| Message ID | 20240416122950.39046-1-hengqi@linux.alibaba.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | ethtool: provide the dim profile fine-tuning channel | expand |
On Tue, 16 Apr 2024 20:29:46 +0800 Heng Qi wrote:
> Please review, thank you very much!
Please stop posting code which does not compile.
The mailing list is for reviewing code, not build testing it.
在 2024/4/17 上午8:38, Jakub Kicinski 写道: > On Tue, 16 Apr 2024 20:29:46 +0800 Heng Qi wrote: >> Please review, thank you very much! > Please stop posting code which does not compile. > The mailing list is for reviewing code, not build testing it. Sorry, three versions v7, RESEND v7, v8 in two days, this may have caused a mess. RESEND v7 has addressed possible bot warnings, and v8 is a fix for Simon's comment. Have you encountered compilation problems in v8? Anyway, I'll bide my time on v8 to avoid another mess. ^^ Thanks.
On Wed, 17 Apr 2024 10:22:52 +0800 Heng Qi wrote:
> Have you encountered compilation problems in v8?
Yes, please try building allmodconfig:
make allmodconfig
make ..
there's many drivers using this API, you gotta build the full kernel..
Hello Jakub, On Wed, Apr 17, 2024 at 10:30 AM Jakub Kicinski <kuba@kernel.org> wrote: > > On Wed, 17 Apr 2024 10:22:52 +0800 Heng Qi wrote: > > Have you encountered compilation problems in v8? > > Yes, please try building allmodconfig: > > make allmodconfig > make .. > > there's many drivers using this API, you gotta build the full kernel.. > About compiling the kernel, I would like to ask one question: what parameters of 'make' do you recommend just like the netdev/build_32bit [1] does? If I run the normal 'make' command without parameters, there is no warning even if the file is not compiled. If running with '-Oline -j 16 W=1 C=1', it will print lots of warnings which makes it very hard to see the useful information related to the commits I just wrote. I want to build and then print all the warnings which are only related to my patches locally, but I cannot find a good way :( [1]: https://netdev.bots.linux.dev/static/nipa/845020/13631720/build_32bit/stderr Thanks, Jason
在 2024/4/17 上午10:53, Jason Xing 写道: > Hello Jakub, > > On Wed, Apr 17, 2024 at 10:30 AM Jakub Kicinski <kuba@kernel.org> wrote: >> On Wed, 17 Apr 2024 10:22:52 +0800 Heng Qi wrote: >>> Have you encountered compilation problems in v8? >> Yes, please try building allmodconfig: >> >> make allmodconfig >> make .. >> >> there's many drivers using this API, you gotta build the full kernel.. >> > About compiling the kernel, I would like to ask one question: what > parameters of 'make' do you recommend just like the netdev/build_32bit > [1] does? Hi Jason, I founded and checked the use of nipa[1] made from Kuba today. I used run.sh in the ./nipa/docker directory locally to run full compilation. If there are additional errors or warnings after applying our own patches than before, we will be given information in the results directory. [1] https://github.com/linux-netdev/nipa/tree/main Thanks. > > If I run the normal 'make' command without parameters, there is no > warning even if the file is not compiled. If running with '-Oline -j > 16 W=1 C=1', it will print lots of warnings which makes it very hard > to see the useful information related to the commits I just wrote. > > I want to build and then print all the warnings which are only related > to my patches locally, but I cannot find a good way :( > > [1]: https://netdev.bots.linux.dev/static/nipa/845020/13631720/build_32bit/stderr > > Thanks, > Jason
On Wed, Apr 17, 2024 at 2:35 PM Heng Qi <hengqi@linux.alibaba.com> wrote: > > > > 在 2024/4/17 上午10:53, Jason Xing 写道: > > Hello Jakub, > > > > On Wed, Apr 17, 2024 at 10:30 AM Jakub Kicinski <kuba@kernel.org> wrote: > >> On Wed, 17 Apr 2024 10:22:52 +0800 Heng Qi wrote: > >>> Have you encountered compilation problems in v8? > >> Yes, please try building allmodconfig: > >> > >> make allmodconfig > >> make .. > >> > >> there's many drivers using this API, you gotta build the full kernel.. > >> > > About compiling the kernel, I would like to ask one question: what > > parameters of 'make' do you recommend just like the netdev/build_32bit > > [1] does? > > Hi Jason, > > I founded and checked the use of nipa[1] made from Kuba today. I used > run.sh in the ./nipa/docker > directory locally to run full compilation. > If there are additional errors or warnings after applying our own > patches than before, we will be given > information in the results directory. > > [1] https://github.com/linux-netdev/nipa/tree/main Great! Thanks for the information:) I'll try it locally. Thanks, Jason > > Thanks. > > > > > If I run the normal 'make' command without parameters, there is no > > warning even if the file is not compiled. If running with '-Oline -j > > 16 W=1 C=1', it will print lots of warnings which makes it very hard > > to see the useful information related to the commits I just wrote. > > > > I want to build and then print all the warnings which are only related > > to my patches locally, but I cannot find a good way :( > > > > [1]: https://netdev.bots.linux.dev/static/nipa/845020/13631720/build_32bit/stderr > > > > Thanks, > > Jason >
