Debian-LTS has issued an advisory today (July 28): http://lwn.net/Alerts/652632/ This is one of the CVEs from the recent Java updates. I would imagine that it's fixed in 55.1, which Cauldron was just updated to. Other versions may need to be fixed. Reproducible: Steps to Reproduce:
Fix was already committed to debian unstable: https://launchpad.net/ubuntu/+source/icu/52.1-10
CC: (none) => cjw
(In reply to Christiaan Welvaart from comment #1) > Fix was already committed to debian unstable: > https://launchpad.net/ubuntu/+source/icu/52.1-10 Thanks! So it wasn't fixed in 55.1 :O Patch now committed in Mageia 4, Mageia 5, and Cauldron.
Whiteboard: (none) => MGA4TOO
Patched packages uploaded for Mageia 4, Mageia 5, and Cauldron. Advisory: ======================== Updated icu packages fixes security vulnerability: It was discovered that ICU Layout Engine was missing multiple boundary checks. These could lead to buffer overflows memory corruption. A specially crafted file could cause an application using ICU to parse untrusted font files to crash and, possibly, execute arbitrary code (CVE-2015-4760). References: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2015-4760 https://bugzilla.redhat.com/show_bug.cgi?id=1242447 http://lwn.net/Alerts/652632/ ======================== Updated packages in core/updates_testing: ======================== icu-52.1-2.5.mga4 icu-data-52.1-2.5.mga4 icu-doc-52.1-2.5.mga4 libicu52-52.1-2.5.mga4 libicu-devel-52.1-2.5.mga4 icu-53.1-12.2.mga5 icu53-data-53.1-12.2.mga5 icu-doc-53.1-12.2.mga5 libicu53-53.1-12.2.mga5 libicu-devel-53.1-12.2.mga5 from SRPMS: icu-52.1-2.5.mga4.src.rpm icu-53.1-12.2.mga5.src.rpm
Assignee: bugsquad => qa-bugsSeverity: normal => critical
MGA4-32 on Acer D620 Xfce No installation issues Ref to bug 1647 Comment 6 for test case At CLI: $ strace -o thund thunderbird not OK. I can insert special characters, but only the ones without a basic letter selection. In the selection box, the lower left button, to choose a "normal" letter as the base for the special character, is not accessible, so I cannot use characters like "ó". and $ grep icu thund open("/lib/libicui18n.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicuuc.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicudata.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/usr/lib/icu/icuplugins52.txt", O_RDONLY) = -1 ENOENT (No such file or directory) stat64("/usr/share/icu/52.1/icudt52l/cnvalias.icu", 0xbff197a0) = -1 ENOENT (No such file or directory) stat64("/usr/share/icu/52.1/icudt52l.dat", {st_mode=S_IFREG|0644, st_size=23505296, ...}) = 0 open("/usr/share/icu/52.1/icudt52l.dat", O_RDONLY) = 17 read(66, "y/RT\nridicule/MGDS\nridiculous/PY"..., 4096) = 4096 Something missing ???
CC: (none) => herman.viaene
In VirtualBox, M4, KDE, 32-bit Install strace thunderbird Package(s) under test: icu default install of icu icu-data libicu52 [root@localhost wilcal]# urpmi icu Package icu-52.1-2.4.mga4.i586 is already installed [root@localhost wilcal]# urpmi icu-data Package icu-data-52.1-2.4.mga4.noarch is already installed [root@localhost wilcal]# urpmi libicu52 Package libicu52-52.1-2.4.mga4.i586 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib/libicui18n.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicuuc.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicudata.so.52", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine install icu from updates_testing [root@localhost wilcal]# urpmi icu Package icu-52.1-2.5.mga4.i586 is already installed [root@localhost wilcal]# urpmi icu-data Package icu-data-52.1-2.5.mga4.noarch is already installed [root@localhost wilcal]# urpmi libicu52 Package libicu52-52.1-2.5.mga4.i586 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib/libicui18n.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicuuc.so.52", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicudata.so.52", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine
CC: (none) => wilcal.int
Advisory committed to svn. Based on comment 5 adding MGA4-64-OK to the whiteboard.
CC: (none) => davidwhodginsWhiteboard: MGA4TOO => MGA4TOO advisory MGA4-64-OK
(In reply to Herman Viaene from comment #4) > not OK. I can insert special characters, but only the ones without a basic > letter selection. In the selection box, the lower left button, to choose a > "normal" letter as the base for the special character, is not accessible, so > I cannot use characters like "ó". Herman this is a security update not a fix for a functional bug. If there is a specific problem with icu and one of the apps that use it please raise a separate bug. Thanks.
Whiteboard: MGA4TOO advisory MGA4-64-OK => MGA4TOO advisory MGA4-32-OK MGA4-64-OK
(In reply to William Kenney from comment #7) > (In reply to Herman Viaene from comment #4) > > > not OK. I can insert special characters, but only the ones without a basic > > letter selection. In the selection box, the lower left button, to choose a > > "normal" letter as the base for the special character, is not accessible, so > > I cannot use characters like "ó". > > Herman this is a security update not a fix for a functional bug. If there > is a specific problem with icu and one of the apps that use it please raise > a separate bug. Thanks. You're telling that without even asking if it's a regression?
