mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "David Wang" <00107082@163.com>
To: "Suren Baghdasaryan" <surenb@google.com>
Cc: kent.overstreet@linux.dev, akpm@linux-foundation.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/2] alloc_tag: add timestamp to codetag iterator
Date: Sat, 10 May 2025 00:16:57 +0800 (CST)	[thread overview]
Message-ID: <204874d5.b00e.196b5d5ea1d.Coremail.00107082@163.com> (raw)
In-Reply-To: <CAJuCfpHRTGvctcEwXHd1bpfUiFa=A--zKmVBJggB5D9huPEdSA@mail.gmail.com>


At 2025-05-10 00:10:01, "Suren Baghdasaryan" <surenb@google.com> wrote:
>On Thu, May 8, 2025 at 10:52 PM David Wang <00107082@163.com> wrote:
>>
>> Codetag iterator use <id,address> pair to guarantee the
>> validness. But both id and address can be reused, there is
>> theoretical possibility when module inserted right after
>> another module removed, kmalloc return an address kfree by
>> previous module and IDR key reuse the key recently removed.
>>
>> Add timestamp to codetag_module and code_iterator, the
>> timestamp is generated by a clock which is strickly
>> incremented whenever a module is loaded. An iterator is
>> valid if and only if its timestamp match codetag_module's.
>>
>> Signed-off-by: David Wang <00107082@163.com>
>> ---
>>  include/linux/codetag.h |  1 +
>>  lib/codetag.c           | 12 ++++++++++--
>>  2 files changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/linux/codetag.h b/include/linux/codetag.h
>> index d14dbd26b370..61d43c3fbd19 100644
>> --- a/include/linux/codetag.h
>> +++ b/include/linux/codetag.h
>> @@ -54,6 +54,7 @@ struct codetag_iterator {
>>         struct codetag_module *cmod;
>>         unsigned long mod_id;
>>         struct codetag *ct;
>> +       unsigned long timestamp;
>
>The way you implement this, it is not really a timestamp, more like a
>unique id or sequence number. I would suggest calling it mod_seq or
>something similar. Maybe: cttype->next_mod_seq, iter->mod_seq and
>cmod->mod_seq.

Copy that~

>
>>  };
>>
>>  #ifdef MODULE
>> diff --git a/lib/codetag.c b/lib/codetag.c
>> index 42aadd6c1454..973bfa5dd5db 100644
>> --- a/lib/codetag.c
>> +++ b/lib/codetag.c
>> @@ -13,6 +13,8 @@ struct codetag_type {
>>         struct idr mod_idr;
>>         struct rw_semaphore mod_lock; /* protects mod_idr */
>
>The comment above should be expanded to say that mod_lock also
>protects accesses to cttype->clock, iter->timestamp and
>cmod->timestamp.

Good catch~!

Thanks
>
>>         struct codetag_type_desc desc;
>> +       /* generates timestamp for module load */
>> +       unsigned long clock;
>>  };
>>
>>  struct codetag_range {
>> @@ -23,6 +25,7 @@ struct codetag_range {
>>  struct codetag_module {
>>         struct module *mod;
>>         struct codetag_range range;
>> +       unsigned long timestamp;
>>  };
>>
>>  static DEFINE_MUTEX(codetag_lock);
>> @@ -48,6 +51,7 @@ struct codetag_iterator codetag_get_ct_iter(struct codetag_type *cttype)
>>                 .cmod = NULL,
>>                 .mod_id = 0,
>>                 .ct = NULL,
>> +               .timestamp = 0,
>>         };
>>
>>         return iter;
>> @@ -91,11 +95,13 @@ struct codetag *codetag_next_ct(struct codetag_iterator *iter)
>>                 if (!cmod)
>>                         break;
>>
>> -               if (cmod != iter->cmod) {
>> +               if (!iter->cmod || iter->timestamp != cmod->timestamp) {
>>                         iter->cmod = cmod;
>> +                       iter->timestamp = cmod->timestamp;
>>                         ct = get_first_module_ct(cmod);
>> -               } else
>> +               } else {
>>                         ct = get_next_module_ct(iter);
>> +               }
>>
>>                 if (ct)
>>                         break;
>> @@ -190,6 +196,8 @@ static int codetag_module_init(struct codetag_type *cttype, struct module *mod)
>>         cmod->range = range;
>>
>>         down_write(&cttype->mod_lock);
>> +       cttype->clock++;
>> +       cmod->timestamp = cttype->clock;
>>         err = idr_alloc(&cttype->mod_idr, cmod, 0, 0, GFP_KERNEL);
>>         if (err >= 0) {
>>                 cttype->count += range_size(cttype, &range);
>> --
>> 2.39.2
>>