在 2024/4/17 下午2:52, Jason Xing 写道: > On Wed, Apr 17, 2024 at 2:35 PM Heng Qi <hengqi@linux.alibaba.com> wrote: >> >> >> 在 2024/4/17 上午10:53, Jason Xing 写道: >>> Hello Jakub, >>> >>> On Wed, Apr 17, 2024 at 10:30 AM Jakub Kicinski <kuba@kernel.org> wrote: >>>> On Wed, 17 Apr 2024 10:22:52 +0800 Heng Qi wrote: >>>>> Have you encountered compilation problems in v8? >>>> Yes, please try building allmodconfig: >>>> >>>> make allmodconfig >>>> make .. >>>> >>>> there's many drivers using this API, you gotta build the full kernel.. >>>> >>> About compiling the kernel, I would like to ask one question: what >>> parameters of 'make' do you recommend just like the netdev/build_32bit >>> [1] does? >> Hi Jason, >> >> I founded and checked the use of nipa[1] made from Kuba today. I used >> run.sh in the ./nipa/docker >> directory locally to run full compilation. >> If there are additional errors or warnings after applying our own >> patches than before, we will be given >> information in the results directory. >> >> [1] https://github.com/linux-netdev/nipa/tree/main > Great! Thanks for the information:) I'll try it locally. To update more information, I checked the source code of nipa. All its test items are located at: 1. nipa/tests/patch, such as build_32bit, build_allmodconfig_war 2. nipa/tests/series, such as fixes_present We can edit the "include" and "exclude" items of [tests] section in the nipa/docker/nipa.config file to specify the specific items we want to test. Regards, Heng > > Thanks, > Jason > >> Thanks. >> >>> If I run the normal 'make' command without parameters, there is no >>> warning even if the file is not compiled. If running with '-Oline -j >>> 16 W=1 C=1', it will print lots of warnings which makes it very hard >>> to see the useful information related to the commits I just wrote. >>> >>> I want to build and then print all the warnings which are only related >>> to my patches locally, but I cannot find a good way :( >>> >>> [1]: https://netdev.bots.linux.dev/static/nipa/845020/13631720/build_32bit/stderr >>> >>> Thanks, >>> Jason
The NetDIM library provides excellent acceleration for many modern network cards. However, the default profiles of DIM limits its maximum capabilities for different NICs, so providing a way which the NIC can be custom configured is necessary. Currently, interaction with the driver is still based on the commonly used "ethtool -C". Since the profile now exists in netdevice, adding a function similar to net_dim_get_rx_moderation_dev() with netdevice as argument is nice, but this would be better along with cleaning up the rest of the drivers, which we can get to very soon after this set. Please review, thank you very much! Changelog ===== v7->v8: - Use kmemdup() instead of kzalloc()/memcpy() in dev_dim_profile_init(). v6->v7: - A new wrapper struct pointer is used in struct net_device. - Add IS_ENABLED(CONFIG_DIMLIB) to avoid compiler warnings. - Profile fields changed from u16 to u32. v5->v6: - Place the profile in netdevice to bypass the driver. The interaction code of ethtool <-> kernel has not changed at all, only the interaction part of kernel <-> driver has changed. v4->v5: - Update some snippets from Kuba, Thanks. v3->v4: - Some tiny updates and patch 1 only add a new comment. v2->v3: - Break up the attributes to avoid the use of raw c structs. - Use per-device profile instead of global profile in the driver. v1->v2: - Use ethtool tool instead of net-sysfs Heng Qi (4): linux/dim: move useful macros to .h file ethtool: provide customized dim profile management virtio-net: refactor dim initialization/destruction virtio-net: support dim profile fine-tuning Documentation/netlink/specs/ethtool.yaml | 33 +++ Documentation/networking/ethtool-netlink.rst | 8 + drivers/net/virtio_net.c | 46 +++-- include/linux/dim.h | 13 ++ include/linux/ethtool.h | 11 +- include/linux/netdevice.h | 24 +++ include/uapi/linux/ethtool_netlink.h | 24 +++ lib/dim/net_dim.c | 10 +- net/core/dev.c | 79 ++++++++ net/ethtool/coalesce.c | 201 ++++++++++++++++++- 10 files changed, 426 insertions(+), 23 deletions(-)