Give test-qnn-python-imports-linux 30 minutes instead of 15 (#24092)

## Problem

`test-qnn-python-imports-linux` has a 15 minute timeout, the shortest in
`pull.yml`. That limit becomes the job's `timeout-minutes`, so it also
covers pulling the large QNN image, checking out submodules and the
cache setup steps, not only the test script.

Those setup steps vary from run to run. Over the last day, successful
runs took 9 to 15 minutes, often over 13. Runs that went past 15 were
killed. 5 of the last 14 runs on main ended this way, with the script
itself taking its usual 7 to 8 minutes. When the job hits the limit,
GitHub cancels it, and the `pull` run on that commit ends as cancelled
instead of green.

## Fix

Raise the limit to 30 minutes, the smallest value already used in this
workflow. A slow runner then no longer fails the job, and a real hang is
still caught.

## Testing

No local test, since this only changes a CI limit. The timings above
come from the job and step durations of recent runs.

Author: shoumikhin

.github/workflows/pull.yml

@@ -1343,7 +1343,7 @@ jobs:
       docker-image: ${{ needs.docker-image.outputs.docker-registry }}/ci-image:executorch-ubuntu-22.04-qnn-sdk-${{ needs.docker-image.outputs.ci-docker-hash }}
       submodules: 'recursive'
       ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-      timeout: 15
+      timeout: 30
       script: |
         CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
         conda activate "${CONDA_ENV}"