From: Jan Beulich <jbeulich@suse.com>
To: "Marek Marczykowski-Górecki" <marmarek@invisiblethingslab.com>,
"Frediano Ziglio" <frediano.ziglio@cloud.com>
Cc: xen-devel@lists.xenproject.org,
"Daniel P. Smith" <dpsmith@apertussolutions.com>
Subject: Re: [PATCH v2] xen/efi: Fix crash with initial empty EFI options
Date: Fri, 29 Aug 2025 09:17:31 +0200 [thread overview]
Message-ID: <e8e3b2c7-39ac-4c3a-a7ff-e2b579f44d2c@suse.com> (raw)
In-Reply-To: <aLEnVSCil2JtDiFY@mail-itl>
On 29.08.2025 06:06, Marek Marczykowski-Górecki wrote:
> On Tue, Jul 08, 2025 at 02:56:58PM +0100, Frediano Ziglio wrote:
>> EFI code path split options from EFI LoadOptions fields in 2
>> pieces, first EFI options, second Xen options.
>> "get_argv" function is called first to get the number of arguments
>> in the LoadOptions, second, after allocating enough space, to
>> fill some "argc"/"argv" variable. However the first parsing could
>> be different from second as second is able to detect "--" argument
>> separator. So it was possible that "argc" was bigger that the "argv"
>> array leading to potential buffer overflows, in particular
>> a string like "-- a b c" would lead to buffer overflow in "argv"
>> resulting in crashes.
>
> I wouldn't call it "buffer overflow" - the argv array is big enough
> here. But if there is "--" in cmdline, it has fewer than argc elements
> initialized. If there is at least one efi option (IOW, "--" is not the
> first one), the sentinel NULL inserted by get_argv() will prevent
> reading past the initialized part. But if "--" is the first one, the
> NULL is inserted into argv[0], which is skipped by the loop in
> efi_start(). Which makes the loop go beyond initialized part of argv
> (crash happens even before it goes beyond end of argv allocation).
>
> So, maybe change it to: bigger than the initialized portion of "argv"
> array, leading to potential uninitialized pointer dereference, ...?
>
>> Using EFI shell is possible to pass any kind of string in
>> LoadOptions.
>>
>> Fixes: bf6501a62e80 ("x86-64: EFI boot code")
>
> Technically, the issue was covered for few months by another issue and
> got re-exposed by 926e680aadde ("EFI: suppress bogus loader warning").
> While it fixed one issue, it also made it possible to put sentinel NULL
> into argv[0] again. But the original EFI code had this issue too, so
> IMO the Fixes tag is correct.
>
> While there is convention to put file name as the first option, I don't
> see anything in the UEFI spec requiring it. So, Xen should not crash
> when it's missing.
Yet if the equivalent of argv[0] is missing from the command line, how
do we even know whether the first token on the command line is an
option (or the -- separator)?
Jan
next prev parent reply other threads:[~2025-08-29 7:17 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-08 13:56 [PATCH v2] xen/efi: Fix crash with initial empty EFI options Frediano Ziglio
2025-07-08 15:41 ` Jan Beulich
2025-07-08 17:22 ` Marek Marczykowski-Górecki
2025-07-09 5:30 ` Jan Beulich
2025-07-09 9:07 ` Frediano Ziglio
2025-07-09 9:11 ` Jan Beulich
2025-07-28 10:39 ` Frediano Ziglio
2025-08-15 10:32 ` Frediano Ziglio
2025-08-29 4:06 ` Marek Marczykowski-Górecki
2025-08-29 7:17 ` Jan Beulich [this message]
2025-08-29 9:15 ` Marek Marczykowski-Górecki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e8e3b2c7-39ac-4c3a-a7ff-e2b579f44d2c@suse.com \
--to=jbeulich@suse.com \
--cc=dpsmith@apertussolutions.com \
--cc=frediano.ziglio@cloud.com \
--cc=marmarek@invisiblethingslab.com \
--cc=xen-devel@lists.xenproject.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.