Linus Torvalds Rejects MMC Changes For Linux 7.0 Cycle: "Complete Garbage"
([Linux Storage] 28 Minutes Ago
MMC Subsystem)
- Reference: 0001612909
- News link: https://www.phoronix.com/news/Linux-7.0-No-MMC-Changes
- Source link:
The Linux MultiMediaCard "MMC" subsystem was set to see some new hardware support, optimized support for secure erase/trim on some eMMCs, and a variety of other improvements. But all of the MMC changes are rejected and will be for the duration of the Linux 7.0 cycle due to an apparent lack of testing and vetting via linux-next that led Linus Torvalds to calling it "complete garbage" and "untested crap".
There were some interesting MMC changes queued up ahead of Linux 7.0 like NXP IW61x device IDs for WiFi chips over SDIO, support for manufacturing dates past 2025, optimizing secure erase/TRIM for some Kingstom eMMCs, DW_MMC code clean-ups, Mediatek MT8189 support in mtk-sd, and various SHDCI driver updates.
The [1]pull request lists the MMC changes that were submitted for Linux 7.0.
But Linus Torvalds found the pull request to not build properly and has lacked sufficient vetting via the likes of linux-next, he commented:
"No.
Those changes are complete garbage and don't even compile. It has apparently never been in linux-next or been build-tested in any way.
When CONFIG_MULTIPLEXER=m, we build that core.o file
obj-$(CONFIG_MULTIPLEXER) += mux-core.o
but in include/linux/mux/consumer.h you have
#ifdef CONFIG_MULTIPLEXER
which won't be true (because what will be defined is CONFIG_MULTIPLEXER_MODULE), so then you get a long stream of things like
drivers/mux/core.c:312:14: error: redefinition of ‘mux_control_states’
because the mux/consumer.h header will have defined the dummy wrapper function.
In other words, that commit ad314348ceb4 ("mux: Add helper functions for getting optional and selected mux-state") is pure unadulterated untested garbage.
I do not want to see a "fixed" pull request from you. This was entirely unacceptable, and I will not be pulling anything more from you this merge window.
Stop sending me untested crap that hasn't been in linux-next and doesn't even pass the most cursory smell test.
You can try again for 7.1, but only if it has been actually in linux-next and properly tested."
So those MMC changes will need to wait now until the Linux 7.1 merge window begins in mid-April following the Linux 7.0 stable debut.
[1] https://lore.kernel.org/lkml/20260209133441.556464-1-ulf.hansson@linaro.org/
There were some interesting MMC changes queued up ahead of Linux 7.0 like NXP IW61x device IDs for WiFi chips over SDIO, support for manufacturing dates past 2025, optimizing secure erase/TRIM for some Kingstom eMMCs, DW_MMC code clean-ups, Mediatek MT8189 support in mtk-sd, and various SHDCI driver updates.
The [1]pull request lists the MMC changes that were submitted for Linux 7.0.
But Linus Torvalds found the pull request to not build properly and has lacked sufficient vetting via the likes of linux-next, he commented:
"No.
Those changes are complete garbage and don't even compile. It has apparently never been in linux-next or been build-tested in any way.
When CONFIG_MULTIPLEXER=m, we build that core.o file
obj-$(CONFIG_MULTIPLEXER) += mux-core.o
but in include/linux/mux/consumer.h you have
#ifdef CONFIG_MULTIPLEXER
which won't be true (because what will be defined is CONFIG_MULTIPLEXER_MODULE), so then you get a long stream of things like
drivers/mux/core.c:312:14: error: redefinition of ‘mux_control_states’
because the mux/consumer.h header will have defined the dummy wrapper function.
In other words, that commit ad314348ceb4 ("mux: Add helper functions for getting optional and selected mux-state") is pure unadulterated untested garbage.
I do not want to see a "fixed" pull request from you. This was entirely unacceptable, and I will not be pulling anything more from you this merge window.
Stop sending me untested crap that hasn't been in linux-next and doesn't even pass the most cursory smell test.
You can try again for 7.1, but only if it has been actually in linux-next and properly tested."
So those MMC changes will need to wait now until the Linux 7.1 merge window begins in mid-April following the Linux 7.0 stable debut.
[1] https://lore.kernel.org/lkml/20260209133441.556464-1-ulf.hansson@linaro.org/