On Tue, 7 Apr 2026 09:57:36 GMT, Kevin Walls <[email protected]> wrote:
> Unfortunate to have two of the same HWCAP_PACA ifndef and comment with URL.
> Though maybe there's no convenient common include file to make this
> definition.
Yes.
We can add new header file to define `HWCAP_PACA` if need of course, but I feel
it is ugly a bit...
> Is this "We can use the caller frame if it is a signal trampoline." ?
Yes.
Signal trampoline is special frame, so we can set sender PC/SP/FP straightly
without DWARF.
> Looks like we do mean to have this check twice here in LinuxAARCH64CFrame.
As I commented, vdso.so on AArch64 Linux might not have `.eh_frame`, then DWARF
parser would fail. Thus I added signal trampoline check in L160.
> src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/debugger/linux/amd64/LinuxAMD64CFrame.java
> line 125:
>
>> 123: : null;
>> 124: } catch (DebuggerException e) {
>> 125: if (linuxDbg().isSignalTrampoline(senderPC)) {
>
> This is the other comment that might be "We can use the caller frame...".
> Do we want the same check above at line 110 when we can't create senderDwarf?
As I wrote in above comment, signal trampoline check you pointed is for lacking
of `.eh_frame` in vdso.so. AFAICS vdso64.so on AMD64 Linux has `.eh_frame` -
Fedora, Ubuntu, Rocky at least. Thus I didn't add the check in
LinuxAMD64CFrame. I'm not sure vDSO in all of Linux distros (kernel
configuration in precsise) is that, but I guess most of are so because major 2
families (Red Hat, Ubuntu) is so.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/29731#issuecomment-4198931724
PR Review Comment: https://git.openjdk.org/jdk/pull/29731#discussion_r3044957017
PR Review Comment: https://git.openjdk.org/jdk/pull/29731#discussion_r3044957436
