[roll] Roll fuchsia [kernel][vm] Fix UAF in PagerProxy

Fix a use-after-free bug in PagerProxy.  Prior to this commit,
PagerProxy would hold a raw const pointer to its PagerDispatcher, even
after the PagerDispatcher had been destroyed.  Now, PagerProxy's
pointer is non-const and is cleared during on_zero_handles (via
PagerProxy::OnDispatcherClose).  See also https://fxbug.dev/498512247

Change pager_ to a non-const pointer.

Set pager_ to null in OnDispatcherClose.

Remove the now redundant pager_dispatcher_closed_ field.

Add const pager_koid_ field so that PagerProxy::GetKoid continues to
function after OZH.  This is a partial revert of
https://fxrev.dev/1337566.

Update Dump to only query the PagerDispatcher name when pager_ is
non-null.

Add regression test.

I have verified that the new test will trigger a kasan panic.

```
[00000.653] 01111:01114> *** Running standalone test directly from userboot ***
[00000.836] 01111:01114> [==========] Flag Values:
[00000.836] 01111:01114>              --gtest_filter = Pager.ProxyUseAfterFree
[00000.836] 01111:01114>              --gtest_shuffle = true
[00000.836] 01111:01114>              --gtest_random_seed = -1
[00000.836] 01111:01114> [==========]
[00000.837] 01111:01114> [==========] Running 1 test from 1 test case.
[00000.837] 01111:01114> [----------] Global test environment set-up.
[00000.837] 01111:01114> [----------] 1 test from Pager
[00000.838] 01111:01114> [ RUN      ] Pager.ProxyUseAfterFree

ZIRCON KERNEL PANIC

UPTIME: 846ms, CPU: 3
Zircon/x64 ELF (EM_X86_64) build ID 7b49f4c67dd7256abbc6341572e457af9f5fba56
zx_system_get_version_string: git-81df98514578575906b4a61a6ac54d7ba239d612-dirty
kernel_debug_level: 2
{{{reset}}}
{{{module:0:vmzircon:elf:7b49f4c67dd7256abbc6341572e457af9f5fba56}}}
{{{mmap:0xffffffff00000000:0x1b3000:load:0:r:0x0}}}
{{{mmap:0xffffffff001b3000:0x5fc000:load:0:rx:0x1b3000}}}
{{{mmap:0xffffffff007af000:0x12000:load:0:r:0x7af000}}}
{{{mmap:0xffffffff007c1000:0x25b000:load:0:rw:0x7c1000}}}
stopping other cpus
halting cpu 1
halting cpu 2
halting cpu 0

KASAN detected a read error: ptr={{{data:0xffffff8041f67bf8}}}, size=0x8, caller: {{{pc:0xffffffff003325c9}}}
Shadow memory state around the buggy address 0xffffffe0083ecf7f:
0xffffffe0083ecf68: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0xffffffe0083ecf70: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0xffffffe0083ecf78: 0x00 0xfa 0xfa 0xfa 0xfa 0xfd 0xfd 0xfd
                                                        ^^
0xffffffe0083ecf80: 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd
0xffffffe0083ecf88: 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd 0xfd

*** KERNEL PANIC (caller pc: 0xffffffff0042007c, stack frame: 0xffffff828000de50):
*** kasan

platform_halt suggested_action 0 reason 4
Halting...
Zircon/x64 ELF (EM_X86_64) build ID 7b49f4c67dd7256abbc6341572e457af9f5fba56
zx_system_get_version_string: git-81df98514578575906b4a61a6ac54d7ba239d612-dirty
kernel_debug_level: 2
{{{reset}}}
{{{module:0:vmzircon:elf:7b49f4c67dd7256abbc6341572e457af9f5fba56}}}
{{{mmap:0xffffffff00000000:0x1b3000:load:0:r:0x0}}}
{{{mmap:0xffffffff001b3000:0x5fc000:load:0:rx:0x1b3000}}}
{{{mmap:0xffffffff007af000:0x12000:load:0:r:0x7af000}}}
{{{mmap:0xffffffff007c1000:0x25b000:load:0:rw:0x7c1000}}}
{{{bt:0:0xffffffff001beff8:ra}}}
{{{bt:1:0xffffffff006c7ddd:ra}}}
{{{bt:2:0xffffffff001b421d:ra}}}
{{{bt:3:0xffffffff0079ebcb:ra}}}
{{{bt:4:0xffffffff0042007c:ra}}}
{{{bt:5:0xffffffff003325c9:ra}}}
{{{bt:6:0xffffffff0051ce4d:ra}}}
{{{bt:7:0xffffffff002a3019:ra}}}
{{{bt:8:0xffffffff002c384c:ra}}}
{{{bt:9:0xffffffff007a23c3:ra}}}
Halted
entering panic shell loop
```

Original-Bug: 498512247
Test: fx core-tests --gtest_filter='Pager.ProxyUseAfterFree'
Original-Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/1584479
Original-Revision: 6774606f2f56ae3ac8acec65670d75db9f3b1b67
GitOrigin-RevId: 3d52526545e93d1bf5551dd10fccbbe486d11390
Change-Id: Iffc954524dd0ab54b01fc82a7236f78d1c22412f