Re: [PATCH v2] ide: add missing hwif->portdev freeing on hwif_init() failure

[Bjorn Helgaas]

[+cc linux-pci for real]

My plan is to take these two additional fixes via the PCI tree, along
with Lorenzo's patch to add a pci_assign_irq() call in
ide_scan_pcidev(), since they're all related.

On Tue, Oct 03, 2017 at 01:18:47PM +0200, Bartlomiej Zolnierkiewicz wrote:
> Recent pci_assign_irq() changes uncovered a problem with missing
> freeing of ide_port class instance on hwif_init() failure in 
> ide_host_register():
> 
> ide0: disabled, no IRQ
> ide0: failed to initialize IDE interface
> ide0: disabling port
> cmd64x 0000:00:02.0: IDE controller (0x1095:0x0646 rev 0x07)
> CMD64x_IDE 0000:00:02.0: BAR 0: can't reserve [io  0x8050-0x8057]
> cmd64x 0000:00:02.0: can't reserve resources
> CMD64x_IDE: probe of 0000:00:02.0 failed with error -16
> ide_generic: please use "probe_mask=0x3f" module parameter for probing
> all legacy ISA IDE ports
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 1 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x94/0xd0
> sysfs: cannot create duplicate filename '/class/ide_port/ide0'
> ...
> 
> Trace:
> [<fffffc00003308a0>] __warn+0x160/0x190
> [<fffffc000048c9f4>] sysfs_warn_dup+0x94/0xd0
> [<fffffc0000330928>] warn_slowpath_fmt+0x58/0x70
> [<fffffc000048c9f4>] sysfs_warn_dup+0x94/0xd0
> [<fffffc0000486d40>] kernfs_path_from_node+0x30/0x60
> [<fffffc00004874ac>] kernfs_put+0x16c/0x2c0
> [<fffffc00004874ac>] kernfs_put+0x16c/0x2c0
> [<fffffc000048d010>] sysfs_do_create_link_sd.isra.2+0x100/0x120
> [<fffffc00005b9d64>] device_add+0x2a4/0x7c0
> [<fffffc00005ba5cc>] device_create_groups_vargs+0x14c/0x170
> [<fffffc00005ba518>] device_create_groups_vargs+0x98/0x170
> [<fffffc00005ba690>] device_create+0x50/0x70
> [<fffffc00005df36c>] ide_host_register+0x48c/0xa00
> [<fffffc00005df330>] ide_host_register+0x450/0xa00
> [<fffffc00005ba2a0>] device_register+0x20/0x50
> [<fffffc00005df330>] ide_host_register+0x450/0xa00
> [<fffffc00005df944>] ide_host_add+0x64/0xe0
> [<fffffc000079b41c>] kobject_uevent_env+0x16c/0x710
> [<fffffc0000310288>] do_one_initcall+0x68/0x260
> [<fffffc00007b13bc>] kernel_init+0x1c/0x1a0
> [<fffffc00007b13a0>] kernel_init+0x0/0x1a0
> [<fffffc0000311868>] ret_from_kernel_thread+0x18/0x20
> [<fffffc00007b13a0>] kernel_init+0x0/0x1a0
> 
> ---[ end trace 24a70433c3e4d374 ]---
> ide0: disabling port
> 
> Fix the problem by adding missing code to ide_host_register().
> 
> Link: http://lkml.kernel.org/r/32ec730f-c1b0-5584-cd35-f8a809122b96@xxxxxxxxxxxx
> Reported-by: Guenter Roeck <linux@xxxxxxxxxxxx>
> Cc: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> Cc: Richard Henderson <rth@xxxxxxxxxxx>
> Cc: Ivan Kokshaysky <ink@xxxxxxxxxxxxxxxxxxxx>
> Cc: Guenter Roeck <linux@xxxxxxxxxxxx>
> Cc: Matt Turner <mattst88@xxxxxxxxx>
> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@xxxxxxxxxxx>
> ---
> This should fix sysfs warning and probing of ide_generic driver
> driven ports in case of hwif_init() failure (i.e. no IRQ issue
> when pci_assign_irq() fix is not present).
> 
> v2:
> - Fixed typo in the patch description: pci_assing_irq() -> pci_assign_irq().
> 
>  drivers/ide/ide-probe.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> Index: b/drivers/ide/ide-probe.c
> ===================================================================
> --- a/drivers/ide/ide-probe.c	2017-10-02 15:08:16.658662459 +0200
> +++ b/drivers/ide/ide-probe.c	2017-10-03 11:59:36.867189557 +0200
> @@ -1451,6 +1451,7 @@ int ide_host_register(struct ide_host *h
>  		if (hwif_init(hwif) == 0) {
>  			printk(KERN_INFO "%s: failed to initialize IDE "
>  					 "interface\n", hwif->name);
> +			device_unregister(hwif->portdev);
>  			device_unregister(&hwif->gendev);
>  			ide_disable_port(hwif);
>  			continue;
>