linux-kernel - Re: Loseing my patience with libata and sata

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, 14 Feb 2014 14:16:32 -0800
From:	Randy Dunlap <rdunlap@...radead.org>
To:	Gene Heskett <gheskett@...v.com>
CC:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Loseing my patience with libata and sata_nv

On 02/14/2014 11:21 AM, Gene Heskett wrote:
> On Friday 14 February 2014, Randy Dunlap wrote:
>> On 02/14/2014 08:31 AM, Gene Heskett wrote:
>>> Which is required for my $290 ASUS M2n-SLI Deluxe motherboard to boot.
>>>
>>> Not finding the option in any kernel tree that exists on my system,
>>> except it appears its been replaced or something.
>>>
>>> This once in a lifetime boot, to 3.12.9, shows from an lsmod:
>>> libata                146855  2 sata_nv,pata_amd
>>> scsi_mod              153579  4 sr_mod,sg,sd_mod,libata
>>>
>>> I haven't a clue how I got it in this boot, and I built it.  And
>>> apparently a make oldconfig, carefully done by hand (takes about an
>>> hour 30 to do) is not capable of adding it to a .config BECAUSE it
>>> does NOT exist in any of the arch/x86/configs/i386_defconfig's present
>>> on this machine.
>>
>> It does not have to exist in any defconfig.  Those are just default
>> config files and SATA_NV is not enabled by default, but you can still
>> enable it.
>>
>>> Here is how I searched:
>>>
>>> gene@...ote:~/src/linux-3.0.69$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.0.69$ cd
>>> ../linux-3.2.40
>>> gene@...ote:~/src/linux-3.2.40$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.2.40$ cd
>>> ../linux-3.4.36
>>> gene@...ote:~/src/linux-3.4.36$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.4.36$ cd
>>> ../linux-3.8.2
>>> gene@...ote:~/src/linux-3.8.2$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.8.2$ cd
>>> ../linux-3.8.3
>>> gene@...ote:~/src/linux-3.8.3$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.8.3$ cd
>>> ../linux-3.12.0
>>> gene@...ote:~/src/linux-3.12.0$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.12.0$ cd
>>> ../linux-3.12.6
>>> gene@...ote:~/src/linux-3.12.6$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.12.6$ cd
>>> ../linux-3.12.9
>>> gene@...ote:~/src/linux-3.12.9$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.12.9$ cd
>>> ../linux-3.13.1
>>> gene@...ote:~/src/linux-3.13.1$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.13.1$ cd
>>> ../linux-3.13.2
>>> gene@...ote:~/src/linux-3.13.2$ grep SATA_NV
>>> arch/x86/configs/i386_defconfig gene@...ote:~/src/linux-3.13.2$
>>>
>>> I've been fighting with this, intermittently for about 6 months. 
>>> Except for this 3.12.9 where I must have been holding my mouth right,
>>> all the rest are boot failures because it can't find the boot drive
>>> with so-and-so for a UUID. sata_nv is on the missing list, even in
>>> this WORKing boots defconfig. "Working" however is a miss-statement,
>>> no multimedia stuff was built.
>>>
>>> There has to be a rule I am violating, but even with Randy's D's help,
>>> it is not becoming obvious.  FWIW libata references also can't be
>>> found, but as can be seen, its in memory right now.
>>
>> libata is built whenever CONFIG_ATA is enabled.
> 
> I don't believe the help states that.

AFAICT the Kconfig help does not tell what builds libata.
Would it help if it did?

Maybe by adding 'libata' to this prompt text:
	"Serial ATA and Parallel ATA drivers"
e.g.,
	"Serial ATA and Parallel ATA drivers (libata)"


> 
>>> So please guys, what is the magic dependency that causes libata and
>>> sata_nv to be included in a .config?
>>
>> The SATA_NV kconfig symbol depends on (requires) the following other
>> kconfig symbols to be enabled:
>> 	ATA_SFF and ATA_BMDMA and PCI and ATA
> 
> I wrote down those deps from the make menuconfig help screen, they are as 
> stated.  I made sure with gedit and grep...
> 
> It gets to:
> 
> Early console decompressing the kernel
> booting the kernel
> 
> Thats it, no disk activity, its hung.
> 
> .config for a 3.13.2 attached.
>>
>> If those are not enabled, then you will need to use 'make <some>config'
>> to enabled them before you can enable SATA_NV.

Please add this string to the kernel command line to see if helps to
narrow down what is happening:
	initcall_debug


-- 
~Randy
--
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/