On Thu, Aug 21, 2025 at 7:58 AM Chia-Yu Chang (Nokia) <[email protected]> wrote: > > > -----Original Message----- > > From: Eric Dumazet <[email protected]> > > Sent: Thursday, August 21, 2025 2:30 PM > > To: Chia-Yu Chang (Nokia) <[email protected]> > > Cc: [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected]; [email protected]; Koen De Schepper (Nokia) > > <[email protected]>; [email protected]; > > [email protected]; [email protected]; > > [email protected]; [email protected]; [email protected]; > > [email protected] > > Subject: Re: [PATCH v15 net-next 10/14] tcp: accecn: AccECN option > > > > > > CAUTION: This is an external email. Please be very careful when clicking > > links or opening attachments. See the URL nok.it/ext for additional > > information. > > > > > > > > On Fri, Aug 15, 2025 at 1:40 AM <[email protected]> wrote: > [...] > > > /* Used for make_synack to form the ACE flags */ diff --git > > > a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h index > > > bdac8c42fa82..53e0e85b52be 100644 > > > --- a/include/uapi/linux/tcp.h > > > +++ b/include/uapi/linux/tcp.h > > > @@ -316,6 +316,13 @@ struct tcp_info { > > > * in milliseconds, including any > > > * unfinished recovery. > > > */ > > > + __u32 tcpi_received_ce; /* # of CE marks received */ > > > + __u32 tcpi_delivered_e1_bytes; /* Accurate ECN byte counters */ > > > + __u32 tcpi_delivered_e0_bytes; > > > + __u32 tcpi_delivered_ce_bytes; > > > + __u32 tcpi_received_e1_bytes; > > > + __u32 tcpi_received_e0_bytes; > > > + __u32 tcpi_received_ce_bytes; > > > }; > > > > > > > We do not add more fields to tcp_info, unless added fields are a multiple > > of 64 bits. > > > > Otherwise a hole is added and can not be recovered. > > Hi Eric, > > Thanks for the feedback. > > Then, would it make sense to add __u32 reserved; here or this is not an > option? >
I would prefer we take the opportunity to export a 32bit field right there, instead of a hole. A reserved field makes it difficult for ss commands to know if a new kernel is using it for a different purpose. Neal, any idea of what would be useful ? I was thinking lately of sk_err_soft, but I am not yet convinced.
