Re: [PATCH v26 00/28] riscv control-flow integrity for usermode

[Deepak Gupta <debug@rivosinc.com>]

On Wed, Jan 21, 2026 at 09:32:20AM +0800, Zong Li wrote:
>On Thu, Jan 8, 2026 at 11:10 PM Deepak Gupta <debug@rivosinc.com> wrote:
>>
>> Hi Paul,
>>
>> I have a bugfix for a bug reported by Jesse Huang (thanks Jesse) in riscv
>> implementation of `map_shadow_stack`.
>>
>> Should I send a new series or only the bugfix-patch for implementation
>> of `map_shadow_stack`
>>
>
>Hi Deepak,
>Not sure if I missed the bugfix patch, I couldn't find it on the
>mailing list. Could I know have you submitted it? If so, could you
>please point me where the patch is?
>Thanks

No you didn't miss anything. I had been busy with job change and even though
change was small. I wanted to do some basic testing before sending fix. I just
sent the fix on this v26 series, patch #10/28.

Sorry about the delay here.

>
>> Let me know. Thanks.
>>
>> -Deepak
>>
>> -Deepak
>>
>>
>>
>> On Fri, Dec 12, 2025 at 10:33 AM Deepak Gupta <debug@rivosinc.com> wrote:
>> >
>> > On Fri, Dec 12, 2025 at 01:30:29AM -0700, Paul Walmsley wrote:
>> > >On Thu, 11 Dec 2025, Deepak Gupta via B4 Relay wrote:
>> > >
>> > >> v26: CONFIG_RISCV_USER_CFI depends on CONFIG_MMU (dependency of shadow stack
>> > >> on MMU). Used b4 to pick tags, apparantly it messed up some tag picks. Fixing it
>> > >
>> > >Deepak: I'm now (at least) the third person to tell you to stop resending
>> > >this entire series over and over again.
>> >
>> > To be very honest I also feel very bad doing and DOSing the lists. Sorry to you
>> > and everyone else.
>> >
>> > But I have been sitting on this patch series for last 3-4 merge windows with
>> > patches being exactly same/similar. So I have been a little more than desperate
>> > to get it in.
>> >
>> > I really haven't had any meaningful feedback on patch series except stalling
>> > just before each merge window for reasons which really shouldn't stall its
>> > merge. Sure that's the nature of open source development and it's maintainer's
>> > call at the end of the day. And I am new to this. I'll improve.
>> >
>> > >
>> > >First, a modified version of the CFI v23 series was ALREADY SITTING IN
>> > >LINUX-NEXT.  So there's no reason you should be resending the entire
>> > >series, UNLESS your intention for me is to drop the entire existing series
>> > >and wait for another merge window.
>> > >
>> > >Second: when someone asks you questions about an individual patch, and you
>> > >want to answer those questions, it's NOT GOOD for you to resend the entire
>> > >28 series as the response!  You are DDOSing a bunch of lists and E-mail
>> > >inboxes.  Just answer the question in a single E-mail.  If you want to
>> > >update a single patch, just send that one patch.
>> >
>> > Noted. I wasn't sure about it. I'll explicitly ask next time if you want me to
>> > send another one.
>> >
>> > >
>> > >If you don't start paying attention to these rules then people are going
>> > >to start ignoring you -- at best! -- and it's going to give the entire
>> > >community a bad reputation.
>> >
>> > Even before this, this patch series has been ignored largely. I don't know
>> > how to get attention. All I wanted was either feedback or get it in. And as I
>> > said I've been desparate to get it in. Also as I said, I'll improve.
>> >
>> > >
>> > >Please acknowledge that you understand this,
>> >
>> > ACKed.
>> >
>> > >
>> > >
>> > >- Paul