Debian 13 kernel 6.12.111 and Broadcom NICs: Proxmox already did this one

Call: Hold on Debian 13 hosts with Broadcom NetXtreme-E (bnxt_en) NICs. This is reported, not confirmed: there are two merged Debian bugs, #1149564 and #1149790, and one of the reporters runs a fleet of identical servers. Stay on 6.12.107+deb13, which the reporter confirms is stable.

Symptoms: AMD-Vi IO_PAGE_FAULT from bnxt_en, “Abandoning msg … firmware status: 0x2000001”, and hwrm errors. One port dies, and later the other. It happens idle or under load. Firmware updates didn’t help.

Why it looks familiar: Proxmox kernel 7.0.14-20/-21 hit the same bnxt_en DMA padding regression in September. Proxmox fixed it in -22 and -23 with the upstream patch “bnxt_en: fix DMA mapping length for padded small packets.” If you run PVE, move to 7.0.14-23 or later and drop your kernel pins. If you run plain Debian, wait for a kernel that carries that patch.

Sources: debian-kernel list, Debian #1149564, Proxmox pve-devel.

Sources

• https://www.mail-archive.com/debian-kernel@lists.debian.org/msg150887.html
• https://bugs.debian.org/1149564
• https://lore.proxmox.com/pve-devel/1791358672.2psqg8mwv6.astroid@yuna.none/
• https://forum.proxmox.com/threads/proxmox-ve-9-2-21-bcm57412-bnxt_en-network-failure-after-upgrade-to-kernel-7-0-14-20-pve.186749/page-4
• https://forum.proxmox.com/threads/regression-in-7-0-14-20-pve-realtek-rtl8168h-r8169-fails-with-pcie-aer-dpc-error-%E2%80%93-7-0-14-19-pve-works.186785/