Re: [OE-core] [dunfell][PATCH v2] vim: fix 2021-3796
From: Steve Sakoman <hidden>
Date: 2021-10-25 18:51:12
On Mon, Oct 25, 2021 at 7:59 AM Steve Sakoman via lists.openembedded.org [off-list ref] wrote:
On Mon, Oct 25, 2021 at 6:43 AM Steve Sakoman via lists.openembedded.org [off-list ref] wrote:quoted
On Sun, Oct 24, 2021 at 9:29 PM Minjae Kim [off-list ref] wrote:quoted
vim is vulnerable to Use After Free Problem: Checking first character of url twice. reference: https://github.com/vim/vim/commit/35a9a00afcb20897d462a766793ff45534810dc3 Signed-off-by: Minjae Kim <redacted> --- .../vim/files/CVE-2021-3796.patch | 50 +++++++++++++++++++ 1 file changed, 50 insertions(+) create mode 100644 meta/recipes-support/vim/files/CVE-2021-3796.patchYou don't seem to be adding the patch to the SRC_URI in the recipe!I fixed that issue, but now once again the patch file fails to apply: stdio: ERROR: vim-8.2-r0 do_patch: Applying patch 'CVE-2021-3796.patch' on target directory '/home/pokybuild/yocto-worker/no-x11/build/build/tmp/work/core2-64-poky-linux/vim/8.2-r0/git' stdio: ERROR: Logfile of failure stored in: /home/pokybuild/yocto-worker/no-x11/build/build/tmp/work/core2-64-poky-linux/vim/8.2-r0/temp/log.do_patch.7612 stdio: ERROR: Task (/home/pokybuild/yocto-worker/no-x11/build/meta/recipes-support/vim/vim_8.2.bb:do_patch) failed with exit code '1'
Perhaps you are having some issues with your mailer that is corrupting the patch? I see that the master version is also having problems. Do you have a git repo I could try pulling from? Steve
Stevequoted
Stevequoted
diff --git a/meta/recipes-support/vim/files/CVE-2021-3796.patch b/meta/recipes-support/vim/files/CVE-2021-3796.patch new file mode 100644 index 0000000000..666bd5c48b --- /dev/null +++ b/meta/recipes-support/vim/files/CVE-2021-3796.patch@@ -0,0 +1,50 @@ +From 6d02e1429771c00046b48f26e53ca4123c3ce4e1 Mon Sep 17 00:00:00 2001 +From: Bram Moolenaar <Bram@vim.org> +Date: Fri, 24 Sep 2021 16:01:09 +0800 +Subject: [PATCH] patch 8.2.3428: using freed memory when replacing + +Problem: Using freed memory when replacing. (Dhiraj Mishra) +Solution: Get the line pointer after calling ins_copychar(). + +Upstream-Status: Backport [https://github.com/vim/vim/commit/35a9a00afcb20897d462a766793ff45534810dc3] +CVE: CVE-2021-3796 + +Signed-off-by: Minjae Kim <flowergom@gmail.com> +--- + src/normal.c | 10 +++++++--- + 1 file changed, 7 insertions(+), 3 deletions(-) + +diff --git a/src/normal.c b/src/normal.c +index c4963e621..305b514bc 100644 +--- a/src/normal.c ++++ b/src/normal.c +@@ -5009,19 +5009,23 @@ nv_replace(cmdarg_T *cap) + { + /* + * Get ptr again, because u_save and/or showmatch() will have +- * released the line. At the same time we let know that the +- * line will be changed. ++ * released the line. This may also happen in ins_copychar(). ++ * At the same time we let know that the line will be changed. + */ +- ptr = ml_get_buf(curbuf, curwin->w_cursor.lnum, TRUE); + if (cap->nchar == Ctrl_E || cap->nchar == Ctrl_Y) + { + int c = ins_copychar(curwin->w_cursor.lnum + + (cap->nchar == Ctrl_Y ? -1 : 1)); ++ ++ ptr = ml_get_buf(curbuf, curwin->w_cursor.lnum, TRUE); + if (c != NUL) + ptr[curwin->w_cursor.col] = c; + } + else ++ { ++ ptr = ml_get_buf(curbuf, curwin->w_cursor.lnum, TRUE); + ptr[curwin->w_cursor.col] = cap->nchar; ++ } + if (p_sm && msg_silent == 0) + showmatch(cap->nchar); + ++curwin->w_cursor.col; +-- +2.17.1 + --2.30.1 (Apple Git-130)-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#157353): https://lists.openembedded.org/g/openembedded-core/message/157353 Mute This Topic: https://lists.openembedded.org/mt/86572003/3620601 Group Owner: openembedded-core+owner@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [steve@sakoman.com] -=-=-=-=-=-=-=-=-=-=-=-