GNU bug report logs - #72278
29.4; Proced reports incorrect results with 16KB page size

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 72278 in the body.
You can then email your comments to 72278 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Rahguzar <rahguzar <at> zohomail.eu>
To: bug-gnu-emacs <at> gnu.org
Subject: 29.4; Proced reports incorrect results with 16KB page size
Date: Wed, 24 Jul 2024 19:22:08 +0200

Dear Emacs Maintainers,

I am using GNU/Linux on an Apple M2 processor using the Asahi Linux
project [1]. They provide a kernel with 16KB page size [2]. This causes proced
to show incorrect values for percentage memory (other memory related
attributes might also be affected). E.g. comparing with top I see that
value reported by proced is a quarter of that related by top. 

[1] https://asahilinux.org
[2] https://discussion.fedoraproject.org/t/switch-to-the-kernel-16k-variant/87711


In GNU Emacs 29.4 (build 1, aarch64-redhat-linux-gnu, GTK+ Version
 3.24.42, cairo version 1.18.0) of 2024-07-08 built on
 c7db9ae0ee0641de8f7c6190b678dcb8
System Description: Fedora Linux Asahi Remix 40 (Forty)

Configured using:
 'configure --build=aarch64-redhat-linux-gnu
 --host=aarch64-redhat-linux-gnu --program-prefix=
 --disable-dependency-tracking --prefix=/usr --exec-prefix=/usr
 --bindir=/usr/bin --sbindir=/usr/sbin --sysconfdir=/etc
 --datadir=/usr/share --includedir=/usr/include --libdir=/usr/lib64
 --libexecdir=/usr/libexec --localstatedir=/var --runstatedir=/run
 --sharedstatedir=/var/lib --mandir=/usr/share/man
 --infodir=/usr/share/info --with-cairo --with-dbus --with-gif
 --with-gpm=no --with-harfbuzz --with-jpeg --with-json --with-modules
 --with-native-compilation=aot --with-pgtk --with-png --with-rsvg
 --with-sqlite3 --with-tiff --with-tree-sitter --with-webp --with-xpm
 --with-xwidgets build_alias=aarch64-redhat-linux-gnu
 host_alias=aarch64-redhat-linux-gnu CC=gcc 'CFLAGS=-DMAIL_USE_LOCKF -O2
 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches
 -pipe -Wall -Werror=format-security
 -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS
 -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong
 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1
 -mbranch-protection=standard -fasynchronous-unwind-tables
 -fstack-clash-protection -fno-omit-frame-pointer
 -mno-omit-leaf-frame-pointer ' LDFLAGS=-Wl,-z,relro
 PKG_CONFIG_PATH=:/usr/lib64/pkgconfig:/usr/share/pkgconfig CXX=g++
 'CXXFLAGS=-O2 -flto=auto -ffat-lto-objects -fexceptions -g
 -grecord-gcc-switches -pipe -Wall -Werror=format-security
 -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS
 -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong
 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1
 -mbranch-protection=standard -fasynchronous-unwind-tables
 -fstack-clash-protection -fno-omit-frame-pointer
 -mno-omit-leaf-frame-pointer ''

Configured features:
ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS GSETTINGS HARFBUZZ JPEG JSON
LIBOTF LIBSELINUX LIBSYSTEMD LIBXML2 MODULES NATIVE_COMP NOTIFY INOTIFY
PDUMPER PGTK PNG RSVG SECCOMP SOUND SQLITE3 THREADS TIFF
TOOLKIT_SCROLL_BARS TREE_SITTER WEBP XIM XWIDGETS GTK3 ZLIB

Important settings:
  value of $LANG: en_US.UTF-8
  locale-coding-system: utf-8-unix

Major mode: ELisp/l

Minor modes in effect:
  pdf-occur-global-minor-mode: t
  shell-dirtrack-mode: t
  undo-fu-session-mode: t
  highlight-quoted-mode: t
  corfu-history-mode: t
  corfu-popupinfo-mode: t
  corfu-mode: t
  evil-surround-mode: t
  server-mode: t
  outline-minor-mode: t
  dirvish-override-dired-mode: t
  satchel-mode: t
  nerd-icons-completion-mode: t
  marginalia-mode: t
  vertico-multiform-mode: t
  vertico-mode: t
  evil-goggles-mode: t
  anzu-mode: t
  evil-mode: t
  evil-local-mode: t
  recentf-mode: t
  savehist-mode: t
  electric-pair-mode: t
  save-place-mode: t
  override-global-mode: t
  flymake-mode: t
  pixel-scroll-precision-mode: t
  repeat-mode: t
  midnight-mode: t
  global-eldoc-mode: t
  eldoc-mode: t
  show-paren-mode: t
  electric-indent-mode: t
  mouse-wheel-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  buffer-read-only: t
  size-indication-mode: t
  line-number-mode: t
  transient-mark-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  abbrev-mode: t
  view-mode: t

