登录
注册
开源
企业版
高校版
搜索
帮助中心
使用条款
关于我们
开源
企业版
高校版
私有云
模力方舟
AI 队友
登录
注册
轻量养虾,开箱即用!低 Token + 稳定算力,Gitee & 模力方舟联合出品的 PocketClaw 正式开售!点击了解详情
代码拉取完成,页面将自动刷新
仓库状态说明
捐赠
捐赠前请先登录
取消
前往登录
扫描微信二维码支付
取消
支付完成
支付提示
将跳转至支付宝完成支付
确定
取消
Watch
不关注
关注所有动态
仅关注版本发行动态
关注但不提醒动态
13
Star
20
Fork
180
src-openEuler
/
glibc
关闭
代码
Issues
9
Pull Requests
0
Wiki
统计
流水线
服务
JavaDoc
PHPDoc
质量分析
Jenkins for Gitee
腾讯云托管
腾讯云 Serverless
悬镜安全
阿里云 SAE
Codeblitz
SBOM
开发画像分析
我知道了,不再自动展开
更新失败,请稍后重试!
移除标识
内容风险标识
本任务被
标识为内容中包含有代码安全 Bug 、隐私泄露等敏感信息,仓库外成员不可访问
回合上游社区补丁,补丁数量:6
已完成
#I9NACB
缺陷
peulerosci
创建于
2024-05-08 09:32
<h3>优先级:</h3><p>次要</p><p></p><h3>原因分析:</h3><p></p><p>【是否新引入rpm包,是否进行了安全编译选项扫描】</p><p></p><p>否</p><p></p><p>【定位分析】</p><p></p><p>例行回合社区bugfix补丁。</p><p></p><table ><tr><td colSpan="1" rowSpan="1">软件包</td><td colSpan="1" rowSpan="1">CommitId</td><td colSpan="1" rowSpan="1">描述</td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">2aa0974d2573441bffd596b07bff8698b1f2f18c</td><td colSpan="1" rowSpan="1">elf: ldconfig should skip temporary files created by package managers<br/><br/>This avoids crashes due to partially written files, after a package<br/>update is interrupted.<br/><br/>Reviewed-by: Adhemerval Zanella </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">cfb5a97a93ea656e3b2263e42142a4032986d9ba</td><td colSpan="1" rowSpan="1">ldconfig: Fixes for skipping temporary files.<br/><br/>Arguments to a memchr call were swapped, causing incorrect skipping<br/>of files.<br/><br/>Files related to dpkg have different names: they actually end in<br/>.dpkg-new and .dpkg-tmp, not .tmp as I mistakenly assumed.<br/><br/>Fixes commit 2aa0974d2573441bffd59 (elf: ldconfig should skip<br/>temporary files created by package managers).</td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">980450f12685326729d63ff72e93a996113bf073</td><td colSpan="1" rowSpan="1">elf: Add TLS modid reuse test for bug 29039<br/><br/>This is a minimal regression test for bug 29039 which only affects<br/>targets with TLSDESC and a reproducer requires that<br/><br/>1) Have modid gaps (closed modules) with old generation.<br/>2) Update a DTV to a newer generation (needs a newer dlopen).<br/>3) But do not update the closed gap entry in that DTV.<br/>4) Reuse the modid gap for a new module (another dlopen).<br/>5) Use dynamic TLSDESC in that new module with old generation (bug).<br/>6) Access TLS via this TLSDESC and the now outdated DTV.<br/><br/>However step (3) in practice rarely happens: during DTV update the<br/>entries for closed modids are initialized to unallocated and then<br/>dynamic TLSDESC calls __tls_get_addr independently of its generation.<br/>The only exception to this is DTV setup at thread creation (gaps are<br/>initialized to NULL instead of unallocated) or DTV resize where the<br/>gap entries are outside the previous DTV array (again NULL instead<br/>of unallocated, and this requires loading > DTV_SURPLUS modules).<br/><br/>So the bug can only cause NULL (+ offset) dereference, not use after<br/>free. And the easiest way to get (3) is via thread creation.<br/><br/>Note that step (5) requires that the newly loaded module has larger<br/>TLS than the remaining optional static TLS. And for (6) there cannot<br/>be other TLS access or dlopen in the thread that updates the DTV.<br/><br/>Tested on aarch64-linux-gnu.<br/><br/>Reviewed-by: Adhemerval Zanella </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">3921c5b40f293c57cb326f58713c924b0662ef59</td><td colSpan="1" rowSpan="1">elf: Fix TLS modid reuse generation assignment (BZ 29039)<br/><br/>_dl_assign_tls_modid() assigns a slotinfo entry for a new module, but<br/>does *not* do anything to the generation counter. The first time this<br/>happens, the generation is zero and map_generation() returns the current<br/>generation to be used during relocation processing. However, if<br/>a slotinfo entry is later reused, it will already have a generation<br/>assigned. If this generation has fallen behind the current global max<br/>generation, then this causes an obsolete generation to be assigned<br/>during relocation processing, as map_generation() returns this<br/>generation if nonzero. _dl_add_to_slotinfo() eventually resets the<br/>generation, but by then it is too late. This causes DTV updates to be<br/>skipped, leading to NULL or broken TLS slot pointers and segfaults.