Power Global

 View Only
Expand all | Collapse all

CVE 2023-30438

  • 1.  CVE 2023-30438

    IBM Champion
    Posted Mon May 22, 2023 09:42 AM

    https://www.itjungle.com/2023/05/22/critical-security-vulnerability-in-powervm-hypervisor/


    The firmware was concurrent for my case.


    I am now at:
    Installed level:  fw1030.01 (030)
    Activated level:  fw1030.10 (058)
    Deferred level:  FW1030.11 (45)
    Is this CVE addressed on my machine or does it still require an IPL to remove the Deferred level and make it the Activated level?  Or does the deferred stuff just refer to other hardware, etc items?



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------


  • 2.  RE: CVE 2023-30438

    Posted Tue May 23, 2023 03:00 PM

    Can you confirm that the Deferred level is really shown as FW1030.11 (045).  We would have expected FW1030.10 (045).



    ------------------------------
    Pete Heyrman
    ------------------------------



  • 3.  RE: CVE 2023-30438

    IBM Champion
    Posted Tue May 23, 2023 03:37 PM



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 4.  RE: CVE 2023-30438

    IBM Champion
    Posted Tue May 23, 2023 03:41 PM

    You should see the other Power 10.  They both started out at the same level but the other one had an issue (case opened) and now it's at the following.

    When I tried to apply the update again (as recommended in the case) it said it would be disruptive.  I had to cancel.



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 5.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 10:13 AM

    I would expect it disruptive for sure.



    ------------------------------
    Alan Fulton Follow me on Twitter - @The_Iron_Monger
    Budd Lake
    2015329657
    ------------------------------



  • 6.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 02:11 PM

    I don't understand why it should be disruptive.  I am trying to go from ML 1030_030 to ML1030_058_026.  Last I knew 030 was in between 058 and 026 and should not be disruptive.  I would understand if I was at a level of 1030 less than 026.



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 7.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 02:28 PM

    True  - but you have only activated on the .10 level



    ------------------------------
    Alan Fulton Follow me on Twitter - @The_Iron_Monger
    Budd Lake
    2015329657
    ------------------------------



  • 8.  RE: CVE 2023-30438

    Posted Thu May 25, 2023 09:22 AM

    I'm sorry if I'm hijacking the thread...but...

    My experience is that even the non-disruptive updates give a message at the end that goes something like: "The update has been applied concurrently.  You must restart each partition to apply the update to that partition."

    Ummmmm...if I have to restart every partition, then I might as well restart the entire frame.  Well...I suppose I could restart the partitions at different times, so there's a little bit more flexibility.  But I just restart the frame every time I do an update so that the update is completely applied.

    Am I missing something?



    ------------------------------
    Anker Lerret
    Blue Cross Blue Shield of Massachusetts
    ------------------------------



  • 9.  RE: CVE 2023-30438

    IBM Champion
    Posted Thu May 25, 2023 09:33 AM

    I don't recall having ever received that message.  Maybe I just blew by it?



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 10.  RE: CVE 2023-30438

    IBM Champion
    Posted Thu May 25, 2023 09:36 AM

    Hi
    when you update any firmware which will be disruptive you will be warned at HMC GUI...with CLI i have not tried



    ------------------------------
    Vincencio Michaelis
    ------------------------------



  • 11.  RE: CVE 2023-30438

    Posted Fri May 26, 2023 04:28 AM

    Yes, it does sound like you're missing something.  The only time I've seen that message is when there isn't a functional RMC connection from the HMC to the LPAR (AIX LPARs).  During the concurrent firmware update process the HMC needs to be able to instruct the LPAR to reset/reload the CPUs that it's using, this is done via RMC.

    Phill.



    ------------------------------
    Phill Rowbottom
    ------------------------------



  • 12.  RE: CVE 2023-30438

    Posted Fri May 26, 2023 08:41 AM

    Thanks, Phill.  I'll keep an eye out for the message the next time I do a firmware upgrade.






