Red Hat 社のパッチを待てない子のための、CVE-2016-5195 対応手抜き手順 for EL5。
※たぶんこの手順だと、正式な RPM が公開されたら、問題なく差し替わるはず。
※というか遅くとも一週間でパッチ出るはずだし、逆にもうしないほうがいいと思う。
※EL5にはそもそもバックポートで実装されてる箇所があるため、一部機能がない。
そのためパッチをだいぶいじっているので、これで無問題か正直不明。
※一応三日ほど色々使ってみて問題なさそうなので公開。
■ビルド環境の準備
$ vi ~/.rpmmacros
※今回はマルチスレッドでビルドしても大丈夫なので %_smp_mflags はコメントアウトしておく。
%_topdir %(echo $HOME)/rpmbuild #%_smp_mflags -j1 %__arch_install_post /usr/lib/rpm/check-rpaths /usr/lib/rpm/check-buildroot %_tmpfilesdir /usr/lib/tmpfiles.d
■Kernel の SourceRPM を取得してくる
$ mkdir -p ~/rpmbuild/SRPMS/ $ cd ~/rpmbuild/SRPMS/ $ yumdownloader --source kernel-`uname -r`
※yumdownloader でダウンロードできない場合、カスタマーポータル or vault.centos.org からダウンロードする。
■SRPM の展開
$ rpm -ivh kernel-2.6.18-412.el5.src.rpm
■SOURCES ディレクトリに、[CVE-2016-5195] のパッチを配備
$ cd ~/rpmbuild/SOURCES/ $ vi fix-remove-gup_flags-FOLL_WRITE-games.patch
※此処の表示をコピペしたらエラーします。↑のリンクをお持ち帰りください。
※EL5 用に一部修正してます。オリジナルはこちら。
From: Linus Torvalds <torvalds@linux-foundation.org> commit 19be0eaffa3ac7d8eb6784ad9bdbc7d67ed8e619 upstream. This is an ancient bug that was actually attempted to be fixed once (badly) by me eleven years ago in commit 4ceb5db9757a ("Fix get_user_pages() race for write access") but that was then undone due to problems on s390 by commit f33ea7f404e5 ("fix get_user_pages bug"). In the meantime, the s390 situation has long been fixed, and we can now fix it by checking the pte_dirty() bit properly (and do it better). The s390 dirty bit was implemented in abf09bed3cce ("s390/mm: implement software dirty bits") which made it into v3.9. Earlier kernels will have to look at the page state itself. Also, the VM has become more scalable, and what used a purely theoretical race back then has become easier to trigger. To fix it, we introduce a new internal FOLL_COW flag to mark the "yes, we already did a COW" rather than play racy games with FOLL_WRITE that is very fundamental, and then use the pte dirty flag to validate that the FOLL_COW flag is still valid. Reported-and-tested-by: Phil "not Paul" Oester <kernel@linuxace.com> Acked-by: Hugh Dickins <hughd@google.com> Reviewed-by: Michal Hocko <mhocko@suse.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Kees Cook <keescook@chromium.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Willy Tarreau <w@1wt.eu> Cc: Nick Piggin <npiggin@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> [wt: s/gup.c/memory.c; s/follow_page_pte/follow_page_mask; s/faultin_page/__get_user_page] Signed-off-by: Willy Tarreau <w@1wt.eu> --- include/linux/mm.h | 1 + mm/memory.c | 14 ++++++++++++-- 2 files changed, 13 insertions(+), 2 deletions(-) diff --git a/include/linux/mm.h b/include/linux/mm.h index 53b0d70..55590f4 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1184,6 +1184,7 @@ struct page *follow_page(struct vm_area_ #define FOLL_TOUCH 0x02 /* mark page accessed */ #define FOLL_GET 0x04 /* do get_page on page */ #define FOLL_ANON 0x08 /* give ZERO_PAGE if no pgtable */ +#define FOLL_COW 0x4000 /* internal GUP flag */ #ifdef CONFIG_XEN typedef int (*pte_fn_t)(pte_t *pte, struct page *pmd_page, unsigned long addr, diff --git a/mm/memory.c b/mm/memory.c index 10cdade..2ca2ee1 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -978,6 +978,16 @@ unsigned long zap_page_range(struct vm_a EXPORT_SYMBOL_GPL(zap_page_range); /* + * FOLL_FORCE can write to even unwritable pte's, but only + * after we've gone through a COW cycle and they are dirty. + */ +static inline bool can_follow_write_pte(pte_t pte, unsigned int flags) +{ + return pte_write(pte) || + ((flags & FOLL_COW) && pte_dirty(pte)); +} + +/* * Do a quick page-table lookup for a single page. */ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, @@ -1024,7 +1034,7 @@ struct page *follow_page(struct vm_area_ pte = *ptep; if (!pte_present(pte)) goto no_page; - if ((flags & FOLL_WRITE) && !pte_write(pte)) + if ((flags & FOLL_WRITE) && !can_follow_write_pte(pte, flags)) goto unlock; page = vm_normal_page(vma, address, pte); if (unlikely(!page)) @@ -1199,7 +1209,7 @@ int get_user_pages(struct task_struct *t * subsequent page lookups as if they were reads. */ if (ret & VM_FAULT_WRITE) - foll_flags &= ~FOLL_WRITE; + foll_flags |= FOLL_COW; switch (ret & ~VM_FAULT_WRITE) { case VM_FAULT_MINOR: -- 2.8.0.rc2.1.gbe9624a
