RFR(XS): 8248878: SA: Implement simple workaround for JDK-8248876

[Yasumasa Suenaga]

Hi Chris,

Your change looks good.


BTW I saw JDK-8248876. I'm not sure, but I guess we can fix this issue if we allow PT_INTERP in L118:

```
105 uintptr_t find_base_address(int fd, ELF_EHDR* ehdr) {
                   :
115   // the base address of a shared object is the lowest vaddr of
116   // its loadable segments (PT_LOAD)
117   for (phdr = phbuf, cnt = 0; cnt < ehdr->e_phnum; cnt++, phdr++) {
118     if (phdr->p_type == PT_LOAD && phdr->p_vaddr < baseaddr) {
119       baseaddr = phdr->p_vaddr;
120     }
121   }
```

/proc/<PID>/maps shows top of `java` is 0x56543b9df000:

56543b9df000-56543b9e0000 r--p 00000000 08:10 55770                      /usr/lib/jvm/java-11-openjdk-amd64/bin/java


`i target` on GDB shows 0x56543b9df000 is .interp section:

Local exec file:
         `/usr/lib/jvm/java-11-openjdk-amd64/bin/java', file type elf64-x86-64.
         Entry point: 0x56543b9e0330
         0x000056543b9df318 - 0x000056543b9df334 is .interp


Thanks,

Yasumasa


On 2020/07/07 13:18, Chris Plummer wrote:
> Hello,
> 
> Please help review the following:
> 
> http://cr.openjdk.java.net/~cjplummer/8248878/webrev.00/index.html
> https://bugs.openjdk.java.net/browse/JDK-8248878
> 
> The explanation of the fix is in the CR. The parent CR, JDK-8248876 [1], explains the issue being addressed.
> 
> There's no test for this fix yet. It requires the changes I'm making for JDK-8247514 [2], which include changes to "findpc" support and the ClhsdbFindPC.java test that trigger this issue.
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8248876
> [2] https://bugs.openjdk.java.net/browse/JDK-8247514
> 
> thanks,
> 
> Chris