linux-kernel - RE: [PATCH] mpt3sas: Do scsi_remove_host() before deleting SAS PHY objects

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 13 May 2016 21:17:35 +0000
From:	"Elliott, Robert (Persistent Memory)" <elliott@....com>
To:	Calvin Owens <calvinowens@...com>,
	Sathya Prakash <sathya.prakash@...adcom.com>,
	Chaitra P B <chaitra.basappa@...adcom.com>,
	Suganath Prabu Subramani 
	<suganath-prabu.subramani@...adcom.com>,
	"James E.J. Bottomley" <jejb@...ux.vnet.ibm.com>,
	"Martin K. Petersen" <martin.petersen@...cle.com>
CC:	"MPT-FusionLinux.pdl@...adcom.com" <MPT-FusionLinux.pdl@...adcom.com>,
	"linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"kernel-team@...com" <kernel-team@...com>
Subject: RE: [PATCH] mpt3sas: Do scsi_remove_host() before deleting SAS PHY
 objects



> -----Original Message-----
> From: linux-kernel-owner@...r.kernel.org [mailto:linux-kernel-
> owner@...r.kernel.org] On Behalf Of Calvin Owens
> Sent: Friday, May 13, 2016 3:28 PM
...
> Subject: [PATCH] mpt3sas: Do scsi_remove_host() before deleting SAS PHY
> objects
> 
...
> The issue is that enclosure_remove_device() expects to be able to re-add
> the device that was previously enclosured: so with SES running, the order
> we unwind things matters in a way it otherwise wouldn't.
> 
> Since mpt3sas deletes the SAS objects before the SCSI hosts, enclosure
> ends up trying to re-parent the SCSI device from the enclosure to the SAS
> PHY which has already been deleted. This obviously ends in sadness.
> 
> The fix appears to be simple: just call scsi_remove_host() before we call
> sas_port_delete() and/or sas_remove_host().
> 
...
> diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
> @@ -8149,6 +8149,8 @@ void scsih_remove(struct pci_dev *pdev)
>  		_scsih_raid_device_remove(ioc, raid_device);
>  	}
> 
> +	scsi_remove_host(shost);
> +
>  	/* free ports attached to the sas_host */
>  	list_for_each_entry_safe(mpt3sas_port, next_port,
>  	   &ioc->sas_hba.sas_port_list, port_list) {
> @@ -8172,7 +8174,6 @@ void scsih_remove(struct pci_dev *pdev)
>  	}
> 
>  	sas_remove_host(shost);
> -	scsi_remove_host(shost);
>  	mpt3sas_base_detach(ioc);
>  	spin_lock(&gioc_lock);
>  	list_del(&ioc->list);

That change matches the pattern of all other drivers calling
sas_remove_host, except for one:
	drivers/message/fusion/mptsas.c

That consensus means this comment in drivers/scsi/scsi_transport_sas.c
is wrong:

/**
 * sas_remove_host  -  tear down a Scsi_Host's SAS data structures
 * @shost:      Scsi Host that is torn down
 *
 * Removes all SAS PHYs and remote PHYs for a given Scsi_Host.
 * Must be called just before scsi_remove_host for SAS HBAs.
 */