linux-kernel - Re: [PATCH v2 4/7] ghes_edac: avoid multiple calls to dmi

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:   Wed, 16 Aug 2017 15:26:04 +0000
From:   "Kani, Toshimitsu" <toshi.kani@....com>
To:     "tony.luck@...el.com" <tony.luck@...el.com>,
        "rostedt@...dmis.org" <rostedt@...dmis.org>,
        "bp@...en8.de" <bp@...en8.de>
CC:     "linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
        "lenb@...nel.org" <lenb@...nel.org>,
        "mchehab@...nel.org" <mchehab@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "rjw@...ysocki.net" <rjw@...ysocki.net>,
        "linux-acpi@...r.kernel.org" <linux-acpi@...r.kernel.org>
Subject: Re: [PATCH v2 4/7] ghes_edac: avoid multiple calls to dmi_walk()

On Wed, 2017-08-16 at 10:29 +0200, Borislav Petkov wrote:
> On Tue, Aug 15, 2017 at 08:48:16AM -0700, Luck, Tony wrote:
> > Won't the user see all their DIMMs reported for each memory
> > controller
> > under /sys/devices/system/edac/mc/mc*/dimm* ?
> > 
> > That sounds confusing.
> 
> Right, and adding the locking was really easy. If only people would
> debate less and actually try to do what they're being advised to.
> But not really: if you wanna have something done, you have to do it
> yourself.

Sorry, but I did not agree on allowing concurrent accesses to mci...

 /* Memory Device - Type 17 of SMBIOS spec */
>  struct memdev_dmi_entry {
> @@ -169,14 +173,11 @@ void ghes_edac_report_mem_error(struct ghes
> *ghes, int sev,
>  	enum hw_event_mc_err_type type;
>  	struct edac_raw_error_desc *e;
>  	struct mem_ctl_info *mci;
> -	struct ghes_edac_pvt *pvt = NULL;
> +	struct ghes_edac_pvt *pvt = ghes_pvt;
> +	unsigned long flags;
>  	char *p;
>  	u8 grain_bits;

I believe you now need to protect from a race condition that a single
mci and pvt can be initialized / consumed from multiple threads.  This
protection is missing in your patch.

Thanks,
-Toshi