From 1d2f2761cd1770bc040596e706335b74572e76de Mon Sep 17 00:00:00 2001
From: Yi Kong <yikong@google.com>
Date: Tue, 20 Jul 2021 16:55:39 +0800
Subject: profcollectd: Remove reports past retention period

Manually generated profcollect reports (through `profcollectctl report`)
and reports that failed to upload for whatever reason need to be
cleaned up periodically, to save disk space as well as to minimise
privacy leak surface.

Test: manual
Bug: 178561556
Change-Id: I54f09d1738f7b3c3b763251af83133bda5c214ae
Merged-In: I54f09d1738f7b3c3b763251af83133bda5c214ae
(cherry picked from commit 8dffc1208417a18870b33ce92c03a8f77da8dbee)
---
 profcollectd/libprofcollectd/config.rs  |  2 ++
 profcollectd/libprofcollectd/report.rs  | 11 ++++++++++-
 profcollectd/libprofcollectd/service.rs | 31 ++++++++++++++++++++++++++++---
 3 files changed, 40 insertions(+), 4 deletions(-)

diff --git a/profcollectd/libprofcollectd/config.rs b/profcollectd/libprofcollectd/config.rs
index c3ab45cf..11098fe1 100644
--- a/profcollectd/libprofcollectd/config.rs
+++ b/profcollectd/libprofcollectd/config.rs
@@ -29,6 +29,8 @@ use std::time::Duration;
 const PROFCOLLECT_CONFIG_NAMESPACE: &str = "profcollect_native_boot";
 const PROFCOLLECT_NODE_ID_PROPERTY: &str = "persist.profcollectd.node_id";
 
+pub const REPORT_RETENTION_SECS: u64 = 14 * 24 * 60 * 60; // 14 days.
+
 lazy_static! {
     pub static ref TRACE_OUTPUT_DIR: &'static Path = Path::new("/data/misc/profcollectd/trace/");
     pub static ref PROFILE_OUTPUT_DIR: &'static Path = Path::new("/data/misc/profcollectd/output/");
diff --git a/profcollectd/libprofcollectd/report.rs b/profcollectd/libprofcollectd/report.rs
index c37993b7..a67d500b 100644
--- a/profcollectd/libprofcollectd/report.rs
+++ b/profcollectd/libprofcollectd/report.rs
@@ -23,7 +23,7 @@ use std::fs::{self, File, Permissions};
 use std::io::{Read, Write};
 use std::os::unix::fs::PermissionsExt;
 use std::path::{Path, PathBuf};
-use std::time::SystemTime;
+use std::time::{Duration, SystemTime};
 use uuid::v1::{Context, Timestamp};
 use uuid::Uuid;
 use zip::write::FileOptions;
@@ -82,3 +82,12 @@ fn get_report_filename(node_id: &MacAddr6) -> Result<String> {
     let uuid = Uuid::new_v1(ts, &node_id.as_bytes())?;
     Ok(uuid.to_string())
 }
+
+/// Get report creation timestamp through its filename (version 1 UUID).
+pub fn get_report_ts(filename: &str) -> Result<SystemTime> {
+    let uuid_ts = Uuid::parse_str(filename)?
+        .to_timestamp()
+        .ok_or_else(|| anyhow!("filename is not a valid V1 UUID."))?
+        .to_unix();
+    Ok(SystemTime::UNIX_EPOCH + Duration::new(uuid_ts.0, uuid_ts.1))
+}
diff --git a/profcollectd/libprofcollectd/service.rs b/profcollectd/libprofcollectd/service.rs
index 0b9cca1e..04f30f89 100644
--- a/profcollectd/libprofcollectd/service.rs
+++ b/profcollectd/libprofcollectd/service.rs
@@ -21,16 +21,17 @@ use binder::public_api::Result as BinderResult;
 use binder::Status;
 use profcollectd_aidl_interface::aidl::com::android::server::profcollect::IProfCollectd::IProfCollectd;
 use std::ffi::CString;
-use std::fs::{copy, create_dir, read_to_string, remove_dir_all, remove_file, write};
+use std::fs::{copy, create_dir, read_dir, read_to_string, remove_dir_all, remove_file, write};
 use std::path::PathBuf;
 use std::str::FromStr;
 use std::sync::{Mutex, MutexGuard};
+use std::time::Duration;
 
 use crate::config::{
     Config, BETTERBUG_CACHE_DIR_PREFIX, BETTERBUG_CACHE_DIR_SUFFIX, CONFIG_FILE,
-    PROFILE_OUTPUT_DIR, REPORT_OUTPUT_DIR, TRACE_OUTPUT_DIR,
+    PROFILE_OUTPUT_DIR, REPORT_OUTPUT_DIR, REPORT_RETENTION_SECS, TRACE_OUTPUT_DIR,
 };
-use crate::report::pack_report;
+use crate::report::{get_report_ts, pack_report};
 use crate::scheduler::Scheduler;
 
 fn err_to_binder_status(msg: Error) -> Status {
@@ -155,6 +156,30 @@ impl ProfcollectdBinderService {
             write(*CONFIG_FILE, &new_config.to_string())?;
         }
 
+        // Clear profile reports out of rentention period.
+        for report in read_dir(*REPORT_OUTPUT_DIR)? {
+            let report = report?.path();
+            let report_name = report
+                .file_stem()
+                .and_then(|f| f.to_str())
+                .ok_or_else(|| anyhow!("Malformed path {}", report.display()))?;
+            let report_ts = get_report_ts(report_name);
+            if let Err(e) = report_ts {
+                log::error!(
+                    "Cannot decode creation timestamp for report {}, caused by {}, deleting",
+                    report_name,
+                    e
+                );
+                remove_file(report)?;
+                continue;
+            }
+            let report_age = report_ts.unwrap().elapsed()?;
+            if report_age > Duration::from_secs(REPORT_RETENTION_SECS) {
+                log::info!("Report {} past rentention period, deleting", report_name);
+                remove_file(report)?;
+            }
+        }
+
         Ok(ProfcollectdBinderService {
             lock: Mutex::new(Lock { scheduler: new_scheduler, config: new_config }),
         })
-- 
cgit v1.2.3