libsnapshot: chunk iov writes

Currently if our iov that we are trying to write is greater than 1024
our write will fail with error "INVALID ARGUMENT". This is because
pwritev() system call takes a max input size of IOV_MAX (which is device
dependant).

With our increased cache size of 1mb or maybe even more (or if user
configures batch size to be large), our write size could be greater than
IOV_MAX, and will fail with an unhelpful error. We should chunk these
writes to ensure they succeed.

Bug: 322279333
Test: cow_api_test + manual testing with large iov write sizes
Change-Id: Ia1fb53cbfc743cfcdfc7256ff9df075ad0c2dd38

1 files changed, 17 insertions, 3 deletions

diff --git a/fs_mgr/libsnapshot/libsnapshot_cow/writer_v3.cpp b/fs_mgr/libsnapshot/libsnapshot_cow/writer_v3.cpp
index ea1da4b53..8cc9964bf 100644
--- a/fs_mgr/libsnapshot/libsnapshot_cow/writer_v3.cpp
+++ b/fs_mgr/libsnapshot/libsnapshot_cow/writer_v3.cpp
@@ -717,13 +717,27 @@ bool CowWriterV3::WriteOperation(std::span<const CowOperationV3> ops,
         return false;
     }
     if (!data.empty()) {
-        const auto ret = pwritev(fd_, data.data(), data.size(), next_data_pos_);
-        if (ret != total_data_size) {
+        int total_written = 0;
+        int i = 0;
+        while (i < data.size()) {
+            int chunk = std::min(static_cast<int>(data.size() - i), IOV_MAX);
+
+            const auto ret = pwritev(fd_, data.data() + i, chunk, next_data_pos_ + total_written);
+            if (ret < 0) {
+                PLOG(ERROR) << "write failed chunk size of: " << chunk
+                            << " at offset: " << next_data_pos_ + total_written << " " << errno;
+                return false;
+            }
+            total_written += ret;
+            i += chunk;
+        }
+        if (total_written != total_data_size) {
             PLOG(ERROR) << "write failed for data of size: " << data.size()
-                        << " at offset: " << next_data_pos_ << " " << ret;
+                        << " at offset: " << next_data_pos_ << " " << errno;
             return false;
         }
     }
+
     header_.op_count += ops.size();
     next_data_pos_ += total_data_size;