Move away from default password
There is some interest in moving OpenBMC away from a default password.
- email: https://lists.ozlabs.org/pipermail/openbmc/2019-June/016515.html (which references a RestrictionMode design: https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21195)
Having a default password is a security risk. Note that changing to a different default password is not a good solution. For example, if a bad actor learns the default password from one device, that actor will likely know the password for many of them.
I am exploring two options for OpenBMC, and want your feedback.
1. Unique password per BMC.
In this approach, there is a way to change the factory default password.
Example flow: assemble the BMC, test it, factory reset, generate unique
password (such as pwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code. Presumably the new factory default would be printed on a sticker, or something.
Are there any other factory settings (settings unique to each device) that would benefit from being set like this?
One downside to this approach is someone who orders 100 systems has to enter 100 unique passwords.
2. Create new account on first access.
Specifically, default OpenBMC to use a new RestrictionMode=SetupAccess which:
- (A) requires users to set up their credentials (at least: remove the
default password) before they can leave that mode, and
- (B) does not let the user do anything else, including other
provisioning or operating the host, while in this mode.
So we could manufacture the BMC with a default password, but require it to be changed as the first step to provision the BMC.
We might want to make network services react to this new mode, for example, the phosphor-webui GUI would likely want to handle this specially, and most REST APIs would be restricted. Note this approach gives an attacker a window of time before the credentials are set up.
I believe both of these approaches represent reasonable security practices which may satisfy laws regarding consumer products. For example, CA Law SB-327 https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB327
Are there guidelines we can follow? Can you find additional vulnerabilities with these approaches? Can we do better than this without requiring additional infrastructure?
I plan to discuss this at the 2019-06-26 Security working group meeting: https://docs.google.com/document/d/1b7x9BaxsfcukQDqbvZsU2ehMq4xoJRQvLxxsDUWmAOI/
and would be happy to see ideas here.
- Joseph
Joseph,
#1 is not yet covered in any docs yet, but https://github.com/openbmc/docs/blob/master/user_management.md#deployment---out-of-factory this covers the out of factory deployment.
Regards,
Richard
Quoted 31 lines · Joseph Reynolds
There is some interest in moving OpenBMC away from a default password.
- email:
https://lists.ozlabs.org/pipermail/openbmc/2019-June/016515.html (which references a RestrictionMode design: https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21195)Having a default password is a security risk. Note that changing to a different default password is not a good solution. For example, if a bad actor learns the default password from one device, that actor will likely know the password for many of them.
I am exploring two options for OpenBMC, and want your feedback.
1. Unique password per BMC.
In this approach, there is a way to change the factory default password. Example flow: assemble the BMC, test it, factory reset, generate unique password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code. Presumably the new factory default would be printed on a sticker, or something.
Are there any other factory settings (settings unique to each device) that would benefit from being set like this?
One downside to this approach is someone who orders 100 systems has to enter 100 unique passwords.2. Create new account on first access.
Specifically, default OpenBMC to use a new RestrictionMode=SetupAccess which:
- (A) requires users to set up their credentials (at least: remove the default password) before they can leave that mode, and
- (B) does not let the user do anything else, including other provisioning or operating the host, while in this mode.
So we could manufacture the BMC with a default password, but require it to be changed as the first step to provision the BMC.
We might want to make network services react to this new mode, for example, the phosphor-webui GUI would likely want to handle this specially, and most REST APIs would be restricted. Note this approach gives an attacker a window of time before the credentials are set up.I believe both of these approaches represent reasonable security practices which may satisfy laws regarding consumer products. For example, CA Law SB-327 https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB327
Are there guidelines we can follow? Can you find additional vulnerabilities with these approaches? Can we do better than this without requiring additional infrastructure?
I plan to discuss this at the 2019-06-26 Security working group meeting: https://docs.google.com/document/d/1b7x9BaxsfcukQDqbvZsU2ehMq4xoJRQvLxxsDUWmAOI/
and would be happy to see ideas here.- Joseph
Quoted 3 lines
Quoted 3 lines
1. Unique password per BMC.
In this approach, there is a way to change the factory default password. Example flow: assemble the BMC, test it, factory reset, generate unique password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code.
How would this new "factory settings" flash partition be protected against being modified by an unauthorized or malicious user?
Quoted 4 lines
Quoted 4 lines
Presumably the new factory default would be printed on a sticker, or something.
Are there any other factory settings (settings unique to each device) that would benefit from being set like this? One downside to this approach is someone who orders 100 systems has to enter 100 unique passwords.
Quoted 11 lines · Adriana Kobylak
Quoted 8 lines
Quoted 8 lines
1. Unique password per BMC.
In this approach, there is a way to change the factory default password. Example flow: assemble the BMC, test it, factory reset, generate unique password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code.How would this new "factory settings" flash partition be protected against being modified by an unauthorized or malicious user?