  • 13.  RE: CVE 2023-30438

    Posted Fri May 26, 2023 05:58 AM

    IIRC the only time I've seen that message is when RMC connection between the HMC and one or more managed LPARs, wasn't working.

    IMHO part of the physical server firmware is part of a LPAR.
    This is based on a memory of a image that contained this information and when I've updated firmware on server that has sufficiently large number of LPARs, the HMC GUI showed how many LPARs had been updated, at the end of "for the Current Release..." firmware update.



    ------------------------------
    Esa Kärkkäinen
    ------------------------------



  • 14.  RE: CVE 2023-30438

    IBM Champion
    Posted Fri May 26, 2023 08:57 AM

    About that RMC connection... Does it matter if RMC connection only shows up for AIX and VIOS lpars and not IBM i?



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 15.  RE: CVE 2023-30438

    IBM Champion
    Posted Fri May 26, 2023 09:07 AM

    IBMi doesn't have/use RMC.  It uses signaling via the hypervisor.



    ------------------------------
    José Pina Coelho
    IT Specialist at Kyndryl
    ------------------------------



  • 16.  RE: CVE 2023-30438

    IBM Champion
    Posted Fri May 26, 2023 06:12 AM

    Some updates do need an LPAR restart because they affect something that is built/run inside the LPAR (like the OF device tree, or some hypervisor-related memory buffer).

    The system update is marked as concurrent (because you don't have to stop the frame), this is good because it allows you to update a system without needing a cross-lpar maintenance window.  However, some of it's functionality is only applied when you reboot the LPARs, which can be restartede on their own independent windows.



    ------------------------------
    José Pina Coelho
    IT Specialist at Kyndryl
    ------------------------------



  • 17.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 11:06 AM

    Hi Robert,

    you have to power off and on the server. Your activated level is 1030.10 so your system is still affected by the vulnerability.

    Regards.



    ------------------------------
    Andres Cordoba
    Unix Specialist
    Dia Corporate
    Madrid
    +34 676934659
    ------------------------------



  • 18.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 12:18 PM

    In the readme for ML1030_058 / FW1030.11 (https://www.ibm.com/support/pages/node/6982699) the following information can be found in the Important Information section:
    NOTE:  The HMC will show the following level once 1030.11 is installed and activated: fw1030.10 (058)



    ------------------------------
    Pete Heyrman
    ------------------------------



  • 19.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 02:36 PM

    Hello Team. We have similar question for that published CVE, but we are running a Power9 server:

    We have firmware level at VH950_092_045 (30)
    Required firmware level is VH950_124_045 (71)

    So this should be an update if am correct, right?

    file name convention from documentation:
    01VHxxx_yyy_zzz 
    xxx is the release level 
    yyy is the service pack level 
    zzz is the last disruptive service pack level 

    An installation is concurrent if: The release level (xxx) is the same, and The service pack level (yyy) currently installed on the system is the same or higher than the last disruptive service pack level (zzz) of the service pack to be installed. 
    (In our case 092 > 045)

    Our question then is, if this is a concurrent update> from HMC GUI (that is, Updates>Change LIC> "for the current release") 
    Would that process not need to power off the server? 

    Also, the LPARs from the server to be updated , should they have to be IPLed? 
    We just want to be sure if this concurrent installation procedures has to be done with any LPAR/server/or any other device to be turned off , or doing it concurrently it is a "transparent" installation and there is no need to power off/IPL anything? 


    Thanks in advance
    Gustavo



    ------------------------------
    Gustavo Orlando de Santis
    ------------------------------



  • 20.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 03:00 PM

    The update may be concurrent in your case.  However you are still vulnerable to the attack if your 'activated' level is not up to snuff.  So, therefore, I would just plan on doing the update as disruptive and power down all your lpars first and IPL the rack after doing the update.