Load-path shadows:
/home/azeem/.local/state/emacs/elpaca/builds/transient/transient hides /usr/share/emacs/29.4/lisp/transient
/home/azeem/.local/state/emacs/elpaca/builds/ef-themes/theme-loaddefs hides /usr/share/emacs/29.4/lisp/theme-loaddefs
/home/azeem/.local/state/emacs/elpaca/builds/eww-hacks/shr hides /usr/share/emacs/29.4/lisp/net/shr
/home/azeem/.local/state/emacs/elpaca/builds/eww-hacks/eww hides /usr/share/emacs/29.4/lisp/net/eww

Features:
(shadow files-x emacsbug evil-nerd-commenter html-mode-expansions
sgml-mode evil-nerd-commenter-operator evil-nerd-commenter-sdk misearch
multi-isearch pdf-roll crm consult-info gnus-fun embark-org
embark-consult embark inspirehep-pdf companion-mode inspirehep reftex
reftex-loaddefs reftex-vars pdf-sync pdf-annot pdf-outline pdf-links
pdf-history pdf-occur tablist tablist-filter semantic/wisent/comp
semantic/wisent semantic/wisent/wisent semantic/util-modes semantic/util
semantic semantic/tag semantic/lex semantic/fw mode-local cedet
pdf-isearch pdf-misc pdf-tools pdf-view pdf-cache pdf-info tq pdf-util
pdf-macs password-store auth-source-pass with-editor shell consult-imenu
checkdoc vc-git vc-dispatcher package-lint-flymake package-lint
let-alist finder finder-inf lisp-mnt package url-handlers
vertico-directory oc-basic url-queue mm-url ol-rmail ol-mhe ol-irc
ol-info ol-gnus nnselect ol-docview doc-view ol-bibtex bibtex ol-bbdb
ol-w3m ol-doi org-link-doi descr-text mastodon-notifications smerge-mode
diff diff-mode link-hint ffap goto-addr avy jka-compr view helpful
cc-langs cc-vars cc-defs imenu trace edebug debug backtrace info-look
info f help-fns radix-tree elisp-refs s dash undo-fu-session
nerd-icons-corfu dabbrev delsel jit-spell ispell cape face-remap
mastodon-media image-mode exif mastodon-profile mastodon-tl url-http
url-gw url-cache url-auth mastodon-auth mastodon-client plstore mastodon
mastodon-search mastodon-toot facemenu mastodon-iso persist
mastodon-http request adaptive-wrap shr-color highlight-quoted
corfu-history corfu-popupinfo corfu evil-embrace embrace expand-region
text-mode-expansions the-org-mode-expansions org-element org-persist xdg
org-id org-refile avl-tree generator er-basic-expansions
expand-region-core expand-region-custom evil-surround qp sort smiley
gnus-cite mm-archive mail-extr textsec uni-scripts idna-mapping
ucs-normalize uni-confusable textsec-check gnus-async gnus-bcklg gnus-ml
disp-table nndraft nnmh vertico-repeat epa-file network-stream nsm
gnus-agent gnus-srvr gnus-score score-mode nnvirtual gnus-msg nntp
gnus-cache fabs consult server mu4e mu4e-org org ob ob-tangle ob-ref
ob-lob ob-table ob-exp org-macro org-src ob-comint org-pcomplete
pcomplete org-list org-footnote org-faces org-entities noutline outline
ob-emacs-lisp ob-core ob-eval org-cycle org-table ol org-fold
org-fold-core org-keys oc org-loaddefs find-func org-version org-compat
org-macs mu4e-notification notifications mu4e-main smtpmail mu4e-view
mu4e-mime-parts gnus-art mm-uu mml2015 mm-view mml-smime smime gnutls
dig gnus-sum gnus-group gnus-undo gnus-start gnus-dbus dbus gnus-cloud
nnimap nnmail mail-source utf7 nnoo parse-time iso8601 gnus-spec
gnus-int gnus-range gnus-win gnus cal-menu calendar cal-loaddefs
mu4e-headers mu4e-thread mu4e-actions mu4e-compose nnheader range
mu4e-search mu4e-lists mu4e-bookmarks mu4e-mark mu4e-message shr
pixel-fill kinsoku url-file svg xml dom browse-url flow-fill mule-util
hl-line mu4e-contacts mu4e-update mu4e-folders mu4e-context
mu4e-query-items mu4e-server mu4e-modeline mu4e-vars mu4e-helpers
mu4e-config mu4e-window ido message sendmail yank-media puny cus-start
dirvish transient format-spec autorevert filenotify dired dired-loaddefs
rfc822 mml mml-sec epa derived epg rfc6068 epg-config gnus-util
time-date mm-decode mm-bodies mm-encode mail-parse rfc2231 rfc2047
rfc2045 mm-util ietf-drums mail-prsvr mailabbrev mail-utils gmm-utils
