[LE-3812] RDMA/mana_ib: use the correct page table index based on hardware page size

[shreeya-patel98]

Commit message

RDMA/mana_ib: use the correct page table index based on hardware page size
jira LE-3812
commit-author Long Li <[email protected]>
commit 9e517a8
upstream-diff There were build failures and conflicts due to the
following missing commits :-
e849aad ("RDMA/mana_ib: Fix bug in creation of dma regions")
071bade ("net: mana: Add support for page sizes other than 4KB on ARM64")
5861cd1 ("net: mana: Enable MANA driver on ARM64 with 4K page size")

MANA hardware uses 4k page size. When calculating the page table index,
it should use the hardware page size, not the system page size.

	Cc: [email protected]
Fixes: 0266a17 ("RDMA/mana_ib: Add a driver for Microsoft Azure Network Adapter")
	Signed-off-by: Long Li <[email protected]>
Link: https://patch.msgid.link/[email protected]
	Signed-off-by: Leon Romanovsky <[email protected]>
(cherry picked from commit 9e517a8)
	Signed-off-by: Shreeya Patel <[email protected]>

----------------------------------------------------------------------------------------------------------
RDMA/mana_ib: use the correct page size for mapping user-mode doorbell page
jira LE-3812
commit-author Long Li <[email protected]>
commit 4a3b99b

When mapping doorbell page from user-mode, the driver should use the system
page size as this memory is allocated via mmap() from user-mode.

	Cc: [email protected]
Fixes: 0266a17 ("RDMA/mana_ib: Add a driver for Microsoft Azure Network Adapter")
	Signed-off-by: Long Li <[email protected]>
Link: https://patch.msgid.link/[email protected]
	Signed-off-by: Leon Romanovsky <[email protected]>
(cherry picked from commit 4a3b99b)
	Signed-off-by: Shreeya Patel <[email protected]>
	

Kernel build logs

/mnt/scratch/workspace/sig-cloud-8/kernel-src-tree
Skipping make mrproper
[TIMER]{MRPROPER}: 0s
x86_64 architecture detected, copying config
'configs/kernel-x86_64.config' -> '.config'
Setting Local Version for build
CONFIG_LOCALVERSION="-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f"
Making olddefconfig
scripts/kconfig/conf  --olddefconfig Kconfig
#
# configuration written to .config
#
Starting Build
scripts/kconfig/conf  --syncconfig Kconfig
  UPD     include/config/kernel.release
  DESCEND objtool
  DESCEND bpf/resolve_btfids
  UPD     include/generated/utsrelease.h
  CALL    scripts/checksyscalls.sh
  CHK     include/generated/compile.h
  CC      init/version.o
  AR      init/built-in.a
  CC      kernel/sys.o
  CC      crypto/fips.o
  CC      drivers/base/firmware_loader/main.o
  AR      crypto/built-in.a
  AR      drivers/base/firmware_loader/built-in.a
  AR      drivers/base/built-in.a
  CC      net/ethtool/ioctl.o
  CC      kernel/trace/trace.o
  
  <--snip-->
  
  INSTALL sound/usb/usx2y/snd-usb-usx2y.ko
  INSTALL sound/virtio/virtio_snd.ko
  INSTALL sound/x86/snd-hdmi-lpe-audio.ko
  INSTALL sound/xen/snd_xen_front.ko
  INSTALL virt/lib/irqbypass.ko
  DEPMOD  4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+
[TIMER]{MODULES}: 8s
Making Install
sh ./arch/x86/boot/install.sh 4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+ arch/x86/boot/bzImage \
	System.map "/boot"
[TIMER]{INSTALL}: 17s
Checking kABI
kABI check passed
Setting Default Kernel to /boot/vmlinuz-4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+ and Index to 1
The default is /boot/loader/entries/753870ec7b134d7582d817dd9bf35992-4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+.conf with index 1 and kernel /boot/vmlinuz-4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+
The default is /boot/loader/entries/753870ec7b134d7582d817dd9bf35992-4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+.conf with index 1 and kernel /boot/vmlinuz-4.18.0-shreeya_sig-cloud-8_4.18.0-553.62.1.el8_10-fb6d8f948c9f+
Generating grub configuration file ...
done
Hopefully Grub2.0 took everything ... rebooting after time metrices
[TIMER]{MRPROPER}: 0s
[TIMER]{BUILD}: 269s
[TIMER]{MODULES}: 8s
[TIMER]{INSTALL}: 17s
[TIMER]{TOTAL} 297s
Rebooting in 10 seconds

kernel-build.log

Kselftests

shreeya@spatel-dev-bom ~/c/w/sig-cloud-8> grep '^ok ' kselftest-before.log | wc -l && grep '^ok ' kselftest-after.log | wc -l
217
217
shreeya@spatel-dev-bom ~/c/w/sig-cloud-8> grep '^not ok ' kselftest-before.log | wc -l && grep '^not ok ' kselftest-after.log | wc -l
57
57

kselftest-after.log
kselftest-before.log

Kernel Commit Check

shreeya@spatel-dev-bom ~/c/w/s/kernel-src-tree-tools (mainline) [1]> python3 check_kernel_commits.py --repo ../kernel-src-tree --pr_branch {shreeya}_sig-cloud-8/4.18.0-553.62.1.el8_10 --base_branch origin/sig-cloud-8/4.18.0-553.62.1.el8_10 --markdown
> ✅ **All referenced commits exist upstream and have no Fixes: tags.**

[Copilot]

Pull Request Overview

This PR addresses page size handling in the MANA RDMA driver to correctly use hardware page size (4KB) instead of system page size for hardware interactions, while maintaining system page size for user-mode mappings.

• Introduces MANA-specific page size constants and macros for 4KB hardware pages
• Updates all DMA region and queue size calculations to use hardware page size
• Fixes user-mode doorbell page mapping to use system page size
• Enables MANA driver support on ARM64 with non-big-endian configurations

Reviewed Changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
include/net/mana/gdma.h	Adds MANA page size constants and macros for 4KB hardware pages
include/net/mana/mana.h	Updates EQ_SIZE to use MANA_PAGE_SIZE instead of PAGE_SIZE
drivers/net/ethernet/microsoft/mana/*.c	Replaces PAGE_* macros with MANA_PAGE_* equivalents for hardware operations
drivers/net/ethernet/microsoft/Kconfig	Enables MANA driver on ARM64 (non-big-endian) platforms
drivers/infiniband/hw/mana/main.c	Fixes page table index calculation and user-mode doorbell mapping

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[PlaidCat]

in the final commit we don't need to know that there was context outside of this commit.

upstream-diff There were build failures and conflicts due to the
following missing commits :-
https://github.com/ctrliq/kernel-src-tree/commit/e849aadfd5936e4e717107cccc811e34280c0bd8 ("RDMA/mana_ib: Fix bug in creation of dma regions")
https://github.com/ctrliq/kernel-src-tree/commit/071badedd6b952ae335c9a5a435e031dcc42793d ("net: mana: Add support for page sizes other than 4KB on ARM64")
https://github.com/ctrliq/kernel-src-tree/commit/5861cd1208d91f2b9adcb99ef2ffd0c209972496 ("net: mana: Enable MANA driver on ARM64 with 4K page size")

This is just for why is the upstream differnt from this commit. The context, correctly called out is for the PR.

[PlaidCat]

[thefossguy-ciq]

🚤

[bmastbergen]

🥌