  reply	other threads:[~2025-05-09 16:17 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-07 17:55 [PATCH] alloc_tag: avoid mem alloc and iter reset when reading allocinfo David Wang
2025-05-07 18:19 ` David Wang
2025-05-07 23:42   ` [PATCH] " Suren Baghdasaryan
2025-05-08  0:01     ` Kent Overstreet
2025-05-08  3:06       ` David Wang
2025-05-08  3:31         ` Kent Overstreet
2025-05-08  3:35           ` David Wang
2025-05-08  4:07             ` Kent Overstreet
2025-05-08  5:51               ` David Wang
2025-05-08 13:33                 ` Kent Overstreet
2025-05-08 16:24                   ` David Wang
2025-05-08 16:34                     ` Kent Overstreet
2025-05-08 16:58                       ` David Wang
2025-05-08 17:17                         ` David Wang
2025-05-08 17:26                           ` Kent Overstreet
2025-05-08  2:24     ` David Wang
2025-05-07 23:36 ` Suren Baghdasaryan
2025-05-08  3:10   ` David Wang
2025-05-08 15:32   ` David Wang
2025-05-08 21:41     ` Suren Baghdasaryan
2025-05-09  5:51 ` [PATCH 1/2] alloc_tag: add timestamp to codetag iterator David Wang
2025-05-09 16:10   ` Suren Baghdasaryan
2025-05-09 16:16     ` David Wang [this message]
2025-05-09  5:53 ` [PATCH 2/2] alloc_tag: keep codetag iterator cross read() calls David Wang
2025-05-09 17:34   ` Suren Baghdasaryan
2025-05-09 17:45     ` David Wang
2025-05-09 17:38 ` [PATCH v2 1/2] alloc_tag: add sequence number for module and iterator David Wang
2025-05-09 19:25   ` Suren Baghdasaryan
2025-06-09  6:42     ` [PATCH v3 " David Wang
2025-06-09 16:41       ` Suren Baghdasaryan
2025-05-09 17:39 ` [PATCH v2 2/2] alloc_tag: keep codetag iterator active between read() calls David Wang
2025-05-09 18:33   ` Tim Chen
2025-05-09 19:36     ` Suren Baghdasaryan
2025-05-09 19:46       ` Tim Chen
2025-05-09 20:46         ` Suren Baghdasaryan
2025-05-09 21:15           ` Suren Baghdasaryan
2025-05-10  3:10             ` David Wang
2025-05-10  3:30               ` Suren Baghdasaryan
2025-05-10  3:58                 ` David Wang
2025-05-10  4:03                   ` Suren Baghdasaryan
2025-06-09  6:44                     ` [PATCH v3 2/2] alloc_tag: keep codetag iterator active between read() David Wang
2025-06-10 16:22                       ` Suren Baghdasaryan
2025-05-10  3:35         ` [PATCH v2 2/2] alloc_tag: keep codetag iterator active between read() calls David Wang
2025-05-10  3:25     ` David Wang

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=204874d5.b00e.196b5d5ea1d.Coremail.00107082@163.com \
    --to=00107082@163.com \
    --cc=akpm@linux-foundation.org \
    --cc=kent.overstreet@linux.dev \
    --cc=linux-kernel@vger.kernel.org \
    --cc=surenb@google.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

all inboxes | Powered by JetHome®