mailheader mu4e-obsolete usher satchel ibuf-ext ibuffer ibuffer-loaddefs
bookmark nerd-icons-completion nerd-icons nerd-icons-faces
nerd-icons-data nerd-icons-data-mdicon nerd-icons-data-flicon
nerd-icons-data-codicon nerd-icons-data-devicon nerd-icons-data-sucicon
nerd-icons-data-wicon nerd-icons-data-faicon nerd-icons-data-powerline
nerd-icons-data-octicon nerd-icons-data-pomicon nerd-icons-data-ipsicon
marginalia vertico-multiform vertico compat compat-30 evil-goggles pulse
color ef-spring-theme ef-themes pdf-loader evil-anzu anzu evil
evil-integration evil-maps evil-commands reveal evil-jumps
evil-command-window evil-types evil-search evil-ex evil-macros
evil-repeat evil-states evil-core comp comp-cstr cl-extra help-mode
advice evil-common rect evil-vars edmacro kmacro orderless
mastodon-autoloads request-autoloads persist-autoloads yeetube-autoloads
filechooser-autoloads wile-autoloads adaptive-wrap-autoloads
usher-autoloads undo-fu-session-autoloads ef-themes-autoloads
jit-spell-autoloads satchel-autoloads pdf-tools-autoloads
tablist-autoloads xr-autoloads link-hint-autoloads avy-autoloads
yaml-mode-autoloads markdown-mode-autoloads magit-autoloads
git-commit-autoloads magit-section-autoloads inspirehep-autoloads
consult-hoogle-autoloads haskell-mode-autoloads geiser-guile-autoloads
geiser-autoloads fabs-autoloads companion-mode-autoloads
eww-hacks-autoloads evil-embrace-autoloads embrace-autoloads
expand-region-autoloads evil-surround-autoloads
evil-nerd-commenter-autoloads evil-goggles-autoloads evil-anzu-autoloads
anzu-autoloads evil-autoloads goto-chg-autoloads
eshell-syntax-highlighting-autoloads eat-autoloads
package-lint-flymake-autoloads package-lint-autoloads
elisp-def-autoloads macrostep-autoloads helpful-autoloads f-autoloads
elisp-refs-autoloads dash-autoloads s-autoloads
highlight-quoted-autoloads dirvish-autoloads transient-autoloads
diredfl-autoloads trashed-autoloads vertico-autoloads
orderless-autoloads nerd-icons-completion-autoloads marginalia-autoloads
embark-consult-autoloads embark-autoloads nerd-icons-corfu-autoloads
nerd-icons-autoloads cape-autoloads corfu-autoloads consult-autoloads
password-store-autoloads with-editor-autoloads compat-autoloads
elpaca-log elpaca-ui url url-proxy url-privacy url-expand url-methods
url-history url-cookie generate-lisp-file url-domsuf url-util mailcap
recentf tree-widget savehist elec-pair saveplace bind-key url-parse
auth-source eieio eieio-core password-cache json map url-vars easy-mmode
cus-edit pp cus-load wid-edit pcase ibuf-macs flymake-proc flymake
project byte-opt bytecomp byte-compile compile text-property-search
comint ansi-osc ansi-color warnings icons thingatpt cl-macs pixel-scroll
cua-base ring subr-x repeat rx midnight gv cl-seq elpaca elpaca-process
cl-loaddefs cl-lib elpaca-autoloads rmc iso-transl tooltip cconv eldoc
paren electric uniquify ediff-hook vc-hooks lisp-float-type elisp-mode
mwheel term/pgtk-win pgtk-win term/common-win pgtk-dnd tool-bar dnd
fontset image regexp-opt fringe tabulated-list replace newcomment
text-mode lisp-mode prog-mode register page tab-bar menu-bar rfn-eshadow
isearch easymenu timer select scroll-bar mouse jit-lock font-lock syntax
font-core term/tty-colors frame minibuffer nadvice seq simple cl-generic
indonesian philippine cham georgian utf-8-lang misc-lang vietnamese
tibetan thai tai-viet lao korean japanese eucjp-ms cp51932 hebrew greek
romanian slovak czech european ethiopic indian cyrillic chinese
composite emoji-zwj charscript charprop case-table epa-hook
jka-cmpr-hook help abbrev obarray oclosure cl-preloaded button loaddefs
theme-loaddefs faces cus-face macroexp files window text-properties
overlay sha1 md5 base64 format env code-pages mule custom widget keymap
hashtable-print-readable backquote threads xwidget-internal dbusbind
inotify dynamic-setting system-font-setting font-render-setting cairo
gtk pgtk multi-tty make-network-process native-compile emacs)

