8361712: Improve ShenandoahAsserts printing

[tstuefe]

Small changes to ShenandoahAsserts and friends to improve fault-tolerance and usefulnes:

• In assert_correct, we now check the forwarding pointer first before extracting the narrowKlass from it. The forwarding pointer may be invalid. To do this, the forwarding-pointer-checking-block was moved up to before Klass* checks. We also use SafeFetch-based checks for some added safety.
• In assert_correct, we test the narrowKlass before decoding it. Otherwise, if its 0, we'd assert, if garbage, the resulting Klass* would be garbage.
• In print_failure, we now check pointers for readability before handing them to print_obj
• In print_obj, we now print the extracted narrowKlass, too (useful on Lilliput with its terse ids)
• In print_obj, we print a small hex dump of the header (32bytes). With little cost we get a glance at the raw header and the first members.
• In print_obj, we print to a single stringStream now and use auto indentation. Avoids having to deal with indentation and newlines manually and having to pay for four stringStreams.
• In assert_correct, I raised the print level for some of the klass-related assertions to safe_oop since the oop, since the hardened version of print_obj now avoids tripping over invalid Klass*

Tests: I manually tested the patch with a number of manually broken oops of various flavours.

Example output, invalid forwardee:

Referenced from:
  no interior location recorded (probably a plain heap scan, or detached oop)

Object:
  0x00000007cd800000 - nk 1798104 klass 0x00000000921b6fd8 [Ljava.lang.Object;

        allocated after mark start
        after update watermark
        marked strong
        marked weak
    not in collection set
    mark:  marked(0xdeaddeaddeaddeaf)
    region: | 3812|R  |Y|BTE    7cd800000,    7cd920bf8,    7cdc00000|TAMS    7cd800000|UWM    7cd800000|U  1154K|T     0B|G     0B|S  1154K|L     0B|CP   0

  0x00000007cd800000:   deaddeaddeaddeaf 00000a96001b6fd8 f9b005cff9b0054d f9b006d3f9b00651   .........o......M.......Q.......
  0x00000007cd800020:   f9b00797f9b00755 f9b00992f9b0087d f9b009a0f9b00999 f9b009a7f9b0098e   U.......}.......................

Forwardee:
  0xdeaddeaddeaddeac - safe print, no details

---

Progress

• Change must be properly reviewed (1 review required, with at least 1 Reviewer)
• Change must not contain extraneous whitespace
• Commit message must refer to an issue

Issue

• JDK-8361712: Improve ShenandoahAsserts printing (Enhancement - P4)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/26237/head:pull/26237
$ git checkout pull/26237

Update a local copy of the PR:
$ git checkout pull/26237
$ git pull https://git.openjdk.org/jdk.git pull/26237/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 26237

View PR using the GUI difftool:
$ git pr show -t 26237

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/26237.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back stuefe! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

[openjdk]

❗ This change is not yet ready to be integrated.
See the Progress checklist in the description for automated requirements.

[openjdk]

@tstuefe The following labels will be automatically applied to this pull request:

• hotspot
• shenandoah

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing lists. If you would like to change these labels, use the /label pull request command.

[mlbridge]

Webrevs

• 01: Full - Incremental (c7a9c26e)
• 00: Full (94fe071e)

[rkennke]

This looks really useful! I have only two comments/questions.

[rkennke]

Does this need to check for heap-end boundary? Otherwise it may crash trying to print bytes from after the heap, which may not be mapped?

[tstuefe]

No, that function is safe to use on any memory. It uses SafeFetch and prints dots for unreadable parts.

[rkennke]

Why do we need/want a default value here?

[tstuefe]

Ease of use, and less code. And it softly enforces 2 as a standard.