(In reply to Samuel VERSCHELDE from comment #8) > (In reply to William Kenney from comment #7) > > (In reply to Herman Viaene from comment #4) > > > > > not OK. I can insert special characters, but only the ones without a basic > > > letter selection. In the selection box, the lower left button, to choose a > > > "normal" letter as the base for the special character, is not accessible, so > > > I cannot use characters like "ó". > > > > Herman this is a security update not a fix for a functional bug. If there > > is a specific problem with icu and one of the apps that use it please raise > > a separate bug. Thanks. > > You're telling that without even asking if it's a regression? Is it a regression?
In VirtualBox, M5, KDE, 32-bit Install strace thunderbird Package(s) under test: icu default install of icu icu53-data libicu53 [root@localhost wilcal]# urpmi icu Package icu-53.1-12.1.mga5.i586 is already installed [root@localhost wilcal]# urpmi icu53-data Package icu53-data-53.1-12.1.mga5.noarch is already installed [root@localhost wilcal]# urpmi libicu53 Package libicu53-53.1-12.1.mga5.i586 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib/libicui18n.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicuuc.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicudata.so.53", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine install icu from updates_testing [root@localhost wilcal]# urpmi icu Package icu-53.1-12.2.mga5.i586 is already installed [root@localhost wilcal]# urpmi icu53-data Package icu53-data-53.1-12.2.mga5.noarch is already installed [root@localhost wilcal]# urpmi libicu53 Package libicu53-53.1-12.2.mga5.i586 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib/libicui18n.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicuuc.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib/libicudata.so.53", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine
Whiteboard: MGA4TOO advisory MGA4-32-OK MGA4-64-OK => MGA4TOO advisory MGA4-32-OK MGA4-64-OK MGA5-32-OK
In VirtualBox, M5, KDE, 64-bit Install strace thunderbird Package(s) under test: icu default install of icu icu53-data lib64icu53 [root@localhost wilcal]# urpmi icu Package icu-53.1-12.1.mga5.x86_64 is already installed [root@localhost wilcal]# urpmi icu53-data Package icu53-data-53.1-12.1.mga5.noarch is already installed [root@localhost wilcal]# urpmi lib64icu53 Package lib64icu53-53.1-12.1.mga5.x86_64 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib64/libicui18n.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib64/libicuuc.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib64/libicudata.so.53", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine install icu from updates_testing [root@localhost wilcal]# urpmi icu Package icu-53.1-12.2.mga5.x86_64 is already installed [root@localhost wilcal]# urpmi icu53-data Package icu53-data-53.1-12.2.mga5.noarch is already installed [root@localhost wilcal]# urpmi lib64icu53 Package lib64icu53-53.1-12.2.mga5.x86_64 is already installed LibreOffice -> Insert -> Special Character works fine strace -o strace.out thunderbird ( worked ) [wilcal@localhost ~]$ grep icu strace.out open("/lib64/libicui18n.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib64/libicuuc.so.53", O_RDONLY|O_CLOEXEC) = 4 open("/lib64/libicudata.so.53", O_RDONLY|O_CLOEXEC) = 4.....etc Displays fine. icu works fine
Whiteboard: MGA4TOO advisory MGA4-32-OK MGA4-64-OK MGA5-32-OK => MGA4TOO advisory MGA4-32-OK MGA4-64-OK MGA5-32-OK MGA5-64-OK
Unless there's a regression caused by this update can we move this on?
Trying to confirm regression now, I need some time to cover possible cases.
(In reply to Herman Viaene from comment #13) > Trying to confirm regression now, I need some time to cover possible cases. Go for it Herman. Just need to push critical security updates ASAP. Thanks.
MGA4-32 on Acer D620 Xfce Same outcome with icu-52.1-2.4 as in Comment 4. Could Xfce be the difference with Bill's results.
Keywords: (none) => validated_updateCC: (none) => sysadmin-bugs
(In reply to Herman Viaene from comment #15) > MGA4-32 on Acer D620 Xfce > Same outcome with icu-52.1-2.4 as in Comment 4. > Could Xfce be the difference with Bill's results. The real question is should we hold up this security update and open a separate bug? What do you think David?
Not a regression, so yeah don't hold this one since it's critical. A separate bug can be filed, but honestly it's unlikely that we'll be able to fix it unless someone can locate an upstream patch that fixes it.
This is good to go. Validated. Thanks David.
This update works fine. Testing complete for mga4&5 32-bit & 64-bit Validating the update. Could someone from the sysadmin team push to updates. Thanks
An update for this issue has been pushed to Mageia Updates repository. http://advisories.mageia.org/MGASA-2015-0297.html
Status: NEW => RESOLVEDResolution: (none) => FIXED