Memory information:
((conses 16 1976204 1127031)
 (symbols 48 94384 7)
 (strings 32 478350 81603)
 (string-bytes 1 31638678)
 (vectors 16 149464)
 (vector-slots 8 3721744 648039)
 (floats 8 2069 5420)
 (intervals 56 197274 14128)
 (buffers 984 162))

Message #8 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Rahguzar <rahguzar <at> zohomail.eu>, 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4;
 Proced reports incorrect results with 16KB page size
Date: Wed, 24 Jul 2024 12:29:25 -0700

Rahguzar via "Bug reports for GNU Emacs, the Swiss army knife of text
editors" <bug-gnu-emacs <at> gnu.org> writes:

> Dear Emacs Maintainers,
>
> I am using GNU/Linux on an Apple M2 processor using the Asahi Linux
> project [1]. They provide a kernel with 16KB page size [2]. This causes proced
> to show incorrect values for percentage memory (other memory related
> attributes might also be affected). E.g. comparing with top I see that
> value reported by proced is a quarter of that related by top.
>
> [1] https://asahilinux.org
> [2] https://discussion.fedoraproject.org/t/switch-to-the-kernel-16k-variant/87711

1. What is the output of running this command?

       grep MemTotal /proc/meminfo

   Does that value seem right?  Does that file make sense in general?

2. Copy this to a buffer somewhere, put point after the last
   parenthesis, and press `C-u C-x C-e':

        (pp (mapcar #'process-attributes
             (seq-take (list-system-processes) 5)))

    Please post the result back here.

    Could you also tell us which values in the resulting list are wrong
    (a.k.a. "different from top")?

Thanks in advance.

Message #11 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Rahguzar <rahguzar <at> zohomail.eu>
To: Stefan Kangas <stefankangas <at> gmail.com>
Cc: 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4; Proced reports incorrect results with 16KB
 page size
Date: Wed, 24 Jul 2024 22:40:53 +0200

Hi Stefan,

Stefan Kangas <stefankangas <at> gmail.com> writes:

> Rahguzar via "Bug reports for GNU Emacs, the Swiss army knife of text
> editors" <bug-gnu-emacs <at> gnu.org> writes:
>
>> Dear Emacs Maintainers,
>>
>> I am using GNU/Linux on an Apple M2 processor using the Asahi Linux
>> project [1]. They provide a kernel with 16KB page size [2]. This causes proced
>> to show incorrect values for percentage memory (other memory related
>> attributes might also be affected). E.g. comparing with top I see that
>> value reported by proced is a quarter of that related by top.
>>
>> [1] https://asahilinux.org
>> [2] https://discussion.fedoraproject.org/t/switch-to-the-kernel-16k-variant/87711
>
> 1. What is the output of running this command?
>
>        grep MemTotal /proc/meminfo
>
>    Does that value seem right?  Does that file make sense in general?

The value and the file generally look correct to me.

