public inbox for drm-ai-reviews@public-inbox.freedesktop.org
 help / color / mirror / Atom feed
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: drm/ttm: Be more aggressive when allocating below protection limit
Date: Tue, 03 Mar 2026 13:29:39 +1000	[thread overview]
Message-ID: <review-patch5-20260302-dmemcg-aggressive-protect-v5-5-ffd3a2602309@gmx.de> (raw)
In-Reply-To: <20260302-dmemcg-aggressive-protect-v5-5-ffd3a2602309@gmx.de>

Patch Review

**Purpose:** The core behavioral change. When an allocation is covered by cgroup protection (below min/low), attempt eviction of unprotected buffers instead of backing off to GTT.

**Logic analysis:**
```c
	may_evict |= dmem_cgroup_below_min(NULL, alloc_state->charge_pool);
	below_low = dmem_cgroup_below_low(NULL, alloc_state->charge_pool);
	alloc_state->only_evict_unprotected = !may_evict && below_low;
```

This creates three tiers:
1. **Below min** (strongest): `may_evict` is forced true → full eviction including low-protected BOs
2. **Below low but not min**: `only_evict_unprotected = true` → evict only unprotected BOs
3. **Not protected**: No change from current behavior

The `only_evict_unprotected` flag prevents the retry-with-low path:
```c
	if (!lret && evict_walk.hit_low && !state->only_evict_unprotected) {
```
This is correct — when we're in "protected allocator" mode, we should not escalate to evicting low-protected BOs because those are peers that deserve their protection.

The extended comment block explaining the protection model is excellent and makes the reasoning very clear.

The condition `(may_evict || below_low)` for converting `-ENOSPC` to `-EBUSY`:
```c
		if (ret == -ENOSPC && (may_evict || below_low))
			ret = -EBUSY;
```
ensures eviction is attempted for any protected allocation.

Note: `dmem_cgroup_below_min/low(NULL, ...)` uses global root for protection calculation. This is correct for determining if the *allocator* is protected. Patch 6 then handles the per-evictee calculation correctly using common ancestors.

---
Generated by Claude Code Patch Reviewer

  parent reply	other threads:[~2026-03-03  3:29 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-02 12:37 [PATCH v5 0/6] cgroup/dmem,drm/ttm: Improve protection in contended cases Natalie Vock
2026-03-02 12:37 ` [PATCH v5 1/6] cgroup/dmem: Add queries for protection values Natalie Vock
2026-03-03  3:29   ` Claude review: " Claude Code Review Bot
2026-03-02 12:37 ` [PATCH v5 2/6] cgroup,cgroup/dmem: Add (dmem_)cgroup_common_ancestor helper Natalie Vock
2026-03-02 14:38   ` Maarten Lankhorst
2026-03-03  3:29   ` Claude review: " Claude Code Review Bot
2026-03-02 12:37 ` [PATCH v5 3/6] drm/ttm: Extract code for attempting allocation in a place Natalie Vock
2026-03-02 15:08   ` Tvrtko Ursulin
2026-03-03  3:29   ` Claude review: " Claude Code Review Bot
2026-03-02 12:37 ` [PATCH v5 4/6] drm/ttm: Split cgroup charge and resource allocation Natalie Vock
2026-03-02 15:25   ` Tvrtko Ursulin
2026-03-03  3:29   ` Claude review: " Claude Code Review Bot
2026-03-02 12:37 ` [PATCH v5 5/6] drm/ttm: Be more aggressive when allocating below protection limit Natalie Vock
2026-03-02 17:02   ` Tvrtko Ursulin
2026-03-03  3:29   ` Claude Code Review Bot [this message]
2026-03-02 12:37 ` [PATCH v5 6/6] drm/ttm: Use common ancestor of evictor and evictee as limit pool Natalie Vock
2026-03-03  3:29   ` Claude review: " Claude Code Review Bot
2026-03-03  3:29 ` Claude review: cgroup/dmem,drm/ttm: Improve protection in contended cases Claude Code Review Bot
  -- strict thread matches above, loose matches on Subject: below --
2026-03-13 11:39 [PATCH v6 0/6] " Natalie Vock
2026-03-13 11:40 ` [PATCH v6 5/6] drm/ttm: Be more aggressive when allocating below protection limit Natalie Vock
2026-03-13 21:10   ` Claude review: " Claude Code Review Bot
2026-02-25 12:10 [PATCH v4 0/6] cgroup/dmem,drm/ttm: Improve protection in contended cases Natalie Vock
2026-02-25 12:10 ` [PATCH v4 5/6] drm/ttm: Be more aggressive when allocating below protection limit Natalie Vock
2026-02-27  3:36   ` Claude review: " Claude Code Review Bot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=review-patch5-20260302-dmemcg-aggressive-protect-v5-5-ffd3a2602309@gmx.de \
    --to=claude-review@example.com \
    --cc=dri-devel-reviews@example.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox