Send a new series and I replace the one in next net On Mon, Feb 2, 2026, 20:01 Dimon <[email protected]> wrote:
> Hi Thomas, > > Thank you for the feedback about splitting the Coverity fixes patch by > issue type. > You're absolutely right - it addresses three distinct problems and should > be separate commits. > > Now the patch is already queued to next-net, and since then five > additional nbl driver fixes have been merged: > net/nbl: improve mailbox exception handling > net/nbl: fix mbuf double-free in queue cleanup > net/nbl: fix mbuf headroom usage in packet Tx > net/nbl: fix memzone leak on queue release > net/nbl: fix hardware stats interrupt nesting issue > > What should I do? Should I ask Stephen to help revert these 6 commits and > then resubmit them? > > > > ------------------------------------------------------------------ > > 发件人：Thomas Monjalon <[email protected]> > 发送时间：2026年2月2日(周一) 04:29 > 收件人：Dimon<[email protected]> > 抄 送：dev<[email protected]>; stable<[email protected]>; Kyo Liu< > [email protected]>; Leon<[email protected]>; Sam< > [email protected]>; Stephen Hemminger<[email protected]> > 主 题：Re: [PATCH v3 1/1] net/nbl: fix issues reported by Coverity > > 21/01/2026 05:36, Stephen Hemminger: > > On Wed, 14 Jan 2026 19:27:44 -0800 > > Dimon Zhao <[email protected]> wrote: > > > > > Address multiple issues reported by Coverity static analysis: > > > > > > Check return values of ioctl() calls for NBL_DEV_USER_GET_BAR_SIZE > > > and NBL_DEV_USER_CLEAR_EVENTFD. The original code tested the return > > > value without capturing it from the ioctl call. > > > > > > Move debug log statement after NULL pointer validation in > > > nbl_dev_configure() to prevent dereference of adapter before the > > > check. > > > > > > Cast u16 operands to u64 before multiplication when calculating > > > buffer sizes and offsets to prevent potential integer overflow. > > > The num_txq_entries, num_rxq_entries, txq_buf_size, and rxq_buf_size > > > fields are all u16 types, and their product could overflow before > > > assignment to size_t or u64 destination variables. > > > > > > Coverity issue: 490942 > > > Coverity issue: 490943 > > > Coverity issue: 490946 > > > Coverity issue: 490947 > > > Coverity issue: 490949 > > > Coverity issue: 490950 > > > Coverity issue: 490951 > > > Coverity issue: 490952 > > > Coverity issue: 490953 > > > Coverity issue: 490954 > > > Coverity issue: 490955 > > > Coverity issue: 490957 > > > Coverity issue: 490958 > > > Coverity issue: 490959 > > > Fixes: a1c5ffa13b2c ("net/nbl: add channel layer") > > > Fixes: dc955cd24c8f ("net/nbl: add coexistence mode") > > > Fixes: 93b38df5a2ec ("net/nbl: support basic configuration") > > > Cc: [email protected] > > > > > > Signed-off-by: Dimon Zhao <[email protected]> > > > > Queued to next-net > > After another look, I think it deserves to be split per type of issue. > Maybe 3 commits? > > >
