[libc][fcntl] fix -Wshorten-64-to-32 for 32b ARM

[nickdesaulniers]

Fixes:

llvm-project/libc/src/__support/OSUtil/linux/fcntl.cpp:63:26: error:
implicit conversion loses integer precision: '__off64_t' (aka 'long long')
to '__off_t' (aka 'long') [-Werror,-Wshorten-64-to-32]
    flk->l_start = flk64.l_start;
                 ~ ~~~~~~^~~~~~~
llvm-project/libc/src/__support/OSUtil/linux/fcntl.cpp:64:24: error:
implicit conversion loses integer precision: '__off64_t' (aka 'long long')
to '__off_t' (aka 'long') [-Werror,-Wshorten-64-to-32]
    flk->l_len = flk64.l_len;
               ~ ~~~~~~^~~~~

We already have an overflow check, just need the cast to be explicit. This
warning was observed on the 32b ARM build in overlay mode.

[llvmbot]

@llvm/pr-subscribers-libc

Author: Nick Desaulniers (paternity leave) (nickdesaulniers)

Changes

Fixes:
llvm-project/libc/src/__support/OSUtil/linux/fcntl.cpp:63:26: error:
implicit conversion loses integer precision: '__off64_t' (aka 'long long')
to '__off_t' (aka 'long') [-Werror,-Wshorten-64-to-32]
flk->l_start = flk64.l_start;
~ ~~~~~~^~~~~~~
llvm-project/libc/src/__support/OSUtil/linux/fcntl.cpp:64:24: error:
implicit conversion loses integer precision: '__off64_t' (aka 'long long')
to '__off_t' (aka 'long') [-Werror,-Wshorten-64-to-32]
flk->l_len = flk64.l_len;
~ ~~~~~~^~~~~

We already have an overflow check, just need the cast to be explicit. This
warning was observed on the 32b ARM build in overlay mode.

---

Full diff: https://github.com/llvm/llvm-project/pull/95945.diff

1 Files Affected:

• (modified) libc/src/__support/OSUtil/linux/fcntl.cpp (+2-2)

diff --git a/libc/src/__support/OSUtil/linux/fcntl.cpp b/libc/src/__support/OSUtil/linux/fcntl.cpp
index 7dc416a7916df..b087f898c395d 100644
--- a/libc/src/__support/OSUtil/linux/fcntl.cpp
+++ b/libc/src/__support/OSUtil/linux/fcntl.cpp
@@ -60,8 +60,8 @@ int fcntl(int fd, int cmd, void *arg) {
     // Now copy back into flk, in case flk64 got modified
     flk->l_type = flk64.l_type;
     flk->l_whence = flk64.l_whence;
-    flk->l_start = flk64.l_start;
-    flk->l_len = flk64.l_len;
+    flk->l_start = static_cast<decltype(flk->l_start)>(flk64.l_start);
+    flk->l_len = static_cast<decltype(flk->l_len)>(flk64.l_len);
     flk->l_pid = flk64.l_pid;
     return retVal;
   }

[nickdesaulniers]

cc @simonzgx

[simonzgx]

Got, thanks a lot.

[SchrodingerZhu]

This patch LGTM.

But @lntue mentioned there may be other problems.

[lntue]

This patch LGTM.

But @lntue mentioned there may be other problems.

The fix for this part is fine for me. The other problem in the same function that I have question is

if (static_cast<unsigned long>(retVal) <= -4096UL)
      return fex.type == F_OWNER_PGRP ? -fex.pid : fex.pid;

which will treat erroneous condition retval < -4096 the same as ok condition retval >= 0. We should continue the discussion of that one in #95570.