linux-kernel - RE: [PATCH 2/2] EDAC/ghes: Modularize ghes_edac driver to remove the dependency on ghes

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]

Message-ID: <DBBPR08MB45389A9DB098F1AC14C19074F76B9@DBBPR08MB4538.eurprd08.prod.outlook.com>
Date:   Tue, 16 Aug 2022 02:20:19 +0000
From:   Justin He <Justin.He@....com>
To:     Borislav Petkov <bp@...en8.de>,
        "toshi.kani@....com" <toshi.kani@....com>
CC:     Ard Biesheuvel <ardb@...nel.org>, Len Brown <lenb@...nel.org>,
        James Morse <James.Morse@....com>,
        Tony Luck <tony.luck@...el.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Robert Richter <rric@...nel.org>,
        Robert Moore <robert.moore@...el.com>,
        "linux-acpi@...r.kernel.org" <linux-acpi@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>,
        "devel@...ica.org" <devel@...ica.org>,
        "Rafael J . Wysocki" <rafael@...nel.org>,
        Shuai Xue <xueshuai@...ux.alibaba.com>,
        Jarkko Sakkinen <jarkko@...nel.org>,
        "linux-efi@...r.kernel.org" <linux-efi@...r.kernel.org>,
        nd <nd@....com>, "stable@...nel.org" <stable@...nel.org>
Subject: RE: [PATCH 2/2] EDAC/ghes: Modularize ghes_edac driver to remove the
 dependency on ghes

Hi Borislav & Toshi
Please see my questions below:

> -----Original Message-----
> From: Borislav Petkov <bp@...en8.de>
> Sent: Friday, August 12, 2022 10:46 PM
> To: Justin He <Justin.He@....com>
> Cc: Ard Biesheuvel <ardb@...nel.org>; Len Brown <lenb@...nel.org>; James
> Morse <James.Morse@....com>; Tony Luck <tony.luck@...el.com>; Mauro
> Carvalho Chehab <mchehab@...nel.org>; Robert Richter <rric@...nel.org>;
> Robert Moore <robert.moore@...el.com>; linux-acpi@...r.kernel.org;
> linux-kernel@...r.kernel.org; linux-edac@...r.kernel.org; devel@...ica.org;
> Rafael J . Wysocki <rafael@...nel.org>; Shuai Xue
> <xueshuai@...ux.alibaba.com>; Jarkko Sakkinen <jarkko@...nel.org>;
> linux-efi@...r.kernel.org; nd <nd@....com>; toshi.kani@....com;
> stable@...nel.org
> Subject: Re: [PATCH 2/2] EDAC/ghes: Modularize ghes_edac driver to remove
> the dependency on ghes
> 
> On Thu, Aug 11, 2022 at 09:17:13AM +0000, Jia He wrote:
> > diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c index
> > d91ad378c00d..ed6519f3d45b 100644
> > --- a/drivers/acpi/apei/ghes.c
> > +++ b/drivers/acpi/apei/ghes.c
> > @@ -94,6 +94,8 @@
> >  #define FIX_APEI_GHES_SDEI_CRITICAL	__end_of_fixed_addresses
> >  #endif
> >
> > +ATOMIC_NOTIFIER_HEAD(ghes_report_chain);
> 
> static. You need function wrappers which call the notifier from the module.
> 
> Also, why atomic? x86_mce_decoder_chain is a blocking one.
> 
> Also, the whole notifier adding thing needs to be a separate patch.
> 
> > @@ -1497,3 +1504,37 @@ void __init acpi_ghes_init(void)
> >  	else
> >  		pr_info(GHES_PFX "Failed to enable APEI firmware first mode.\n");
> > }
> > +
> > +/*
> > + * Known x86 systems that prefer GHES error reporting:
> > + */
> > +static struct acpi_platform_list plat_list[] = {
> > +	{"HPE   ", "Server  ", 0, ACPI_SIG_FADT, all_versions},
> > +	{ } /* End */
> > +};
> > +
> > +struct list_head *ghes_get_devices(bool force) {
> > +	int idx = -1;
> > +
> > +	if (IS_ENABLED(CONFIG_X86)) {
> > +		idx = acpi_match_platform_list(plat_list);
> > +		if (idx < 0 && !force)
> > +			return NULL;
> > +	}
> > +
> > +	return &ghes_devs;
> > +}
> > +EXPORT_SYMBOL_GPL(ghes_get_devices);
> 
> And yes, as Toshi points out, the other EDAC drivers - sb_edac, skx_edac and
> amd64_edac - should call this function in their init functions so that it can get
> selected which driver to load on HPE server platforms.
> 
I assume that all those edac drivers which used owner checking are
impacted, right? So the impacted list should be:
drivers/edac/pnd2_edac.c:1532:  if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))
drivers/edac/sb_edac.c:3513:    if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))
drivers/edac/amd64_edac.c:4333: if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))
drivers/edac/i10nm_base.c:552:  if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))
drivers/edac/skx_base.c:657:    if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))
drivers/edac/igen6_edac.c:1275: if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR)))

Furthermore, should I totally remove the owner check in above driver
XX_init()? Because after the new helper ghes_get_device() checking, those
drivers can't be initialized after ghes_edac is loaded. If ghes_edac is NOT
loaded, the edac_mc_owner check in edac_mc_add_mc_with_groups() can guarantee
that there is only one regular edac driver.

What do you think of it?

--
Cheers,
Justin (Jia He)