BMCWeb payloads larger than than 64MB

[Joseph Reynolds]

On 8/5/20 6:26 PM, Joseph Reynolds wrote:
> On 8/3/20 4:09 PM, Joseph Reynolds wrote:
>> This is a reminder of the OpenBMC Security Working Group meeting 
>> scheduled for this Wednesday August 5 at 10:00am PDT.
>>
>> We'll discuss current development items, and anything else that comes 
>> up.

...snip...

>> 4. Is there interest in enhancing OpenBMC firmware image update 
>> uploads using the Redfish-specified multipart HTTP push updates (that 
>> is, support the MultipartHttpPushUri property?
> Sounds good, but nobody is working on it.  We also discussed use cases 
> for golden/primary/active/alternate images.
>
Ed,

You mentioned "For any payloads larger than 64MB, this stuff needs 
revisited" on Jul 22 in 
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/34972 and I would 
like to learn more about what you think the direction should be.

In the mean time it seems like the current design can tolerate a 128Mb 
payload.  Do you foresee additional problems other than we've already 
seen?  Examples: resource use, slow connections, and the attending 
security vulnerabilities.

- Joseph