> 2. Copy this to a buffer somewhere, put point after the last
>    parenthesis, and press `C-u C-x C-e':
>
>         (pp (mapcar #'process-attributes
>              (seq-take (list-system-processes) 5)))
>
>     Please post the result back here.
>
>     Could you also tell us which values in the resulting list are wrong
>     (a.k.a. "different from top")?

The processes that end up in that list have memory footprint so small
that top ends up showing percentage memory as 0. I modified the snippet
to obtain 5 processes started by firefox:

(((args . "/usr/lib64/firefox/firefox")
  (pmem . 3.018895533065192)
  (rss . 727876)
  (vsize . 17513856)
  (thcount . 105)
  (nice . 0)
  (pri . 20)
  (pcpu . 1.0718871382569872)
  (etime 12 34487 450000 0)
  (start 26260 56544 260000 0)
  (ctime 0 8651 310000 0)
  (cstime 0 393 900000 0)
  (cutime 0 8257 410000 0)
  (time 0 8799 330000 0)
  (stime 0 2500 340000 0)
  (utime 0 6298 990000 0)
  (cmajflt . 83)
  (cminflt . 10288762)
  (majflt . 1571)
  (minflt . 20999882)
  (tpgid . -1)
  (ttname . "")
  (sess . 1239)
  (pgrp . 1239)
  (ppid . 1)
  (state . "S")
  (comm . "firefox")
  (group . "azeem")
  (egid . 1000)
  (user . "azeem")
  (euid . 1000))
 ((args . "/usr/lib64/firefox/firefox -contentproc -parentBuildID 20240712194222 -prefsLen 42067 -prefMapSize 262227 -appDir /usr/lib64/firefox/browser {6dc0e492-94e8-413e-83e0-ac2b970bde23} 1241 socket")
  (pmem . 0.05839737689600688)
  (rss . 14080)
  (vsize . 576016)
  (thcount . 6)
  (nice . 0)
  (pri . 20)
  (pcpu . 5.359847034838214e-05)
  (etime 12 34486 950000 0)
  (start 26260 56544 770000 0)
  (ctime 0 0 0 0)
  (cstime 0 0 0 0)
  (cutime 0 0 0 0)
  (time 0 0 440000 0)
  (stime 0 0 180000 0)
  (utime 0 0 260000 0)
  (cmajflt . 0)
  (cminflt . 0)
  (majflt . 9)
  (minflt . 1669)
  (tpgid . -1)
  (ttname . "")
  (sess . 1239)
  (pgrp . 1239)
  (ppid . 1241)
  (state . "S")
  (comm . "Socket Process")
  (group . "azeem")
  (egid . 1000)
  (user . "azeem")
  (euid . 1000))
 ((args . "/usr/lib64/firefox/firefox -contentproc -childID 1 -isForBrowser -prefsLen 42067 -prefMapSize 262227 -jsInitLen 234780 -parentBuildID 20240712194222 -greomni /usr/lib64/firefox/omni.ja -appomni /usr/lib64/firefox/browser/omni.ja -appDir /usr/lib64/firefox/browser {2280679f-70cf-4f2d-887a-33665cc52830} 1241 tab")
  (pmem . 0.9880603908509892)
  (rss . 238228)
  (vsize . 22750640)
  (thcount . 27)
  (nice . 0)
  (pri . 20)
  (pcpu . 0.13563582493620777)
  (etime 12 34486 810000 0)
  (start 26260 56544 920000 0)
  (ctime 0 0 0 0)
  (cstime 0 0 0 0)
  (cutime 0 0 0 0)
  (time 0 1113 460000 0)
  (stime 0 75 970000 0)
  (utime 0 1037 490000 0)
  (cmajflt . 0)
  (cminflt . 0)
  (majflt . 0)
  (minflt . 1710810)
  (tpgid . -1)
  (ttname . "")
  (sess . 1239)
  (pgrp . 1239)
  (ppid . 1241)
  (state . "R")
  (comm . "WebExtensions")
  (group . "azeem")
  (egid . 1000)
  (user . "azeem")
  (euid . 1000))
 ((args . "/usr/lib64/firefox/firefox -contentproc -childID 2 -isForBrowser -prefsLen 43016 -prefMapSize 262227 -jsInitLen 234780 -parentBuildID 20240712194222 -greomni /usr/lib64/firefox/omni.ja -appomni /usr/lib64/firefox/browser/omni.ja -appDir /usr/lib64/firefox/browser {c7fd5dcc-7f18-4677-8cd3-ff49efab3e75} 1241 tab")
  (pmem . 0.10299173743477577)
  (rss . 24832)
  (vsize . 4366000)
  (thcount . 28)
  (nice . 0)
  (pri . 20)
  (pcpu . 0.0032487989036966473)
  (etime 12 34486 770000 0)
  (start 26260 56544 960000 0)
  (ctime 0 0 0 0)
  (cstime 0 0 0 0)
  (cutime 0 0 0 0)
  (time 0 26 670000 0)
  (stime 0 6 580000 0)
  (utime 0 20 90000 0)
  (cmajflt . 0)
  (cminflt . 0)
  (majflt . 0)
  (minflt . 2708)
  (tpgid . -1)
  (ttname . "")
  (sess . 1239)
  (pgrp . 1239)
  (ppid . 1241)
  (state . "S")
  (comm . "Privileged Cont")
  (group . "azeem")
  (egid . 1000)
  (user . "azeem")
  (euid . 1000))
 ((args . "/usr/lib64/firefox/firefox -contentproc -childID 3 -isForBrowser -prefsLen 29688 -prefMapSize 262227 -jsInitLen 234780 -parentBuildID 20240712194222 -greomni /usr/lib64/firefox/omni.ja -appomni /usr/lib64/firefox/browser/omni.ja -appDir /usr/lib64/firefox/browser {77722f8d-f95f-4796-9765-00d7017c5796} 1241 tab")
  (pmem . 1.0487306202000508)
  (rss . 252856)
  (vsize . 6381408)
  (thcount . 31)
  (nice . 0)
  (pri . 20)
  (pcpu . 0.48562684148940605)
  (etime 12 34486 380000 0)
  (start 26260 56545 350000 0)
  (ctime 0 0 0 0)
  (cstime 0 0 0 0)
  (cutime 0 0 0 0)
  (time 0 3986 600000 0)
  (stime 0 651 720000 0)
  (utime 0 3334 880000 0)
  (cmajflt . 0)
  (cminflt . 0)
  (majflt . 5)
  (minflt . 1335797)
  (tpgid . -1)
  (ttname . "")
  (sess . 1239)
  (pgrp . 1239)
  (ppid . 1241)
  (state . "S")
  (comm . "Isolated Web Co")
  (group . "azeem")
  (egid . 1000)
  (user . "azeem")
  (euid . 1000)))

