OpenVZ / Virtuozzo kernel development (devel@openvz.org)
 help / color / mirror / Atom feed
From: Konstantin Khorenko <khorenko@virtuozzo.com>
To: Vladimir Riabchun <vladimir.riabchun@virtuozzo.com>
Cc: devel@openvz.org
Subject: Re: [Devel] [PATCH hci-8.1 v3 0/6] ve/printk: Fix logs leakage via ve_printk called from IRQ context
Date: Thu, 12 Nov 2026 17:35:40 +0100	[thread overview]
Message-ID: <179450134078.4060562.16522511515702111377.b4-review@b4> (raw)
In-Reply-To: <20261012-ve-printk-irq-v3-0-efc95fd5c78a@virtuozzo.com>

I won't apply this version, please send v4 with the changes below.

On Mon, 12 Oct 2026 00:52:27 +0200, Vladimir Riabchun <vladimir.riabchun@virtuozzo.com> wrote:
> Vladimir Riabchun (6):
>       ve/printk: Add sanity checks in ve_printk functions
>       ve/printk: Report potentially leaking usage of ve_printk functions
>       ve/printk: Specify ve in printks
>       ve/printk: Fix time virtualization
>       ve/failcount: Use ve_printk with specified ve
>       ve: add neigh_tbl failcounter

The approach is fine: after the series nothing outside task context
picks a container log from get_exec_env(). What to change for v4:

1. Patch 3: rename the local "__ve" in ve_printk_ratelimited() to a
   name no caller uses, and write "(dst)" in the macro.

2. Patch 3: make __ve_log_state() static, it is used only in printk.c.
   Drop its prototype from internal.h in patch 4.

3. Patch 3: rewrite the message. Say what users see change: the
   conntrack table full, expectation table full and TCPMSS messages
   now go to the log of net->owner_ve, and ve_printk_ratelimited()
   with VE_LOG outside task context warns and prints to the host log.
   Add Fixes tags for a335738ad21cf and bc279343e9663.

4. Patch 4: rewrite the message around the real broken case, and say
   that records stored while the CT has no ve_nsproxy keep host time.
   Add Fixes tags for 1249fbadd45b3 and 8b42370ae936e.

5. Patch 6: count and warn only when !ve_is_super(ve), as Vasileios
   asked.

6. Patch 6: keep one Feature: line, "per-ve failcounters".

7. Patch 2: add a Fixes tag for 8c24627ed5346 ("ve/printk: Introduce
   ve_printk()").

8. Patch 5: double space in "Use new  __ve_printk helpers".

9. Patch 1: say why the condition is reordered (so that patch 2 can
   put the WARN before the get_exec_env() check), or fold the reorder
   into patch 2.

10. Cover letter:
    - "No tests crashed" says nothing here: these WARNs do not crash
      without panic_on_warn. Say whether dmesg showed a WARNING from
      these functions in the neoltp runs.
    - "failcounts can be now used in any context": not under the
      scheduler locks, __ve_printk() is not deferred. "can now be
      used" in softirq and hardirq.
    - Typos: "there an no", "GET)VE_LOG".

Separate patches, not needed for v4:

1. WARN_ON_ONCE() sits inside the shared functions, so only the first
   offending call site per boot gets a backtrace. A once flag per
   caller would report all of them.

2. ve_offline() does kfree(ve->ve_name) without RCU, while conntrack
   and now neigh_alloc() read ve_name(net->owner_ve) from softirq. The
   netns can outlive the offline cgroup.

-- 
Konstantin Khorenko,
Virtuozzo Linux Kernel Team

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

      parent reply	other threads:[~2026-11-12 16:37 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-11 22:52 Vladimir Riabchun
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 1/6] ve/printk: Add sanity checks in ve_printk functions Vladimir Riabchun
2026-10-11 22:57   ` no-reply
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 2/6] ve/printk: Report potentially leaking usage of " Vladimir Riabchun
2026-10-11 22:59   ` no-reply
2026-10-26 15:21     ` Vladimir Riabchun
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 3/6] ve/printk: Specify ve in printks Vladimir Riabchun
2026-10-11 23:02   ` no-reply
2026-10-26 15:27     ` Vladimir Riabchun
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 4/6] ve/printk: Fix time virtualization Vladimir Riabchun
2026-10-11 23:03   ` no-reply
2026-10-26 15:36     ` Vladimir Riabchun
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 5/6] ve/failcount: Use ve_printk with specified ve Vladimir Riabchun
2026-10-11 23:05   ` no-reply
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-11 22:52 ` [Devel] [PATCH hci-8.1 v3 6/6] ve: add neigh_tbl failcounter Vladimir Riabchun
2026-10-11 23:04   ` no-reply
2026-10-12  7:59   ` Vasileios Almpanis
2026-11-12 16:35   ` Konstantin Khorenko
2026-10-12  7:59 ` [Devel] [PATCH hci-8.1 v3 0/6] ve/printk: Fix logs leakage via ve_printk called from IRQ context Vasileios Almpanis
2026-11-12 16:35 ` Konstantin Khorenko [this message]

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=179450134078.4060562.16522511515702111377.b4-review@b4 \
    --to=khorenko@virtuozzo.com \
    --cc=devel@openvz.org \
    --cc=vladimir.riabchun@virtuozzo.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