executor: increase coverage buffer size

The coverage buffer frequently overflows.
We cannot increase it radically b/c they consume lots of memory
(num procs x num kcovs x buffer size) and lead to OOM kills
(at least with 8 procs and 2GB KASAN VM).
So increase it 2x and slightly reduce number of threads/kcov descriptors.
However, in snapshot mode we can be more aggressive (only 1 proc).

This reduces number of overflows by ~~2-4x depending on syscall.

1 files changed, 8 insertions, 1 deletions

diff --git a/executor/executor_linux.h b/executor/executor_linux.h
index f2e9d4df6..e952e6ea9 100644
--- a/executor/executor_linux.h
+++ b/executor/executor_linux.h
@@ -13,6 +13,12 @@
 
 static bool pkeys_enabled;
 
+// The coverage buffer can realistically overflow. In the non-snapshot mode we cannot afford
+// very large buffer b/c there are usually multiple procs, and each of them consumes
+// significant amount of memory. In snapshot mode we have only one proc, so we can have
+// larger coverage buffer.
+const int kSnapshotCoverSize = 1024 << 10;
+
 const unsigned long KCOV_TRACE_PC = 0;
 const unsigned long KCOV_TRACE_CMP = 1;
 
@@ -101,7 +107,8 @@ static void cover_open(cover_t* cov, bool extra)
 		failmsg("filed to dup cover fd", "from=%d, to=%d", fd, cov->fd);
 	close(fd);
 	const int kcov_init_trace = is_kernel_64_bit ? KCOV_INIT_TRACE64 : KCOV_INIT_TRACE32;
-	const int cover_size = extra ? kExtraCoverSize : kCoverSize;
+	const int cover_size = extra ? kExtraCoverSize : flag_snapshot ? kSnapshotCoverSize
+								       : kCoverSize;
 	if (ioctl(cov->fd, kcov_init_trace, cover_size))
 		fail("cover init trace write failed");
 	cov->mmap_alloc_size = cover_size * (is_kernel_64_bit ? 8 : 4);