The percentage memory values in top for these processes are 12.1 0.2 4.0
0.4 and 4.2 so 4 times those returned by proced.

> Thanks in advance.

From discussion with Eli at the help-gnu-emacs thread I think there is
some hard coded assumption about 4KB page size in sysdep.c. E.g. here at
line 3741 

	  pmem = 4.0 * 100 * rss / procfs_get_total_memory ();

Thanks,
Rahguzar

Message #14 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Rahguzar <rahguzar <at> zohomail.eu>, Paul Eggert <eggert <at> cs.ucla.edu>
Cc: 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4;
 Proced reports incorrect results with 16KB page size
Date: Thu, 25 Jul 2024 08:06:44 +0300

> Date: Wed, 24 Jul 2024 19:22:08 +0200
> From:  Rahguzar via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs <at> gnu.org>
> 
> I am using GNU/Linux on an Apple M2 processor using the Asahi Linux
> project [1]. They provide a kernel with 16KB page size [2]. This causes proced
> to show incorrect values for percentage memory (other memory related
> attributes might also be affected). E.g. comparing with top I see that
> value reported by proced is a quarter of that related by top. 
> 
> [1] https://asahilinux.org
> [2] https://discussion.fedoraproject.org/t/switch-to-the-kernel-16k-variant/87711

Paul, it sounds like we need to multiply by getpagesize() instead of
by a fixed value of 4?  Some implementations of
system_process_attributes already do that, although not the one which
reads "/proc" on Linux.

Message #19 received at 72278-done <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Rahguzar <rahguzar <at> zohomail.eu>, 72278-done <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4; Proced reports incorrect results with 16KB page
 size
Date: Wed, 24 Jul 2024 23:38:34 -0700

[Message part 1 (text/plain, inline)]

On 2024-07-24 22:06, Eli Zaretskii wrote:
> Paul, it sounds like we need to multiply by getpagesize() instead of
> by a fixed value of 4?

