Gentoo Archives: gentoo-dev

From: Tobias Klausmann <klausman@g.o>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] [rfc] enable USE=xattr by default
Date: Thu, 15 Oct 2015 11:22:31
Message-Id: 20151015112218.GA96246@skade.schwarzvogel.de
In Reply to: Re: [gentoo-dev] [rfc] enable USE=xattr by default by Jason Zaman
1 Hi!
2
3 On Thu, 15 Oct 2015, Jason Zaman wrote:
4 > Can you try this:
5 >
6 > # getfattr -d -m- /bin/ping
7 > security.capability=0sAQAAAgAgAAAAAAAAAAAAAAAAAAA=
8 > # setfattr -n user.test -v "foo" ./ping
9 > # setfattr -n user.pax.flags -v "me" ./ping
10 > # getfattr -d -m- /bin/ping
11 > security.capability=0sAQAAAgAgAAAAAAAAAAAAAAAAAAA=
12 > user.pax.flags="me"
13 > user.test="foo"
14 >
15 > If this works then something else is causing those messages and we
16 > should look into it further.
17
18 # cd /bin/
19 # getfattr -d -m- ./ping
20 # file: ping
21 security.capability=0sAQAAAgAgAAAAAAAAAAAAAAAAAAA=
22
23 # setfattr -n user.test -v "foo" ./ping
24 # setfattr -n user.pax.flags -v "me" ./ping
25 # getfattr -d -m- ./ping
26 # file: ping
27 security.capability=0sAQAAAgAgAAAAAAAAAAAAAAAAAAA=
28 user.pax.flags="me"
29 user.test="foo"
30
31 But note that emerging iputils does _not_ result in the PaX
32 messages I described. Here's an incomplete list of packages that
33 have triggered them in the past:
34
35 app-emulation/qemu-2.4.0.1
36 dev-lang/ghc-7.10.2-r1
37 dev-lang/python-3.5.0-r1
38 media-video/mpv-0.11.0
39 sys-devel/llvm-3.7.0-r2
40 www-client/firefox-bin-41.0.1
41 www-client/google-chrome-beta-46.0.2490.64_p1
42
43 I have no clue what they do differently from iputils to result in
44 those messages.
45
46 So I looked at a file mentioned during the llvm merge,
47 /usr/bin/llvm-rtdyld:
48
49 # getfattr ./llvm-rtdyld
50 # setfattr -n user.test -v "foo" ./llvm-rtdyld
51 # setfattr -n user.pax.flags -v "me" ./llvm-rtdyld
52 # getfattr -d -m- ./llvm-rtdyld
53
54 Then I remembered that the setfattr might not happen in the live
55 system, but on the tmpfs mounted on /var/tmp/portage.
56
57 So I tried it there:
58
59 # cp ./llvm-rtdyld /var/tmp/portage
60 # cd /var/tmp/portage
61 # getfattr -d -m- ./llvm-rtdyld
62 # setfattr -n user.test -v "foo" ./llvm-rtdyld
63 setfattr: ./llvm-rtdyld: Operation not supported
64 # setfattr -n user.pax.flags -v "me" ./llvm-rtdyld
65 setfattr: ./llvm-rtdyld: Operation not supported
66
67 So it's not a BTRFS problem, but one of tmpfs. So I wondered if I
68 maybe had missed to activate xattr suport for tmpfs, but no:
69
70 # zgrep -i tmpfs /proc/config.gz
71 CONFIG_DEVTMPFS=y
72 CONFIG_DEVTMPFS_MOUNT=y
73 CONFIG_TMPFS=y
74 CONFIG_TMPFS_POSIX_ACL=y
75 CONFIG_TMPFS_XATTR=y
76 #
77
78 I have no clue what's missing, since I can't find a mount option
79 or anything that I might have missed:
80
81 # mount -l -t tmpfs |grep portage
82 none on /var/tmp/portage type tmpfs (rw,relatime,size=8388608k)
83
84 On a whim, I tried user_xattr, something older ext* FSs support,
85 but alas:
86
87 # mount -o remount,user_xattr /var/tmp/portage
88 mount: /var/tmp/portage not mounted or bad option
89
90 In some cases useful info is found in syslog - try
91 dmesg | tail or so.
92 # dmesg|tail -n1
93 [ 8623.541053] tmpfs: No value for mount option 'user_xattr'
94
95 Providing =something just results in this:
96 [ 8701.215926] tmpfs: Bad mount option user_xattr
97
98 So I'm fresh out of ideas.
99
100 Regards,
101 Tobias
102
103
104
105
106
107 --
108 "Sendmail is the sort of tool that gave UNIX its bad reputation."
109 -- _System Performance Tuning_

Replies

Subject Author
Re: [gentoo-dev] [rfc] enable USE=xattr by default Rich Freeman <rich0@g.o>