<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/140610>140610</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [lldb] MachProcess::GetPlatform fails to detect platform in macOS 15.4
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            new issue
      </td>
    </tr>

    <tr>
      <th>Assignees</th>
      <td>
      </td>
    </tr>

    <tr>
      <th>Reporter</th>
      <td>
          talkeren
      </td>
    </tr>
</table>

<pre>
    Hey, there is an issue to debug some binaries since the latest mac version, so I investigated it and found part of the root cause.

It should be reproducible with Chromium's test binaries.
```
gn gen out/Default
ninja -C out/Default unit_tests
```

And than this will fail
```
(lldb) target create out/Default/unit_tests
Current executable set to '/Users/tkeren/chromium/src/out/Default/unit_tests' (arm64).
(lldb) process launch
error: process resume at entry point failed: Resume timed out.
Process 54485 exited with status = -1 (0xffffffff) lost connection
```

The actual failure happens because `qProcessInfo` fails due to a timeout. This MachO binary has 500+ load commands, for each one `MachProcess::GetMachOInformationFromMemory` calls `MachProcess::GetDeploymentInfo` which calls `MachProcess::GetPlatform`. 

`MachProcess::GetPlatform` cache the result in m_platform, but it just check if the value is not zero, if `MachProcess::GetProcessPlatformViaDYLDSPI` fails to return a valid value, it will call it every time, which is what happens now. This function is slow - it takes about ~60ms for each call. so 60ms * 500 * 2 - we are talking about a minute to handle that command, which causes timeout.

I didn't investigate why `m_dyld_process_info_get_platform` returns 0 (which probably start to happen in macOS 15.4 because this is when the issue started), but I do have a fix for this issue which is just caching the result wether it is valid or not.
</pre>
<img width="1" height="1" alt="" src="http://email.email.llvm.org/o/eJyEVV-PWrcT_TTmZbTIGC4sDzyQRfx-KzVK1KSV-rSaaw9cJ742tcew9KGfvbLvZTebKulqJQSeP8fnnBljSvboiTaieSea3QQzdyFuGN1XiuQnbTDXzf_pKtQDcEeRwCZADzalTMABDLX5CCn0BK31GC0lSNZrKuHgkCkx9KjhTDHZ4EuhFOARrD9TYntEJgOWAb2BQ8jewAkjQzjUAjEEBo050VTIrZDbR4bUhewMtASRTjGYrG3rCC6WO3joYuht7oVaJai9b6hq_lKO_3J79HAkDyGzUPsdHTA7FnLrrf-CcPfw9gCyt_xU6qXvygi53XoD3KEH7myCi3UODmjd94Hq3jnTCrUGxngkBh0Jmb5DoPZvWj3kGMkz0DPpzFjumYgL8UKthNr_ligmofZc5RJqr18I2Keohdr_pLxagVD3GPvlQqj19C3IUwyaUgKH2etOyC3FGKKYb19OIqXcEyADeY5XOAXruV6dTIn7dThn25MptywNPo65zWJx3wA926J-VS4xck4g5ju4mxVc8vkw_hU4LiQGHbwnzcVG_1Lhc0eAmjMO7OdI0OHpRD5BS9VBIJbyzxHAoz8EsZQ1NIEZvIwVa0EKn4uU71F3HwYDXaHDBI2UQr0DF9CADn2P3qRi6EOIQKg7CL52KYljIzHfivn2f8S1WGkbeyw32MfQv6c-xGvBodG59KPUHZ1cuPbk-Qb70lnd_Tzpo0MuzcRSTmGg6L8jQaPuhtEt6joG66F_Ot0i1AO0mcu4fslFj470V7DDqJ7R5boefGD4i2Io0fbwQ3zD91vz3y3u_vhl9-nj46ssHCAS5-gBS3Vrhh61Lg-DVigoX-hM8VrlK6cDPWUYO-QXG_hwGYU9ZF9dVEKSCxe4KzUYv1ICbENm-Hsp-_Sqa2kzLXur_izUtlihfiq4gwsBRoKyNK0_jhUQeuszV2N16I0rtCLffPMKs3ozvVhvXHNgrPFCrfjbRQmX7lr47J_M1ZmncRCfrD-EpyPxq05LOTKXoMC8HzqdYmixddcya5EHYIWaKjLqD59g1kwXL-NS11klkXxVeFj6NZmMUOubHR7BlFJnAoSDfa60jckl4UWNwTOou8LSNx67UHlaigQ2jUKHWGw0nZjN3Kzna5zQZrZarJb3zbJpJt1mtZJ0r8nog1qhVrRqG7PEmV7N7_VyJnFiN0qqRjaztZKzxayZqhWt9Uyu1RxRts1KLCT1aN3UuXM_DfE4qWg3s4VczuTEYUsu1WdRKU-X4S5CqfJKxk1JumvzMYmFdDZxei3Dll19T-smbXbws5l7dbohJs1wk_CtJpMc3aZjPtUKai_U_mi5y-1Uh7LqS_Px4-4UwxfSZdVXyOV1GO903qh_AgAA__8YVqoF">