Yes, that sounds right. I installed the attached patch into Emacs master 
on Savannah, and am boldly closing the bug as fixed. We can reopen it if 
I'm wrong.

Thanks, Rahguzar, for reporting the bug.

[0001-Fix-process-attributes-rss-and-pmem-on-GNU-Linux.patch (text/x-patch, attachment)]

Message #22 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Rahguzar <rahguzar <at> zohomail.eu>
Cc: 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4;
 Proced reports incorrect results with 16KB page size
Date: Thu, 25 Jul 2024 09:12:37 +0000

Rahguzar <rahguzar <at> zohomail.eu> writes:

> From discussion with Eli at the help-gnu-emacs thread I think there is
> some hard coded assumption about 4KB page size in sysdep.c. E.g. here at
> line 3741
>
> 	  pmem = 4.0 * 100 * rss / procfs_get_total_memory ();

Makes sense to me.

AFAIK, the page size can be different for different processes, so I
guess we'd need to find the page size per process in something like
/proc/<pid>/status?  Is that the right way to do it?

I'm not on a GNU/Linux machine right now so I can't check.

Message #25 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Rahguzar <rahguzar <at> zohomail.eu>
To: Stefan Kangas <stefankangas <at> gmail.com>
Cc: 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4; Proced reports incorrect results with 16KB
 page size
Date: Thu, 25 Jul 2024 11:39:06 +0200

Hi Stefan,

Stefan Kangas <stefankangas <at> gmail.com> writes:

> Rahguzar <rahguzar <at> zohomail.eu> writes:
>
>> From discussion with Eli at the help-gnu-emacs thread I think there is
>> some hard coded assumption about 4KB page size in sysdep.c. E.g. here at
>> line 3741
>>
>> 	  pmem = 4.0 * 100 * rss / procfs_get_total_memory ();
>
> Makes sense to me.
>
> AFAIK, the page size can be different for different processes, so I
> guess we'd need to find the page size per process in something like
> /proc/<pid>/status?  Is that the right way to do it?
>
> I'm not on a GNU/Linux machine right now so I can't check.

Paul Eggert's changes fixed the issue for me.

Rahguzar

Message #28 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Paul Eggert <eggert <at> cs.ucla.edu>, Eli Zaretskii <eliz <at> gnu.org>
Cc: Rahguzar <rahguzar <at> zohomail.eu>, 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4;
 Proced reports incorrect results with 16KB page size
Date: Thu, 25 Jul 2024 02:55:32 -0700

Paul Eggert <eggert <at> cs.ucla.edu> writes:

> From 42cdfe7f60ef7ae3bccb5b1e43349c3a6b54ac4a Mon Sep 17 00:00:00 2001
> From: Paul Eggert <eggert <at> cs.ucla.edu>
> Date: Wed, 24 Jul 2024 23:32:24 -0700
> Subject: [PATCH] Fix process-attributes rss and pmem on GNU/Linux
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
>
> Problem reported by Rahguzar <https://bugs.gnu.org/72278>.
> * src/sysdep.c (system_process_attributes):
> [GNU_LINUX || CYGWIN || __ANDROID__]: When computing rss and pmem,
> don’t assume pagesize is 4096; it could be greater.
> ---
>  src/sysdep.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/src/sysdep.c b/src/sysdep.c
> index d916a695155..3955d796ca5 100644
> --- a/src/sysdep.c
> +++ b/src/sysdep.c
> @@ -3548,6 +3548,7 @@ procfs_ttyname (int rdev)
>  }
>  # endif	/* GNU_LINUX || __ANDROID__ */
>
> +/* Total usable RAM in KiB.  */
>  static uintmax_t
>  procfs_get_total_memory (void)
>  {
> @@ -3737,8 +3738,13 @@ system_process_attributes (Lisp_Object pid)
>  	  attrs = Fcons (Fcons (Qnice, make_fixnum (niceness)), attrs);
>  	  attrs = Fcons (Fcons (Qthcount, INT_TO_INTEGER (thcount)), attrs);
>  	  attrs = Fcons (Fcons (Qvsize, INT_TO_INTEGER (vsize / 1024)), attrs);
> -	  attrs = Fcons (Fcons (Qrss, INT_TO_INTEGER (4 * rss)), attrs);
> -	  pmem = 4.0 * 100 * rss / procfs_get_total_memory ();
> +
> +	  /* RSS in KiB.  */
> +	  uintmax_t rssk = rss;
> +	  rssk *= getpagesize () >> 10;
> +
> +	  attrs = Fcons (Fcons (Qrss, INT_TO_INTEGER (rssk)), attrs);
> +	  pmem = 100.0 * rssk / procfs_get_total_memory ();

Thanks, that looks like an improvement.

