linux-kernel - Re: [Patch v2 2/3] firmware: dmi_scan: use full dmi version for SMBIOS3

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: <20150413185850.5756dea3@endymion.delvare>
Date:	Mon, 13 Apr 2015 18:58:50 +0200
From:	Jean Delvare <jdelvare@...e.de>
To:	Ivan Khoronzhuk <ivan.khoronzhuk@...aro.org>
Cc:	matt.fleming@...el.com, ard.biesheuvel@...aro.org,
	leif.lindholm@...aro.org, linux-kernel@...r.kernel.org
Subject: Re: [Patch v2 2/3] firmware: dmi_scan: use full dmi version for
 SMBIOS3

Hi Ivan,

Sorry for the very late review.

On Wed, 18 Feb 2015 13:33:20 +0200, Ivan Khoronzhuk wrote:
> New SMBIOS3 spec adds additional field for versioning - docrev.
> The docrev identifies the revision of a specification implemented in
> the table structures, so display SMBIOSv3 versions in format,
> like "3.22.1".

I'm not sure if that will be terribly useful in practice, but it
certainly can't hurt, so why not.

> In case of only 32 bit entry point for versions > 3 display
> dmi version like "3.22.x" as we don't know the docrev.
> 
> In other cases display version like it was.

I don't like this part, see below.

> Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@...aro.org>
> ---
>  drivers/firmware/dmi_scan.c | 20 ++++++++++++--------
>  1 file changed, 12 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c
> index 07d2960..3f3470f 100644
> --- a/drivers/firmware/dmi_scan.c
> +++ b/drivers/firmware/dmi_scan.c
> @@ -17,7 +17,7 @@
>   */
>  static const char dmi_empty_string[] = "        ";
>  
> -static u16 __initdata dmi_ver;
> +static u32 dmi_ver __initdata;
>  /*
>   * Catch too early calls to dmi_check_system():
>   */
> @@ -200,7 +200,7 @@ static void __init dmi_save_uuid(const struct dmi_header *dm, int slot,
>  	 * the UUID are supposed to be little-endian encoded.  The specification
>  	 * says that this is the defacto standard.
>  	 */
> -	if (dmi_ver >= 0x0206)
> +	if (dmi_ver >= 0x020600)
>  		sprintf(s, "%pUL", d);
>  	else
>  		sprintf(s, "%pUB", d);
> @@ -472,7 +472,7 @@ static void __init dmi_format_ids(char *buf, size_t len)
>   */
>  static int __init dmi_present(const u8 *buf)
>  {
> -	int smbios_ver;
> +	u32 smbios_ver;

This makes little sense, as the code handles smbios_ver as a 16-bit
value. You should either leave this alone (this change is not needed
AFAICS?), or declare it as a u16. It only makes sense to make it a u32
if you change the code to store 24-bit values in it as you ultimately
do in dmi_ver.

>  
>  	if (memcmp(buf, "_SM_", 4) == 0 &&
>  	    buf[5] < 32 && dmi_checksum(buf, buf[5])) {
> @@ -507,8 +507,9 @@ static int __init dmi_present(const u8 *buf)
>  		if (dmi_walk_early(dmi_decode) == 0) {
>  			if (smbios_ver) {
>  				dmi_ver = smbios_ver;
> -				pr_info("SMBIOS %d.%d present.\n",
> -				       dmi_ver >> 8, dmi_ver & 0xFF);
> +				pr_info("SMBIOS %d.%d%s present.\n",
> +					dmi_ver >> 8, dmi_ver & 0xFF,
> +					(dmi_ver < 0x0300) ? "" : ".x");

I see zero value in this change. The trailing .x adds no information
for the reader, and if anyone tries to parse that line, this is more
work as they have 3 different formats to handle instead of 2.

>  			} else {
>  				smbios_header_size = 15;
>  				memcpy(smbios_header, buf, smbios_header_size);
> @@ -517,6 +518,7 @@ static int __init dmi_present(const u8 *buf)
>  				pr_info("Legacy DMI %d.%d present.\n",
>  				       dmi_ver >> 8, dmi_ver & 0xFF);
>  			}
> +			dmi_ver <<= 8;

I understand this was the easiest way to implement the change, but I'm
not really comfortable with dmi_ver having different formats for
different parts of the code. This is rather error prone for future
changes, even if the code is currently correct.

>  			dmi_format_ids(dmi_ids_string, sizeof(dmi_ids_string));
>  			printk(KERN_DEBUG "DMI: %s\n", dmi_ids_string);
>  			return 0;
> @@ -534,7 +536,8 @@ static int __init dmi_smbios3_present(const u8 *buf)
>  {
>  	if (memcmp(buf, "_SM3_", 5) == 0 &&
>  	    buf[6] < 32 && dmi_checksum(buf, buf[6])) {
> -		dmi_ver = get_unaligned_be16(buf + 7);
> +		dmi_ver = get_unaligned_be32(buf + 6);
> +		dmi_ver &= 0xFFFFFF;

That would easily fit on a single line.

>  		dmi_len = get_unaligned_le32(buf + 12);
>  		dmi_base = get_unaligned_le64(buf + 16);
>  		smbios_header_size = buf[6];
> @@ -553,8 +556,9 @@ static int __init dmi_smbios3_present(const u8 *buf)
>  		dmi_num = dmi_len / 4;
>  
>  		if (dmi_walk_early(dmi_decode) == 0) {
> -			pr_info("SMBIOS %d.%d present.\n",
> -				dmi_ver >> 8, dmi_ver & 0xFF);
> +			pr_info("SMBIOS %d.%d.%d present.\n",
> +				dmi_ver >> 16, (dmi_ver >> 8) & 0xFF,
> +				dmi_ver & 0xFF);
>  			dmi_format_ids(dmi_ids_string, sizeof(dmi_ids_string));
>  			pr_debug("DMI: %s\n", dmi_ids_string);
>  			return 0;


-- 
Jean Delvare
SUSE L3 Support
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/