lkml.org 
[lkml]   [2004]   [Nov]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCH] kobject: fix double kobject_put in kobject_unregister()
    On Wed, Nov 10, 2004 at 02:54:21PM -0800, Greg KH wrote:
    > On Wed, Nov 10, 2004 at 02:19:23PM -0800, Keshavamurthy Anil S wrote:
    > > Hi Greg,
    > >
    > > This patch fixes the problem where in kobject resources were getting
    > > freed when those kobject were still in use due to double kobject_put()
    > > getting called in the kobject_unregister() code path.
    > >
    > > With out this patch kobject_unregister() will have some serious side effects.
    > >
    > > Please apply.
    > >
    > > signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
    > >
    > >
    > > linux-2.6.10-rc1-mm4-askeshav/lib/kobject.c | 3 ++-
    > > 1 files changed, 2 insertions(+), 1 deletion(-)
    > >
    > > diff -puN lib/kobject.c~kobject_unregister_fix lib/kobject.c
    > > --- linux-2.6.10-rc1-mm4/lib/kobject.c~kobject_unregister_fix 2004-11-10 13:43:42.243877455 -0800
    > > +++ linux-2.6.10-rc1-mm4-askeshav/lib/kobject.c 2004-11-10 13:46:30.788797265 -0800
    > > @@ -301,6 +301,8 @@ void kobject_del(struct kobject * kobj)
    > > {
    > > kobject_hotplug(kobj, KOBJ_REMOVE);
    > > sysfs_remove_dir(kobj);
    > > +
    > > + /* unlink does kobject_put() for us */
    > > unlink(kobj);
    > > }
    > >
    > > @@ -313,7 +315,6 @@ void kobject_unregister(struct kobject *
    > > {
    > > pr_debug("kobject %s: unregistering\n",kobject_name(kobj));
    > > kobject_del(kobj);
    > > - kobject_put(kobj);
    > > }
    >
    > No, this is wrong. Count the add and put in the sequence of:
    > kobject_register()
    > kobject_unregister()
    >
    > they are balanced.
    >
    > You mention you are seeing problems. Have a trace? Example code?

    Here is the call trace when I tried to do kobject_unregiser()
    Call Trace:
    [<a000000100016de0>] show_stack+0x80/0xa0
    sp=e00000000e3cf730 bsp=e00000000e3c9230
    [<a000000100017640>] show_regs+0x840/0x880
    sp=e00000000e3cf900 bsp=e00000000e3c91c8
    [<a000000100023f10>] die+0x150/0x1c0
    sp=e00000000e3cf910 bsp=e00000000e3c9188
    [<a000000100023fc0>] die_if_kernel+0x40/0x60
    sp=e00000000e3cf910 bsp=e00000000e3c9158
    [<a0000001000240f0>] ia64_fault+0x110/0x1060
    sp=e00000000e3cf910 bsp=e00000000e3c9100
    [<a00000010000dae0>] ia64_leave_kernel+0x0/0x260
    sp=e00000000e3cfb20 bsp=e00000000e3c9100
    [<a0000001002743e0>] strlen+0x20/0x140
    sp=e00000000e3cfcf0 bsp=e00000000e3c90a8
    [<a00000010026a270>] kobject_get_path+0x150/0x1e0
    sp=e00000000e3cfcf0 bsp=e00000000e3c9068
    [<a00000010026a770>] kobject_hotplug+0x350/0x6e0
    sp=e00000000e3cfcf0 bsp=e00000000e3c9008
    [<a0000001002695b0>] kobject_del+0x30/0x80
    sp=e00000000e3cfd10 bsp=e00000000e3c8fe0
    [<a000000100269620>] kobject_unregister+0x20/0x60
    sp=e00000000e3cfd10 bsp=e00000000e3c8fc0
    [<a000000100302420>] acpi_device_unregister+0x320/0x340
    sp=e00000000e3cfd10 bsp=e00000000e3c8fa0
    [<a000000100302b70>] acpi_bus_remove+0x2f0/0x340
    sp=e00000000e3cfd10 bsp=e00000000e3c8f68
    [<a000000100302ec0>] acpi_bus_trim+0x300/0x440
    sp=e00000000e3cfd30 bsp=e00000000e3c8f18
    [<a000000100303110>] acpi_eject_store+0x110/0x240
    sp=e00000000e3cfde0 bsp=e00000000e3c8ee0
    [<a000000100301fb0>] acpi_device_attr_store+0x70/0xa0
    sp=e00000000e3cfe20 bsp=e00000000e3c8ea8
    [<a0000001001864b0>] sysfs_write_file+0x270/0x300
    sp=e00000000e3cfe20 bsp=e00000000e3c8e58
    [<a0000001001052e0>] vfs_write+0x200/0x2c0

    I will try to explain the what is happening here.

    I have a sysfs tree like this.
    .../_SB/LSB0
    .../_SB/LSB0/CPU5
    .../_SB/LSB0/MEM1
    .../_SB/LSB0/MEM2

    Say LSB0 is kobj1, CPU5 is kobj2, MEM1 is kobj3, MEM2 is kobj4

    All I am trying to do is first I will do kobject_unregister(kobj2),then
    kobject_unregister(kobj3), then kobject_unregister(kobj4) and now
    when I try to do kobject_unregister(kobj1) i.e when I am trying to remove
    LSB0 directory, I am seeing the above stack trace.

    The patch I posted fixed this problem.

    Let me know if you need more data.

    -Anil



    >
    > thanks,
    >
    > gre k-h
    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 14:07    [W:0.026 / U:181.092 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site