-
Notifications
You must be signed in to change notification settings - Fork 3.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Upgrade grpc netty due to security risks #8077
Comments
Not sure.. we are using grpc-netty-shaded and it seems to bring in io.netty:netty-codec-http:4.1.52.Final ? |
netty will be upgraded to "one" version in the near future based on all such requirements. You talked about upgrading to 4.1.53.Final but it could be 4.1.60.Final so it is best to track this one issue and you can also add your comments in that one issue. |
Yes, 4.1.60.Final will be good. |
It is unlikely to be a patch release like 1.37.1 and also I don't know how soon this will be done. CC @ejona86 who can throw some light on this |
gRPC is not impacted by that CVE, so there's no immediate need to update Netty within grpc-netty-shaded. If you are using Netty elsewhere in your application and are using grpc-netty, then there is the need to let you upgrade your application to a newer version to avoid the CVE in your application (because you could be using multipart). To that end grpc-java 1.35.1, 1.36.1, and 1.37.0 are compatible with Netty 4.1.60.Final. Netty 4.1.60 was incompatible with some Bazel environments that build tcnative and netty-transport-epoll, so we'll need to upgrade to Netty 4.1.61+. That work is scheduled to happen soon, but the multipart CVE has no real impact on that. |
Ok.. we only depend on grpc-netty-shaded. "Netty contains a flaw in the AbstractDiskHttpData.delete() function in handler/codec/http/multipart/AbstractDiskHttpData.java that is triggered as temporary file entries are added to the 'DeleteOnExitHook' object but not properly removed when processing POST requests that are 16 kB. This may allow a remote attacker to exhaust available memory resources, potentially resulting in a denial of service." |
Yes. gRPC does not process multipart messages and does not use AbstractDiskHttpData. |
If you use grpc-netty-shaded, you shouldn't need a dependency on io.netty:netty-codec-http. You may investigate how that dependency is being introduced into your build. |
We only have a dependency to grpc-netty-shaded. |
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was brought up in grpc#8077.
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was noticed in in grpc#8077.
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was noticed in in #8077.
With #8165 removing the pom.properties, it seems there's no further issues here. |
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was noticed in in grpc#8077.
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was noticed in in grpc#8077.
The pom.properties are apparently present to allow tooling to know what Maven artifact cooresponds to a JAR, just by looking at the JAR. Since we shade Netty, that produces inaccurate results. This was noticed in in #8077.
Hi!
In grpc 1.37.0 it seems like there is a security issue as below.
io.netty:netty-codec-http:4.1.52.Final:
"Netty contains a flaw in the AbstractDiskHttpData.delete() function in handler/codec/http/multipart/AbstractDiskHttpData.java that is triggered as temporary
file entries are added to the 'DeleteOnExitHook' object but not properly removed when processing POST requests that are 16 kB. This may allow a remote attacker
to exhaust available memory resources, potentially resulting in a denial of service."
Do you have any plan for upgrade grpc-netty (or grpc-netty-shaded) to a new version for netty-codec-http:4.1.53.Final?
The text was updated successfully, but these errors were encountered: