Re: [PATCH] PCI: export MSI mode using attributes, not kobjects

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Thu, Dec 12, 2013 at 04:56:20PM -0700, Bjorn Helgaas wrote:
> On Thu, Dec 12, 2013 at 4:17 PM, Yinghai Lu <yinghai@kernel.org> wrote:
> > On Sat, Dec 7, 2013 at 12:41 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> >> On Wed, Nov 27, 2013 at 11:46 AM, Greg Kroah-Hartman
> >> <gregkh@linuxfoundation.org> wrote:
> >>> From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> >>>
> >>> The PCI MSI sysfs code is a mess with kobjects for things that don't
> >>> really need to be kobjects.  This patch creates attributes dynamically
> >>> for the MSI interrupts instead of using kobjects.
> >>>
> >>> Note, this removes a directory from the current MSI interrupt sysfs
> >>> code:
> >>>
> >>> old MSI kobjects:
> >>> pci_device
> >>>    └── msi_irqs
> >>>        └── 40
> >>>            └── mode
> >>>
> >>> new MSI attributes:
> >>> pci_device
> >>>    └── msi_irqs
> >>>        └── 40
> >>>
> >>> As there was only one file "mode" with the kobject model, the interrupt
> >>> number is now a file that returns the "mode" of the interrupt (msi vs.
> >>> msix).
> >>>
> >>> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> >>
> >> I added the acks from Neil and Veaceslav, folded in the
> >> Documentation/ABI update, and applied the whole thing to my pci/misc
> >> branch for v3.14.  Thanks!
> >
> > got:
> > [   71.429735] BUG: key ffff887fcf082a58 not in .data!
> > [   71.429737] ------------[ cut here ]------------
> > [   71.429742] WARNING: CPU: 0 PID: 4 at kernel/locking/lockdep.c:2987 lockdep_i
> > nit_map+0x127/0x5b0()
> > [   71.429743] DEBUG_LOCKS_WARN_ON(1)
> > [   71.429744] Modules linked in:
> > [   71.429747] CPU: 0 PID: 4 Comm: kworker/0:0 Tainted: G          I  3.13.0-rc3
> > -yh-01187-ge0e4e4e-dirty #45
> > [   71.429757] Workqueue: events work_for_cpu_fn
> > [   71.429792]  0000000000000009 ffff881fcf95da58 ffffffff82020475 ffff881fcf95d
> > aa0
> > [   71.429800]  ffff881fcf95da90 ffffffff81097c3d ffff887fcf082a58 ffff88dfcec1a
> > d28
> > [   71.429809]  0000000000000000 0000000000000000 ffff889fcf2eecd0 ffff881fcf95d
> > af0
> > [   71.429809] Call Trace:
> > [   71.429814]  [<ffffffff82020475>] dump_stack+0x45/0x56
> > [   71.429818]  [<ffffffff81097c3d>] warn_slowpath_common+0x7d/0xa0
> > [   71.429822]  [<ffffffff81097cac>] warn_slowpath_fmt+0x4c/0x50
> > [   71.429826]  [<ffffffff810e3d97>] lockdep_init_map+0x127/0x5b0
> > [   71.429835]  [<ffffffff81242fe8>] ? sysfs_new_dirent+0x98/0x140
> > [   71.429839]  [<ffffffff81242033>] sysfs_add_file_mode_ns+0x63/0xc0
> > [   71.429843]  [<ffffffff8124491d>] internal_create_group+0x18d/0x260
> > [   71.429853]  [<ffffffff81556345>] ? populate_msi_sysfs+0x185/0x1d0
> > [   71.429857]  [<ffffffff81244a52>] sysfs_create_groups+0x42/0xa0
> > [   71.429861]  [<ffffffff81556367>] populate_msi_sysfs+0x1a7/0x1d0
> > [   71.429865]  [<ffffffff81556b67>] pci_enable_msi_block+0x1f7/0x2a0
> > [   71.429870]  [<ffffffff81549795>] pcie_port_device_register+0x335/0x520
> > [   71.429874]  [<ffffffff81549f08>] pcie_portdrv_probe+0x68/0xa0
> > [   71.429883]  [<ffffffff8153cf75>] local_pci_probe+0x45/0xa0
> > [   71.429887]  [<ffffffff810b0af4>] work_for_cpu_fn+0x14/0x20
> > [   71.429891]  [<ffffffff810b431b>] process_one_work+0x28b/0x4a0
> > [   71.429895]  [<ffffffff810b4292>] ? process_one_work+0x202/0x4a0
> > [   71.429899]  [<ffffffff810b55db>] worker_thread+0x26b/0x3a0
> > [   71.429903]  [<ffffffff810e262d>] ? trace_hardirqs_on+0xd/0x10
> > [   71.429906]  [<ffffffff810b5370>] ? manage_workers.isra.17+0x340/0x340
> > [   71.429912]  [<ffffffff810bc031>] kthread+0x111/0x120
> > [   71.429919]  [<ffffffff810cddab>] ? local_clock+0x2b/0x40
> > [   71.429923]  [<ffffffff810bbf20>] ? kthread_stop+0xf0/0xf0
> > [   71.429927]  [<ffffffff8203842c>] ret_from_fork+0x7c/0xb0
> > [   71.429931]  [<ffffffff810bbf20>] ? kthread_stop+0xf0/0xf0
> > [   71.429933] ---[ end trace c511e3d74efea94e ]---
> >
> > looks like Greg forgot adding attr init.
> >
> > Can you fold attached patch into Greg's patch ?
> 
> Sure.  I don't know how to reproduce it, so I couldn't verify the fix,
> but I added it to Greg's patch.  Thanks!

I think lockdep has to be enabled for this to show up, I wasn't running
with that enabled when I tested the code, my fault, sorry.

greg k-h