Closed Bug 1093025 Opened 6 years ago Closed 5 years ago

Support separate listen and stream sockets for low-level I/O

Categories

(Firefox OS Graveyard :: GonkIntegration, defect)

All
Gonk (Firefox OS)
defect
Not set

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: tzimmermann, Assigned: tzimmermann)

References

Details

Attachments

(5 files)

Currently, the code around |UnixSocketConsumer| allows to listen for connection requests. When accepting the request, the I/O code closes the listen socket.

This behavior leads to race conditions in the connect/disconnect prodecure. We should have support for separate listen and stream sockets.
Apply patch to gaia, build, then flash, a "Certificate Test" app should appear in home screen.

"Import Certificates" will import a test certificate and ready to be read by keystore.
"Delete Certificates" will delete the imported certificate.
"Certificate List" will show full name of certificate used for keystore read command.

Use "adb shell keystore_cli g WIFI_SERVERCERT_ServerCert" to read certificate from keystore socket, the expected result should be:
-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
Assignee: nobody → tzimmermann
Status: NEW → ASSIGNED
Shawn,

it seems like Kyle is away for the rest of the month. Would you have time to review these I/O patches?
Flags: needinfo?(shuang)
I would like to help but it's better to ask IPC module owner :bent is a better choice.
Flags: needinfo?(shuang)
Ben, do you have time to review this patch set?
Flags: needinfo?(bent.mozilla)
Hi Thomas, unfortunately no, I'm pretty swamped at the moment :(
Flags: needinfo?(bent.mozilla)
OMFG you had to wait for me to get back from a 2 month vacation for this to get reviewed? D:

Well, anyways, I'm back. Does this need to be rebased before I start reviewing since it's been almost a month ( D: D: D: ), or can I get started now?
(In reply to Kyle Machulis (OUT OCT 1-NOV 30) [:kmachulis] [:qdot] (USE NEEDINFO?) from comment #10)
> OMFG you had to wait for me to get back from a 2 month vacation for this to
> get reviewed? D:

Don't worry. :) Right, it's only required for fixing some test cases.

> Well, anyways, I'm back. Does this need to be rebased before I start
> reviewing since it's been almost a month ( D: D: D: ), or can I get started
> now?

Just tested and the patch set applied cleanly.
Attachment #8517440 - Flags: review?(kyle) → review+
Attachment #8517441 - Flags: review?(kyle) → review+
Attachment #8517442 - Flags: review?(kyle) → review+
Attachment #8517443 - Flags: review?(kyle) → review+
You need to log in before you can comment on or make changes to this bug.