Wrong len in FUSEHeaderIn when sending write request with a payload
YaroslavLitvinov opened this issue · comments
Description
When sending FUSEHeaderIn request with a FUSE_WRITE data containing a payload, current implementation of fuse in gvisor incorrectly calculates FUSEHeaderIn.Len field, taking into account a fixed size of marshaled data structure only, which is - payload.SizeBytes(), and not using here size of actual payload, though sending it.
For instance: When writing 100 bytes file, the overall packet length should be 40 + 40 + 100, where 40 is the size of FUSEHeaderIn, 40 - FUSEWriteIn, 100 - file size, and not just 80 bytes as currently.
It looks to me that the test/fuse/linux/write_test.cc test is testing exactly this behavior.
This leads to errors related to receiving excessive data on server side when it attempts to read that data.
Point me if I am wrong.
Steps to reproduce
No response
runsc version
No response
docker version (if using docker)
No response
uname
No response
kubectl (if using Kubernetes)
No response
repo state (if built from source)
No response
runsc debug logs (if available)
No response
@avagin Hi again, I'm not sure if fuse support is any priority. But if so, what's the supposed write support next steps?
A friendly reminder that this issue had no activity for 120 days.
@avagin seems like you already have a fix for this? Do you want to rebase & merge that?