My guess would be it'd be protected the same way that the default password is today: not at all. If an attacker can write to flash, the only way to reset the box is to dediprog the BMC flash chip.
Signature · 2 lines
Stewart Smith OPAL Architect, IBM.
Quoted 10 lines · Stewart Smith
Quoted 9 lines · Adriana Kobylak
Quoted 8 lines
Quoted 8 lines
1. Unique password per BMC.
In this approach, there is a way to change the factory default password. Example flow: assemble the BMC, test it, factory reset, generate unique password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code.
Quoted 6 lines
Quoted 2 lines
How would this new "factory settings" flash partition be protected against being modified by an unauthorized or malicious user?
My guess would be it'd be protected the same way that the default password is today: not at all. If an attacker can write to flash, the only way to reset the box is to dediprog the BMC flash chip.
Access to the flash would be protected from attack by network agents via password access to the BMC at two critical points.
In this scenario:
1. The factory assembles and tests the BMC, then changes its password to a new value. The password hash is stored on the flash "factory settings" partition. The BMC is then shipped to its new owner with the new password. At this point, only the owner has password access to the BMC (unless the factory keeps a record of the new password).
2. The owner installs the BMC and configures it, including its network. For example, change the password, creates new accounts, and set up IP.
At this point, only the owner and owner's agents have password access to the BMC.
At this point, one of the owner's agents could use ssh to access the flash partition. (But why would they need to?)
Quoted 10 lines · Joseph Reynolds
There is some interest in moving OpenBMC away from a default password.
- email:
https://lists.ozlabs.org/pipermail/openbmc/2019-June/016515.html (which references a RestrictionMode design:
https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21195)Having a default password is a security risk. Note that changing to a different default password is not a good solution. For example, if a bad actor learns the default password from one device, that actor will likely know the password for many of them.
and it makes it pretty easy to use something like Shodan to find all the possible OpenBMCs connected to the Internet (hopefully by accident) and pop a root shell on them.
Mind you, in a lab environment, it's *really* useful.
Quoted 16 lines
I am exploring two options for OpenBMC, and want your feedback.
1. Unique password per BMC.
In this approach, there is a way to change the factory default password.
Example flow: assemble the BMC, test it, factory reset, generate unique
password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code.
Presumably the new factory default would be printed on a sticker, or something.
Are there any other factory settings (settings unique to each device) that would benefit from being set like this?
One downside to this approach is someone who orders 100 systems has to enter 100 unique passwords.
There's also a downside for those of us who often work remotely from machines, we may have to wait for someone to be awake and be able to physically go and check what the default password is.
Quoted 6 lines
2. Create new account on first access.
Specifically, default OpenBMC to use a new RestrictionMode=SetupAccess which:
- (A) requires users to set up their credentials (at least: remove the
default password) before they can leave that mode, and
- (B) does not let the user do anything else, including other
Quoted 8 lines
provisioning or operating the host, while in this mode.
So we could manufacture the BMC with a default password, but require it to be changed as the first step to provision the BMC.
We might want to make network services react to this new mode, for example, the phosphor-webui GUI would likely want to handle this specially, and most REST APIs would be restricted.
Note this approach gives an attacker a window of time before the credentials are set up.
In a lab environment though, we'd have to ensure we had a *very* reliably way to reset the BMC when we switched who was using the machine.
Quoted 6 lines
I believe both of these approaches represent reasonable security practices which may satisfy laws regarding consumer products. For example, CA Law SB-327
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB327
Quoted 7 lines
Are there guidelines we can follow? Can you find additional vulnerabilities with these approaches? Can we do better than this without requiring additional infrastructure?
I plan to discuss this at the 2019-06-26 Security working group meeting:
https://docs.google.com/document/d/1b7x9BaxsfcukQDqbvZsU2ehMq4xoJRQvLxxsDUWmAOI/
and would be happy to see ideas here.
I had an idea that provides less security, but still may be valuable:
make the default password at manufacturing be linked to the MAC address of the BMC. This prevents people not on a local network to the machine from gaining access (e.g. I have no idea what the MAC address of 8.8.8.8 is), but if I'm on the same ethernet network, then I can still work it out. It would also allow someone buying 100 machines to programatically work out what the password should be.
Signature · 2 lines
Stewart Smith OPAL Architect, IBM.
Quoted 12 lines · Stewart Smith
Quoted 10 lines · Joseph Reynolds
There is some interest in moving OpenBMC away from a default password.
- email:
https://lists.ozlabs.org/pipermail/openbmc/2019-June/016515.html (which references a RestrictionMode design:
https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21195)Having a default password is a security risk. Note that changing to a different default password is not a good solution. For example, if a bad actor learns the default password from one device, that actor will likely know the password for many of them.
