From: "H.J. Lu" <hjl.tools@gmail.com>
To: Florian Weimer <fweimer@redhat.com>
Cc: Szabolcs Nagy <szabolcs.nagy@arm.com>,
Rongwei Wang <rongwei.wang@linux.alibaba.com>,
GNU C Library <libc-alpha@sourceware.org>,
Adhemerval Zanella <adhemerval.zanella@linaro.org>,
xuyu@linux.alibaba.com, gavin.dg@linux.alibaba.com
Subject: Re: [PATCH v6 1/2] elf: Properly align PT_LOAD segments [BZ #28676]
Date: Mon, 13 Dec 2021 05:34:47 -0800 [thread overview]
Message-ID: <CAMe9rOqYZfpUJs1pq9o8i-MjjEQq-xNcZbiFtD4h6XtsPF_sqA@mail.gmail.com> (raw)
In-Reply-To: <87ee6gzldv.fsf@oldenburg.str.redhat.com>
On Mon, Dec 13, 2021 at 5:26 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> * H. J. Lu:
>
> > On Mon, Dec 13, 2021 at 3:59 AM Florian Weimer <fweimer@redhat.com> wrote:
> >>
> >> * Szabolcs Nagy:
> >>
> >> >> What are the p_align values of the involved objects? I would not expect
> >> >> any changes on 32-bit Arm because p_align and the run-time page size
> >> >> should match there.
> >> >
> >> > p_align is 64K
> >> >
> >> > and i see a lot of close to 64K PROT_NONE mappings
> >> > left behind after many dlclose which creates a lot of
> >> > vm fragmentation when dlopen/dlclose is called in a loop.
> >> >
> >> > (i think the mapping is at the beginning or end of
> >> > the lib as some kind of padding and left behind after
> >> > dlclose, but haven't confirmed this yet)
> >>
> >> Oh. So why there is a real bug here, I think we need to discuss what
> >> the change means for 64K p_align binaries on 4K kernels. Do the
> >> additional munmap calls matter for startup performance?
> >
> > We should align munmap arguments.
>
> Sorry, I meant “while there is a real bug here”, that is: even if we fix
> that, the additional munmap calls could hurt startup performance.
True.
> >> (I understand this is very much a correctness fix, but startup
> >> performance matters as well.)
> >>
> >
> > The kernel loader doesn't call munmap in this case. Should
> > we be concerned about the unused pages?
>
> Which approach leads to fewer mappings that count against the map limit
> (64K by default I believe)?
Without munmap, some pages will be wasted.
Should we change the linker to add a new DT_XXX entry to request
this new behavior? Linker will set it only when setting p_align > page
size due to variable alignment or a command-line option.
--
H.J.
next prev parent reply other threads:[~2021-12-13 13:35 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-04 4:58 [PATCH RFC 0/1] make ld.so map .text LOAD ssegments and aligned by p_align Rongwei Wang
2021-12-04 4:58 ` [PATCH RFC 1/1] elf: align the mapping address of LOAD segments with p_align Rongwei Wang
2021-12-04 18:10 ` Florian Weimer
2021-12-06 2:47 ` Rongwei Wang
2021-12-06 14:48 ` H.J. Lu
2021-12-08 2:14 ` Rongwei Wang
2021-12-08 2:33 ` H.J. Lu
2021-12-08 3:04 ` Rongwei Wang
2021-12-08 23:52 ` H.J. Lu
2021-12-09 1:43 ` Rongwei Wang
2021-12-10 12:39 ` [PATCH v5 0/2] fix p_align on PT_LOAD segment in DSO isn't honored Rongwei Wang
2021-12-10 12:39 ` [PATCH v5 1/2] elf: Properly align PT_LOAD segments Rongwei Wang
2021-12-10 15:43 ` H.J. Lu
2021-12-10 15:45 ` Florian Weimer
2021-12-10 18:54 ` H.J. Lu
2021-12-10 18:57 ` H.J. Lu
2021-12-10 12:39 ` [PATCH v5 2/2] Add a testcase to check alignment of PT_LOAD segment Rongwei Wang
2021-12-10 13:48 ` Adhemerval Zanella
2021-12-10 15:41 ` H.J. Lu
2021-12-10 18:56 ` H.J. Lu
2021-12-10 20:05 ` Adhemerval Zanella
2021-12-10 20:24 ` H.J. Lu
2021-12-10 21:34 ` Adhemerval Zanella
2021-12-10 13:13 ` [PATCH v5 0/2] fix p_align on PT_LOAD segment in DSO isn't honored H.J. Lu
2021-12-10 13:58 ` Rongwei Wang
2021-12-13 2:51 ` [PATCH v6 " Rongwei Wang
2021-12-13 2:51 ` [PATCH v6 1/2] elf: Properly align PT_LOAD segments [BZ #28676] Rongwei Wang
2021-12-13 11:05 ` Szabolcs Nagy
2021-12-13 11:17 ` Florian Weimer
2021-12-13 11:35 ` Szabolcs Nagy
2021-12-13 11:59 ` Florian Weimer
2021-12-13 13:20 ` H.J. Lu
2021-12-13 13:26 ` Florian Weimer
2021-12-13 13:34 ` H.J. Lu [this message]
2021-12-13 11:46 ` Andreas Schwab
2021-12-13 11:52 ` Szabolcs Nagy
2021-12-13 14:51 ` Rongwei Wang
2021-12-13 17:37 ` Szabolcs Nagy
2021-12-13 17:50 ` Florian Weimer
2021-12-13 2:51 ` [PATCH v6 2/2] Add a testcase to check alignment of PT_LOAD segment Rongwei Wang
2021-12-14 2:03 ` [PATCH v6 0/2] fix p_align on PT_LOAD segment in DSO isn't honored Fangrui Song
2021-12-14 3:56 ` H.J. Lu
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=CAMe9rOqYZfpUJs1pq9o8i-MjjEQq-xNcZbiFtD4h6XtsPF_sqA@mail.gmail.com \
--to=hjl.tools@gmail.com \
--cc=adhemerval.zanella@linaro.org \
--cc=fweimer@redhat.com \
--cc=gavin.dg@linux.alibaba.com \
--cc=libc-alpha@sourceware.org \
--cc=rongwei.wang@linux.alibaba.com \
--cc=szabolcs.nagy@arm.com \
--cc=xuyu@linux.alibaba.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).