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 21.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 02:38 PM

    Hello, 

    We have similar question for that published CVE, but we are running a Power9 server:

    We have firmware level at VH950_092_045 (30)
    Required firmware level is VH950_124_045 (71)

    So this should be an update if am correct, right?

    file name convention from documentation:
    01VHxxx_yyy_zzz 
    xxx is the release level 
    yyy is the service pack level 
    zzz is the last disruptive service pack level 

    An installation is concurrent if: The release level (xxx) is the same, and The service pack level (yyy) currently installed on the system is the same or higher than the last disruptive service pack level (zzz) of the service pack to be installed. 
    (In our case 092 > 045)

    Our question then is, if this is a concurrent update> from HMC GUI (that is, Updates>Change LIC> "for the current release") 
    Would that process not need to power off the server? 

    Also, the LPARs from the server to be updated , should they have to be IPLed? 
    We just want to be sure if this concurrent installation procedures has to be done with any LPAR/server/or any other device to be turned off , or doing it concurrently it is a "transparent" installation and there is no need to power off/IPL anything? 


    Thanks in advance







    ------------------------------
    Gustavo Orlando de Santis
    ------------------------------



  • 22.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 03:08 PM

    A service pack can be applied concurrently (while the server and partitions are active) if the release level and the last disruptive service pack level is the same.  From the example,  updating from HV950_092_045 to HV950_124_045 can be applied concurrently since the release is the same (950) and the last disruptive service pack level (045) is the same.  A successful concurrent apply and activate will be running with all the latest fixes that are concurrent.  There is no need to reboot any on the partition on the server.  The fix packs may contain deferred fixes and those require a server reboot to apply.  The security fix that went into HV950_124_045 is concurrent so once the service pack is activated successfully, the fix will have been applied.



    ------------------------------
    Pete Heyrman
    ------------------------------



  • 23.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 03:11 PM

    Hi Pete
    same valid for P10 that security patch will be activated concurrent ?
    thx vince



    ------------------------------
    Vincencio Michaelis
    ------------------------------



  • 24.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 03:13 PM

    I have the same question that Vincencio has.  Especially since Andres said otherwise.



    ------------------------------
    Robert Berendt IBMChampion
    ------------------------------



  • 25.  RE: CVE 2023-30438

    Posted Wed May 24, 2023 04:05 PM

    Yes, all the POWER9 and Power10 SPs for CVE 2023-30438 can be applied conncurrently from the last disruptive service pack.



    ------------------------------
    Pete Heyrman
    ------------------------------



  • 26.  RE: CVE 2023-30438

    IBM Champion
    Posted Wed May 24, 2023 04:37 PM

    To clarify , can be applied conncurrently from the last ACTIVATED disruptive service pack. So as not to confuse with deferred, 



    ------------------------------
    Alan Fulton Follow me on Twitter - @The_Iron_Monger
    Budd Lake
    2015329657
    ------------------------------



  • 27.  RE: CVE 2023-30438

    Posted Thu May 25, 2023 11:23 AM

    Sorry, to be sure, so as @Pete Heyrman said, and as I commented before from my side:

    We have firmware level at VH950_092_045 (FW950.30)
    Required firmware level is VH950_124_045 (71)

    Here is the HMC update information screenshot:

    And in FLRT we have



    And taking into account that:

    An installation is concurrent if: The release level (xxx) is the same, and The service pack level (yyy) currently installed on the system is the same or higher than the last disruptive service pack level (zzz) of the service pack to be installed. 
    (In our case 092 > 045)

    So if we update, this will be concurrent, NO need to reboot the server neither the LPARs, since the release is the same (950) and the last disruptive service pack level (045) is the same. is that correct?

    Thanks again team. I already opened a skill case to IBM Support, but still waiting for their reply on this.

    Gustavo





    ------------------------------
    Gustavo Orlando de Santis
    ------------------------------



  • 28.  RE: CVE 2023-30438

    IBM Champion
    Posted Fri May 26, 2023 03:11 AM

    Hi Gustavo,

    I did exactly the same update yesterday - from 950.30 to 950.71 and it worked online for me.



    ------------------------------
    Andrey Klyachkin

    https://www.power-devops.com
    ------------------------------