<br/><br/>Fix this by resetting the generation to zero in _dl_assign_tls_modid(),<br/>so it behaves the same as the first time a slot is assigned.<br/>_dl_add_to_slotinfo() will still assign the correct static generation<br/>later during module load, but relocation processing will no longer use<br/>an obsolete generation.<br/><br/>Note that slotinfo entry (aka modid) reuse typically happens after a<br/>dlclose and only TLS access via dynamic tlsdesc is affected. Because<br/>tlsdesc is optimized to use the optional part of static TLS, dynamic<br/>tlsdesc can be avoided by increasing the glibc.rtld.optional_static_tls<br/>tunable to a large enough value, or by LD_PRELOAD-ing the affected<br/>modules.<br/><br/>Fixes bug 29039.<br/><br/>Reviewed-by: Szabolcs Nagy </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">434eca873f14f618d6c2279b54fb809fb56f2c50</td><td colSpan="1" rowSpan="1">elf: Fix _dl_debug_vdprintf to work before self-relocation<br/><br/>The strlen might trigger and invalid GOT entry if it used before<br/>the process is self-relocated (for instance on dl-tunables if any<br/>error occurs).<br/><br/>For i386, _dl_writev with PIE requires to use the old 'int $0x80'<br/>syscall mode because the calling the TLS register (gs) is not yet<br/>initialized.<br/><br/>Checked on x86_64-linux-gnu.<br/>Reviewed-by: Siddhesh Poyarekar </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">1cce91d8aed5c3eca2b6f47767c82d9ed3e9e33f</td><td colSpan="1" rowSpan="1">elf: Check objname before calling fatal_error<br/><br/>_dl_signal_error may be called with objname == NULL. _dl_exception_create<br/>checks objname == NULL. But fatal_error doesn't. Check objname before<br/>calling fatal_error. This fixes BZ #31596.<br/>Reviewed-by: Sunil K Pandey </td></tr></table><p></p><p>【对外部是否有影响】</p><p></p><p>否</p><p></p><p></p>
<h3>优先级:</h3><p>次要</p><p></p><h3>原因分析:</h3><p></p><p>【是否新引入rpm包,是否进行了安全编译选项扫描】</p><p></p><p>否</p><p></p><p>【定位分析】</p><p></p><p>例行回合社区bugfix补丁。</p><p></p><table ><tr><td colSpan="1" rowSpan="1">软件包</td><td colSpan="1" rowSpan="1">CommitId</td><td colSpan="1" rowSpan="1">描述</td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">2aa0974d2573441bffd596b07bff8698b1f2f18c</td><td colSpan="1" rowSpan="1">elf: ldconfig should skip temporary files created by package managers<br/><br/>This avoids crashes due to partially written files, after a package<br/>update is interrupted.<br/><br/>Reviewed-by: Adhemerval Zanella </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">cfb5a97a93ea656e3b2263e42142a4032986d9ba</td><td colSpan="1" rowSpan="1">ldconfig: Fixes for skipping temporary files.<br/><br/>Arguments to a memchr call were swapped, causing incorrect skipping<br/>of files.<br/><br/>Files related to dpkg have different names: they actually end in<br/>.dpkg-new and .dpkg-tmp, not .tmp as I mistakenly assumed.<br/><br/>Fixes commit 2aa0974d2573441bffd59 (elf: ldconfig should skip<br/>temporary files created by package managers).</td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">980450f12685326729d63ff72e93a996113bf073</td><td colSpan="1" rowSpan="1">elf: Add TLS modid reuse test for bug 29039<br/><br/>This is a minimal regression test for bug 29039 which only affects<br/>targets with TLSDESC and a reproducer requires that<br/><br/>1) Have modid gaps (closed modules) with old generation.<br/>2) Update a DTV to a newer generation (needs a newer dlopen).<br/>3) But do not update the closed gap entry in that DTV.<br/>4) Reuse the modid gap for a new module (another dlopen).<br/>5) Use dynamic TLSDESC in that new module with old generation (bug).<br/>6) Access TLS via this TLSDESC and the now outdated DTV.<br/><br/>However step (3) in practice rarely happens: during DTV update the<br/>entries for closed modids are initialized to unallocated and then<br/>dynamic TLSDESC calls __tls_get_addr independently of its generation.