugSquad_.3F

I've moved to Debian, so I no longer have Ubuntu installed.

[Geoffrey Paul (geoffro17)]

Quinta, you could also ping the package maintainers for libimobiledevice. Last I checked, my patch was present in version 1.1.2. The latest Ubuntu version is 1.1.1-4.