Server crash on receiving query

[Scott Bradner]

thanks for figuring this out 

Scott

> On Nov 4, 2024, at 11:35 AM, Ondřej Surý <ondrej at isc.org> wrote:
> 
> I spent last two hours crashing my computer while bisecting the issue,
> and you would not have guessed the commit that causes this:
> 
> commit f063ee276e4a7f6cfccbefa969e8db8df952348b
> Author: Ondřej Surý <ondrej at isc.org>
> Date:   2022-05-18 14:10:58 +0200
> 
>    Use libuv CFLAGS and LIBS when checking for features
> 
>    When autoconf was checking for libuv features, the LIBUV_CFLAGS was not
>    added to CFLAGS and LIBUV_LIBS to LIBS which resulted in false
>    negatives.
> 
>    Use AX_SAVE_FLAGS and AX_RESTORE_FLAGS to temporarily add LIBUV_CFLAGS
>    and LIBUV_LIBS to their respective variables.
> 
> diff --git a/configure.ac b/configure.ac
> index 8dc5aca4c4..84108c9820 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -557,6 +557,10 @@ AC_MSG_CHECKING([for libuv])
> PKG_CHECK_MODULES([LIBUV], [libuv >= 1.0.0], [],
>   [AC_MSG_ERROR([libuv not found])])
> 
> +AX_SAVE_FLAGS([libuv])
> +CFLAGS="$CFLAGS $LIBUV_CFLAGS"
> +LIBS="$LIBS $LIBUV_LIBS"
> +
> # libuv recvmmsg support
> AC_CHECK_DECLS([UV_UDP_MMSG_FREE, UV_UDP_MMSG_CHUNK], [], [], [[#include <uv.h>]])
> AC_MSG_CHECKING([whether struct msghdr uses padding for alignment])
> @@ -571,6 +575,8 @@ AC_COMPILE_IFELSE([AC_LANG_PROGRAM([#include <sys/socket.h>],
> # libuv recverr support
> AC_CHECK_DECLS([UV_UDP_LINUX_RECVERR], [], [], [[#include <uv.h>]])
> 
> +AX_RESTORE_FLAGS([libuv])
> +
> # [pairwise: --enable-doh --with-libnghttp2=auto, --enable-doh --with-libnghttp2=yes, --disable-doh]
> AC_ARG_ENABLE([doh],
>       [AS_HELP_STRING([--disable-doh], [enable DNS over HTTPS, requires libnghttp2 (default=yes)])],
> 
> The bug has been reported to Apple now as a security issue.
> 
> Ondrej
> --
> Ondřej Surý (He/Him)
> ondrej at isc.org
> 
> My working hours and your working hours may be different. Please do not feel obligated to reply outside your normal working hours.
> 
>> On 4. 11. 2024, at 15:03, Scott Bradner <sob at sobco.com> wrote:
>> 
>> for what its worth - I am NOT framing this as a BIND 9 fault  as in BIND should
>> change what it is doing 
>> 
>> but BIND 9 is doing something that Apple does not like  and it would
>> help Apple to know what that is
>> 
>> Scott
>> 
>>> On Nov 4, 2024, at 9:59 AM, Ondřej Surý <ondrej at isc.org> wrote:
>>> 
>>> Please stop framing this as this is BIND 9 fault. I’ll report this to Apple as I can reproduce this on my machine too on unprivileged account.
>>> 
>>> Ondrej
>>> --
>>> Ondřej Surý — ISC (He/Him)
>>> 
>>> My working hours and your working hours may be different. Please do not feel obligated to reply outside your normal working hours.
>>> 
>>>> On 4. 11. 2024, at 14:23, Scott Bradner <sob at sobco.com> wrote:
>>>> 
>>>> does named so something in buffering previous queries that would
>>>> trigger a different process on the third query?  (seems strange that
>>>> its reboot from another computer is repeatable on the 3rd query for the
>>>> same site?)
>>>> 
>>>> note that many queries from another computer are just fine as long as its
>>>> not for the same site
>>>> 
>>>> Scott
>>>> 
>>>>>> On Nov 4, 2024, at 9:16 AM, Ondřej Surý <ondrej at isc.org> wrote:
>>>>>> 
>>>>>> maybe named is using some special system call
>>>>> 
>>>>> Named is not using anything special. All the stuff is just libraries and standard API.
>>>>> 
>>>>> Ondrej
>>>>> --
>>>>> Ondřej Surý — ISC (He/Him)
>>>>> 
>>>>> My working hours and your working hours may be different. Please do not feel obligated to reply outside your normal working hours.
>>>>> 
>>>>>>> On 4. 11. 2024, at 14:14, Scott Bradner <sob at sobco.com> wrote:
>>>>>> 
>>>>>> maybe named is using some special system call
>>>>> 
>>>> 
>>> 
>> 
>