<br/>The only exception to this is DTV setup at thread creation (gaps are<br/>initialized to NULL instead of unallocated) or DTV resize where the<br/>gap entries are outside the previous DTV array (again NULL instead<br/>of unallocated, and this requires loading > DTV_SURPLUS modules).<br/><br/>So the bug can only cause NULL (+ offset) dereference, not use after<br/>free. And the easiest way to get (3) is via thread creation.<br/><br/>Note that step (5) requires that the newly loaded module has larger<br/>TLS than the remaining optional static TLS. And for (6) there cannot<br/>be other TLS access or dlopen in the thread that updates the DTV.<br/><br/>Tested on aarch64-linux-gnu.<br/><br/>Reviewed-by: Adhemerval Zanella </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">3921c5b40f293c57cb326f58713c924b0662ef59</td><td colSpan="1" rowSpan="1">elf: Fix TLS modid reuse generation assignment (BZ 29039)<br/><br/>_dl_assign_tls_modid() assigns a slotinfo entry for a new module, but<br/>does *not* do anything to the generation counter. The first time this<br/>happens, the generation is zero and map_generation() returns the current<br/>generation to be used during relocation processing. However, if<br/>a slotinfo entry is later reused, it will already have a generation<br/>assigned. If this generation has fallen behind the current global max<br/>generation, then this causes an obsolete generation to be assigned<br/>during relocation processing, as map_generation() returns this<br/>generation if nonzero. _dl_add_to_slotinfo() eventually resets the<br/>generation, but by then it is too late. This causes DTV updates to be<br/>skipped, leading to NULL or broken TLS slot pointers and segfaults.<br/><br/>Fix this by resetting the generation to zero in _dl_assign_tls_modid(),<br/>so it behaves the same as the first time a slot is assigned.<br/>_dl_add_to_slotinfo() will still assign the correct static generation<br/>later during module load, but relocation processing will no longer use<br/>an obsolete generation.<br/><br/>Note that slotinfo entry (aka modid) reuse typically happens after a<br/>dlclose and only TLS access via dynamic tlsdesc is affected. Because<br/>tlsdesc is optimized to use the optional part of static TLS, dynamic<br/>tlsdesc can be avoided by increasing the glibc.rtld.optional_static_tls<br/>tunable to a large enough value, or by LD_PRELOAD-ing the affected<br/>modules.<br/><br/>Fixes bug 29039.<br/><br/>Reviewed-by: Szabolcs Nagy </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">434eca873f14f618d6c2279b54fb809fb56f2c50</td><td colSpan="1" rowSpan="1">elf: Fix _dl_debug_vdprintf to work before self-relocation<br/><br/>The strlen might trigger and invalid GOT entry if it used before<br/>the process is self-relocated (for instance on dl-tunables if any<br/>error occurs).<br/><br/>For i386, _dl_writev with PIE requires to use the old 'int $0x80'<br/>syscall mode because the calling the TLS register (gs) is not yet<br/>initialized.<br/><br/>Checked on x86_64-linux-gnu.<br/>Reviewed-by: Siddhesh Poyarekar </td></tr><tr><td colSpan="1" rowSpan="1">openEuler:glibc</td><td colSpan="1" rowSpan="1">1cce91d8aed5c3eca2b6f47767c82d9ed3e9e33f</td><td colSpan="1" rowSpan="1">elf: Check objname before calling fatal_error<br/><br/>_dl_signal_error may be called with objname == NULL. _dl_exception_create<br/>checks objname == NULL. But fatal_error doesn't. Check objname before<br/>calling fatal_error. This fixes BZ #31596.<br/>Reviewed-by: Sunil K Pandey </td></tr></table><p></p><p>【对外部是否有影响】</p><p></p><p>否</p><p></p><p></p>
评论 (
4
)
登录
后才可以发表评论
状态
已完成
待办的
已挂起
修复中
已确认
已完成
已验收
已取消
负责人
未设置
桐小哥
tong_1001
负责人
协作者
+负责人
+协作者
标签
sig/Computing
未设置
项目
未立项任务
未立项任务
里程碑
未关联里程碑
未关联里程碑
Pull Requests
未关联
未关联
关联的 Pull Requests 被合并后可能会关闭此 issue
分支
未关联
分支 (
-
)
标签 (
-
)
开始日期   -   截止日期
-
置顶选项
不置顶
置顶等级:高
置顶等级:中
置顶等级:低
优先级
不指定
严重
主要
次要
不重要
预计工期
(小时)
参与者(3)
1
https://gitee.com/src-openeuler/glibc.git
git@gitee.com:src-openeuler/glibc.git
src-openeuler
glibc
glibc
点此查找更多帮助
搜索帮助
Git 命令在线学习
如何在 Gitee 导入 GitHub 仓库
Git 仓库基础操作
企业版和社区版功能对比
SSH 公钥设置
如何处理代码冲突
仓库体积过大,如何减小?
如何找回被删除的仓库数据
Gitee 产品配额说明
GitHub仓库快速导入Gitee及同步更新
什么是 Release(发行版)
将 PHP 项目自动发布到 packagist.org
评论
仓库举报
回到顶部
登录提示
该操作需登录 Gitee 帐号,请先登录后再操作。
立即登录
没有帐号,去注册