From mboxrd@z Thu Jan  1 00:00:00 1970
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: cgroup/dmem: introduce a peak file
Date: Sat, 16 May 2026 11:20:49 +1000
Message-ID: <review-patch1-20260513-dmem_peak-v2-1-dac06999db9e@igalia.com>
In-Reply-To: <20260513-dmem_peak-v2-1-dac06999db9e@igalia.com>
References: <20260513-dmem_peak-v2-1-dac06999db9e@igalia.com> <20260513-dmem_peak-v2-1-dac06999db9e@igalia.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Patch Review

**Approach:** The patch adds a `get_resource_peak()` getter that reads `poo=
l->cnt.watermark`, a `dmem_cgroup_region_peak_show()` display function, and=
 a cftype entry =E2=80=94 all following the exact same pattern as `get_reso=
urce_current()` / `dmem_cgroup_region_current_show()`.

**Positive:**

1. The `READ_ONCE` usage in `get_resource_peak()` is correct and consistent=
 with how `page_counter_charge()` in `mm/page_counter.c` uses `WRITE_ONCE` =
to update the watermark. This is the same deliberate-racy-but-acceptable pa=
ttern documented in the page counter code.

2. The null check (`pool ? ... : 0`) mirrors `get_resource_current()` exact=
ly.

3. `CFTYPE_NOT_ON_ROOT` is the correct flag =E2=80=94 this matches `memory.=
peak` in the memory controller and is consistent with how `dmem.min`, `dmem=
.low`, and `dmem.max` are configured.

**Observations:**

1. **`current` lacks `CFTYPE_NOT_ON_ROOT` but `peak` has it.** Looking at t=
he existing tree, `dmem.current` does *not* set `CFTYPE_NOT_ON_ROOT`, so it=
 appears on root. The new `dmem.peak` does set `CFTYPE_NOT_ON_ROOT`, so it =
won't appear on root. This is consistent with how `memory.current` (no flag=
) and `memory.peak` (has `CFTYPE_NOT_ON_ROOT`) work in `mm/memcontrol.c`, s=
o the asymmetry is intentional and correct.

2. **No write/reset support.** The commit message and documentation say "re=
ad-only for now," which is fine for an initial version. The `memory.peak` c=
ontroller has grown `peak_open`/`peak_release`/`memory_peak_write` for per-=
fd local watermark tracking and reset support. If that's planned as a follo=
w-up, it might be worth mentioning in the commit message for future referen=
ce, but this is a nit =E2=80=94 the v1=E2=86=92v2 changelog already documen=
ts the deliberate decision to start read-only.

3. **Documentation placement is correct.** The `dmem.peak` entry is inserte=
d between `dmem.max/min/low` and `dmem.capacity`, which is a natural locati=
on grouping the per-cgroup files together before the root-only `dmem.capaci=
ty`.

4. **Watermark initialization.** The `page_counter_init()` function zeroes =
`usage` and sets `max` but does not explicitly initialize `watermark` or `l=
ocal_watermark`. Since `page_counter` structs in `dmem_cgroup_pool_state` a=
re allocated with `kzalloc` (which zero-initializes), the watermark starts =
at 0, which is correct for a "max seen so far" semantic. This is fine but w=
orth noting for anyone reading the patch.

**No functional issues found.** The patch is a straightforward, correct ext=
ension of the existing dmem cgroup interface.

---
Generated by Claude Code Patch Reviewer