Quoted 5 lines
and it makes it pretty easy to use something like Shodan to find all the possible OpenBMCs connected to the Internet (hopefully by accident) and pop a root shell on them.
Mind you, in a lab environment, it's *really* useful.
I imagine for the forseeable future, OpenBMC would continue to have a default userid and password (and I hope each development lab is using a different default password than the well-known password). But I think development labs are subject to attack, so we need to eventually move away from default passwords even in the development labs.
At this time, I am looking for options to move away from this model, but do not anticipate changing the default.
Quoted 38 lines
Quoted 17 lines
I am exploring two options for OpenBMC, and want your feedback.
1. Unique password per BMC.
In this approach, there is a way to change the factory default password.
Example flow: assemble the BMC, test it, factory reset, generate
unique
password (such aspwgen), then use a new function “save factory default settings” which would save the current setting into a new “factory settings” flash partition. After that, a factory reset would reset to the factory installed password, not to the setting in the source code.
Presumably the new factory default would be printed on a sticker, or something.
Are there any other factory settings (settings unique to each device) that would benefit from being set like this?
One downside to this approach is someone who orders 100 systems has to enter 100 unique passwords.There's also a downside for those of us who often work remotely from machines, we may have to wait for someone to be awake and be able to physically go and check what the default password is.
Quoted 7 lines
2. Create new account on first access.
Specifically, default OpenBMC to use a new RestrictionMode=SetupAccess which:
- (A) requires users to set up their credentials (at least: remove
the
default password) before they can leave that mode, and
- (B) does not let the user do anything else, including otherQuoted 8 lines
provisioning or operating the host, while in this mode.
So we could manufacture the BMC with a default password, but require it to be changed as the first step to provision the BMC.
We might want to make network services react to this new mode, for example, the phosphor-webui GUI would likely want to handle this specially, and most REST APIs would be restricted.
Note this approach gives an attacker a window of time before the credentials are set up.
Quoted 2 lines
In a lab environment though, we'd have to ensure we had a *very* reliably way to reset the BMC when we switched who was using the machine.
Agreed. Gaining initial access and recovering access the BMC is one of the issues.
Quoted 13 lines
Quoted 5 lines
I believe both of these approaches represent reasonable security practices which may satisfy laws regarding consumer products. For example, CA Law SB-327
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB327Quoted 7 lines
Are there guidelines we can follow? Can you find additional vulnerabilities with these approaches? Can we do better than this without requiring additional infrastructure?
I plan to discuss this at the 2019-06-26 Security working group meeting:
https://docs.google.com/document/d/1b7x9BaxsfcukQDqbvZsU2ehMq4xoJRQvLxxsDUWmAOI/
and would be happy to see ideas here.
Quoted 7 lines
I had an idea that provides less security, but still may be valuable:
make the default password at manufacturing be linked to the MAC address of the BMC. This prevents people not on a local network to the machine from gaining access (e.g. I have no idea what the MAC address of 8.8.8.8 is), but if I'm on the same ethernet network, then I can still work it out. It would also allow someone buying 100 machines to programatically work out what the password should be.
Thanks, I understand this idea. It may satisfy the letter of the law, and perhaps also its intent (disclaimer: not a legal opinion), so it has some merit. But this scheme not as secure as random passwords (specifically, if the attacker learns the MAC addresses).
- Joseph
Quoted 21 lines · Joseph Reynolds
Quoted 13 lines · Stewart Smith
Quoted 11 lines · Joseph Reynolds
There is some interest in moving OpenBMC away from a default password.
- email:
https://lists.ozlabs.org/pipermail/openbmc/2019-June/016515.html
(which
references a RestrictionMode design:
https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21195)Having a default password is a security risk. Note that changing to a different default password is not a good solution. For example, if a bad actor learns the default password from one device, that actor will likely know the password for many of them.
Quoted 6 lines
and it makes it pretty easy to use something like Shodan to find all the
possible OpenBMCs connected to the Internet (hopefully by accident) and pop a root shell on them.Mind you, in a lab environment, it's *really* useful.
Quoted 8 lines
I imagine for the forseeable future, OpenBMC would continue to have a default userid and password (and I hope each development lab is using a different default password than the well-known password). But I think development labs are subject to attack, so we need to eventually move away from default passwords even in the development labs.
At this time, I am looking for options to move away from this model, but do not anticipate changing the default.
I admire your optimism :)
I could probably ruin everybody's day with a simple nmap invocation and for loop in shell across the IBM class A :)
Having something that wasn't the same everywhere would probably be an improvement in the lab, and make it harder for someone to accidentally do something to a machine they didn't intend to.
Signature · 2 lines
Stewart Smith OPAL Architect, IBM.