■SPEC にパッチを埋め込む
$ cd ~/rpmbuild/SPECS/ $ cp -piav kernel.spec kernel.spec.org $ vi kernel.spec
※変更点いくつかあるので、Diffを載せてます。
●Red Hat Enterprise Linux 5 [kernel-2.6.18-412.el5]
--- kernel.spec.rh 2016-08-01 21:02:08.000000000 +0900 +++ kernel.spec 2016-10-22 11:21:34.000000000 +0900 @@ -58,7 +58,7 @@ Summary: The Linux kernel (the core of t # After branching, please hardcode these values as the # %dist and %rhel tags are not reliable in rhel5-era rpm # For example dist -> .el5 and rhel -> 5 -%define dist .el5 +%define dist .el5.tsc %define rhel 5 # Values used for RHEL version info in version.h @@ -427,6 +427,7 @@ Patch3: xen-2.6.18-redhat.patch # empty final patch file to facilitate testing of kernel patches Patch99999: linux-kernel-test.patch +Patch2000: fix-remove-gup_flags-FOLL_WRITE-games.patch BuildRoot: %{_tmppath}/kernel-%{KVERREL}-root @@ -686,6 +687,7 @@ cd linux-%{KVERREL}.%{_target_cpu} %if %([ -s %{PATCH99999} ] && echo 1 || echo 0) %patch99999 -p1 %endif +%patch2000 -p1 -E cp %{SOURCE10} Documentation/ @@ -1646,6 +1648,9 @@ This is required to use SystemTap with % %endif %changelog +* Sat Oct 22 2016 Makoto Satou <makopi sgv417 jp> [2.6.18-412.el5.tsc] +- [security] mm: remove gup_flags FOLL_WRITE games from __get_user_pages() [1384344] {CVE-2016-5195} + * Mon Aug 01 2016 Alexander Gordeev <agordeev@redhat.com> [2.6.18-412.el5] - [s390] kernel: Fix tlb flushing with idte. (Hendrik Brueckner) [1350541] - [fs] Do not double unlock the dcache_lock in d_materialise_unique (Miklos Szeredi) [1198315]
●CentOS 5 [kernel-2.6.18-412.el5]
--- kernel.spec.org 2016-09-06 05:30:45.000000000 +0900 +++ kernel.spec 2016-10-22 10:37:46.000000000 +0900 @@ -58,7 +58,7 @@ Summary: The Linux kernel (the core of t # After branching, please hardcode these values as the # %dist and %rhel tags are not reliable in rhel5-era rpm # For example dist -> .el5 and rhel -> 5 -%define dist .el5 +%define dist .el5.tsc %define rhel 5 # Values used for RHEL version info in version.h @@ -427,6 +427,7 @@ Patch3: xen-2.6.18-redhat.patch # empty final patch file to facilitate testing of kernel patches Patch99999: linux-kernel-test.patch +Patch2000: fix-remove-gup_flags-FOLL_WRITE-games.patch BuildRoot: %{_tmppath}/kernel-%{KVERREL}-root @@ -686,6 +687,7 @@ cd linux-%{KVERREL}.%{_target_cpu} %if %([ -s %{PATCH99999} ] && echo 1 || echo 0) %patch99999 -p1 %endif +%patch2000 -p1 -E cp %{SOURCE10} Documentation/ @@ -1646,6 +1648,9 @@ This is required to use SystemTap with % %endif %changelog +* Sat Oct 22 2016 Makoto Satou <makopi sgv417 jp> [2.6.18-412.el5.tsc] +- [security] mm: remove gup_flags FOLL_WRITE games from __get_user_pages() [1384344] {CVE-2016-5195} + * Mon Sep 05 2016 Johnny Hughes <johnny@centos.org> [2.6.18-412.el5] - Roll in CentOS Branding
■SourceRPM の作成
$ cd ~/rpmbuild/SPECS/ $ rpmbuild -bs --define 'dist .el5.tsc' kernel.spec
→ここで「–define」で 「dist .el5.tsc」と入れることで、
本当は .el5.tsc という文字列が、RPM に刻まれます。
ただ、CentOS 5 (RHEL 5) の kernel.spec では、–define で dist 指定しても
ハードコートされているため反映されません。
そのため、kernel.spec 内の dist 部に、.tsc をハードコートをしてます。
別に入れなくてもいいんですが、自前ビルドしたものがどれなのか明示的に分かる様に
あえて入れております。
SPECファイルのChangelog や define などで .tsc とつけてるのが、自分のサインです。
■Pre-Build の実行
$ rpmbuild -bp --define 'dist .el5.tsc' kernel.spec
→ error: Failed build dependencies が出た場合、
それらをインストールしなければ、ビルドできません。
また、【exit 0】で終了しない場合、何かがおかしいです。
・Patch Fileの内容
・SPEC の記述内容
これらを確認。
■RPM のビルド
$ rpmbuild --rebuild --define 'dist .el5.tsc' ~/rpmbuild/SRPMS/kernel-2.6.18-412.el5.tsc.src.rpm $ rpmbuild --rebuild --target=noarch --define 'dist .el5.tsc' ~/rpmbuild/SRPMS/kernel-2.6.18-412.el5.tsc.src.rpm
→ここで「–define」で(ry)
■ビルドした Kernel のインストール
$ su - # cd ~admin/rpmbuild/RPMS/ # yum localupdate ./*/*2.6.18-412.el5.tsc*