On Linux, page sizes can vary per process ("multiple page size
support").  Should we bother with that?

I also note that the man page for getpagesize says this:

    This call first appeared in 4.2BSD.  SVr4, 4.4BSD, SUSv2.  In
    SUSv2 the getpagesize() call was labeled LEGACY, and it was
    removed in POSIX.1-2001.

>  	  if (pmem > 100)
>  	    pmem = 100;
>  	  attrs = Fcons (Fcons (Qpmem, make_float (pmem)), attrs);
> --
> 2.43.0

Message #31 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Stefan Kangas <stefankangas <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>
Cc: Rahguzar <rahguzar <at> zohomail.eu>, 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4; Proced reports incorrect results with 16KB page
 size
Date: Thu, 25 Jul 2024 07:47:56 -0700

On 2024-07-25 02:55, Stefan Kangas wrote:
> On Linux, page sizes can vary per process ("multiple page size
> support").  Should we bother with that?

I thought that support didn't change what getpagesize returns or what 
/proc/PID/stat reports, at least not in mainline Linux kernels. Am I 
missing something?


> I also note that the man page for getpagesize says this:
> 
>      This call first appeared in 4.2BSD.  SVr4, 4.4BSD, SUSv2.  In
>      SUSv2 the getpagesize() call was labeled LEGACY, and it was
>      removed in POSIX.1-2001.

Emacs uses getpagesize in other places. I suppose we should change them 
all to sysconf (_SC_PAGESIZE) on the off chance that there's a platform 
somewhere where page sizes exceed 1 GiB. I don't know of any such 
platform now, probably because so much software still uses getpagesize....

Message #34 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Paul Eggert <eggert <at> cs.ucla.edu>, Eli Zaretskii <eliz <at> gnu.org>
Cc: Rahguzar <rahguzar <at> zohomail.eu>, 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4;
 Proced reports incorrect results with 16KB page size
Date: Fri, 26 Jul 2024 18:33:09 -0400

Paul Eggert <eggert <at> cs.ucla.edu> writes:

> On 2024-07-25 02:55, Stefan Kangas wrote:
>> On Linux, page sizes can vary per process ("multiple page size
>> support").  Should we bother with that?
>
> I thought that support didn't change what getpagesize returns or what
> /proc/PID/stat reports, at least not in mainline Linux kernels. Am I
> missing something?

I don't know to be honest, which is why I'm asking.  If you think using
getpagesize is all that we need then that's good enough for me.

Message #37 received at 72278 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Stefan Kangas <stefankangas <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>
Cc: Rahguzar <rahguzar <at> zohomail.eu>, 72278 <at> debbugs.gnu.org
Subject: Re: bug#72278: 29.4; Proced reports incorrect results with 16KB page
 size
Date: Fri, 26 Jul 2024 18:40:30 -0700

On 2024-07-26 15:33, Stefan Kangas wrote:
> Paul Eggert<eggert <at> cs.ucla.edu> writes:
> 
>> On 2024-07-25 02:55, Stefan Kangas wrote:
>>> On Linux, page sizes can vary per process ("multiple page size
>>> support").  Should we bother with that?
>> I thought that support didn't change what getpagesize returns or what
>> /proc/PID/stat reports, at least not in mainline Linux kernels. Am I
>> missing something?
> I don't know to be honest, which is why I'm asking.

For recent Asahi Linux user-mode pages are always 16 KiB; see 
<https://github.com/AsahiLinux/docs/wiki/Broken-Software>.

I'd be surprised if any Linux distro lets getpagesize calls differ among 
processes running atop the same kernel. Although Linux does support huge 
pages via (for example) mmap with MAP_HUGETLB, I would expect this to 
not affect getpagesize or /proc/PID/stat. I haven't verified this by 
running code or by scouting through kernel source, though.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.