netdev - Re: [Intel-wired-lan] [PATCH iwl-net] i40e: fix hot issue NVM content is corrupted after nvmupdate

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: Mon, 10 Jun 2024 12:18:44 +0200
From: Paul Menzel <pmenzel@...gen.mpg.de>
To: Aleksandr Loktionov <aleksandr.loktionov@...el.com>,
 Kelvin Kang <kelvin.kang@...el.com>
Cc: Jan Sokolowski <jan.sokolowski@...el.com>, netdev@...r.kernel.org,
 Arkadiusz Kubalewski <arkadiusz.kubalewski@...el.com>,
 Anthony L Nguyen <anthony.l.nguyen@...el.com>,
 intel-wired-lan@...ts.osuosl.org, Leon Romanovsky <leonro@...dia.com>
Subject: Re: [Intel-wired-lan] [PATCH iwl-net] i40e: fix hot issue NVM content
 is corrupted after nvmupdate

Dear Aleksandr,


Thank you for your prompt reply.

Am 10.06.24 um 12:14 schrieb Loktionov, Aleksandr:

>> -----Original Message-----
>> From: Paul Menzel <pmenzel@...gen.mpg.de>
>> Sent: Monday, June 10, 2024 11:45 AM

[…]

>> Am 10.06.24 um 11:20 schrieb Aleksandr Loktionov:
>>> After 230f3d53a547 patch all I/O errors are being converted into
>>> EAGAIN which leads to retries until timeout so nvmupdate sometimes
>>> fails after more than 20 minutes!
>>>
>>> Remove misleading EIO to EGAIN conversion and pass all errors as is.
>>>
>>> Fixes: 230f3d53a547 ("i40e: remove i40e_status")
>>
>> This commit is present since v6.6-rc1, released September last year
>> (2023). So until now, nobody noticed this?
>>
> Really, really. The regression affects users only when they update F/W,
> and not all F/W are affected, only that generate I/O errors while update.

Thank you for the clarification.

> Not cars are affected, but the consequences are serous as in subj.

Please details this also in the commit message (body) as asked for below.

>>> Co-developed-by: Kelvin Kang <kelvin.kang@...el.com>
>>> Signed-off-by: Kelvin Kang <kelvin.kang@...el.com>
>>> Reviewed-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@...el.com>
>>> Signed-off-by: Aleksandr Loktionov <aleksandr.loktionov@...el.com>
>>
>> Please give more details about your test setup. For me it’s also
>> not clear, how the NVM content gets corrupted as stated in the
>> summary/title. Could you please elaborate that in the commit
>> message.
>>
>>> ---
>>>    drivers/net/ethernet/intel/i40e/i40e_adminq.h | 4 ----
>>>    1 file changed, 4 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>>> b/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>>> index ee86d2c..55b5bb8 100644
>>> --- a/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>>> +++ b/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>>> @@ -109,10 +109,6 @@ static inline int i40e_aq_rc_to_posix(int aq_ret, int aq_rc)
>>>    		-EFBIG,      /* I40E_AQ_RC_EFBIG */
>>>    	};
>>>
>>> -	/* aq_rc is invalid if AQ timed out */
>>> -	if (aq_ret == -EIO)
>>> -		return -EAGAIN;
>>> -
>>>    	if (!((u32)aq_rc < (sizeof(aq_to_posix) / sizeof((aq_to_posix)[0]))))
>>>    		return -ERANGE;
>>
>> The referenced commit 230f3d53a547 does:
>>
>> ```
>> diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>> b/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>> index ee394aacef4d..267f2e0a21ce 100644
>> --- a/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>> +++ b/drivers/net/ethernet/intel/i40e/i40e_adminq.h
>> @@ -5,7 +5,6 @@
>>    #define _I40E_ADMINQ_H_
>>
>>    #include "i40e_osdep.h"
>> -  #include "i40e_status.h"
>>    #include "i40e_adminq_cmd.h"
>>
>>    #define I40E_ADMINQ_DESC(R, i)   \
>> @@ -117,7 +116,7 @@ static inline int i40e_aq_rc_to_posix(int
>> aq_ret, int aq_rc)
>>           };
>>
>>           /* aq_rc is invalid if AQ timed out */
>> -       if (aq_ret == I40E_ERR_ADMIN_QUEUE_TIMEOUT)
>> +       if (aq_ret == -EIO)
>>                   return -EAGAIN;
>>
>>           if (!((u32)aq_rc < (sizeof(aq_to_posix) /
>> sizeof((aq_to_posix)[0]))))
>> ```
>>
>> So I do not see yet, why removing the whole hunk is the solution.


Kind regards,

Paul