Skip to contentSkip to conversation

Driver core changes for 5.19-rc1

GIT PULL In review Greg KH 5 participants 10 messages Jun 3, 2022 – Jun 4, 2022 Driver core

lore
Share & tools
lore

Greg KH

author maintainer
#0 lore raw

The following changes since commit 672c0c5173427e6b3e2a9bbb7be51ceeec78093a:

Linux 5.18-rc5 (2022-05-01 13:57:58 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git tags/driver-core-5.19-rc1

for you to fetch changes up to b232b02bf3c205b13a26dcec08e53baddd8e59ed:

driver core: fix deadlock in __device_attach (2022-05-19 19:37:08 +0200)

----------------------------------------------------------------
Driver core changes for 5.19-rc1

Here is the set of driver core changes for 5.19-rc1.

Note, I'm not really happy with this pull request as-is, see below for details, but overall this is all good for everything but a small set of systems, which we have a fix for already.

Lots of tiny driver core changes and cleanups happened this cycle, but the two major things were:

- firmware_loader reorganization and additions including the ability to have XZ compressed firmware images and the ability for userspace to initiate the firmware load when it needs to, instead of being always initiated by the kernel. FPGA devices specifically want this ability to have their firmware changed over the lifetime of the system boot, and this allows them to work without having to come up with yet-another-custom-uapi interface for loading firmware for them.
- physical location support added to sysfs so that devices that know this information, can tell userspace where they are located in a common way. Some ACPI devices already support this today, and more bus types should support this in the future.

Smaller changes included:
- driver_override api cleanups and fixes
- error path cleanups and fixes
- get_abi script fixes
- deferred probe timeout changes.

It's that last change that I'm the most worried about. It has been reported to cause boot problems for a number of systems, and I have a tested patch series that resolves this issue. But I didn't get it merged into my tree before 5.18-final came out, so it has not gotten any linux-next testing.

I'll send the fixup patches (there are 2) as a follow-on series to this pull request if you want to take them directly, _OR_ I can just revert the probe timeout changes and they can wait for the next -rc1 merge cycle. Given that the fixes are tested, and pretty simple, I'm leaning toward that choice. Sorry this all came at the end of the merge window, I should have resolved this all 2 weeks ago, that's my fault as it was in the middle of some travel for me.

All have been tested in linux-next for weeks, with no reported issues other than the above-mentioned boot time outs.

----------------------------------------------------------------
Bagas Sanjaya (2):
Documentation: dd: Use ReST lists for return values of driver_deferred_probe_check_state()
firmware_loader: describe 'module' parameter of firmware_upload_register()

Christophe JAILLET (1):
drivers/base/memory: Fix an unlikely reference counting issue in __add_memory_block()

Dan Carpenter (2):
driver core: location: Check for allocations failure
test_firmware: fix end of loop test in upload_read_show()

Dietmar Eggemann (1):
topology: Remove unused cpu_cluster_mask()

Greg Kroah-Hartman (3):
Revert "firmware_loader: use kernel credentials when reading firmware" Merge 5.18-rc5 into driver-core-next
export: fix string handling of namespace in EXPORT_SYMBOL_NS

Hao Luo (1):
kernfs: Separate kernfs_pr_cont_buf and rename_lock.

Haowen Bai (1):
firmware: edd: Remove redundant condition

Imran Khan (2):
kernfs: Remove reference counting for kernfs_open_node.
kernfs: Rename kernfs_put_open_node to kernfs_unlink_open_file.

Krzysztof Kozlowski (14):
driver: platform: Add helper for safer setting of driver_override
amba: Use driver_set_override() instead of open-coding
fsl-mc: Use driver_set_override() instead of open-coding
hv: Use driver_set_override() instead of open-coding
PCI: Use driver_set_override() instead of open-coding s390/cio: Use driver_set_override() instead of open-coding
spi: Use helper for safer setting of driver_override
vdpa: Use helper for safer setting of driver_override
clk: imx: scu: Fix kfree() of static memory on setting driver_override
slimbus: qcom-ngd: Fix kfree() of static memory on setting driver_override
rpmsg: Constify local variable in field store macro
rpmsg: Fix kfree() of static memory on setting driver_override
rpmsg: Fix calling device_lock() on non-initialized device
rpmsg: use local 'dev' variable

Luis Chamberlain (1):
MAINTAINERS: add Russ Weight as a firmware loader maintainer

Lukasz Luba (1):
arch_topology: Trace the update thermal pressure

Mark-PK Tsai (1):
driver core: Prevent overriding async driver of a device before it probe

Miaohe Lin (1):
drivers/base/node.c: fix compaction sysfs file leak

Russ Weight (10):
firmware_loader: Clear data and size in fw_free_paged_buf
firmware_loader: Check fw_state_is_done in loading_store
firmware_loader: Split sysfs support from fallback
firmware_loader: Add firmware-upload support
firmware_loader: Add sysfs nodes to monitor fw_upload
test_firmware: Add test support for firmware upload
test_firmware: Error injection for firmware upload
selftests: firmware: Add firmware upload selftests
firmware_loader: Fix configs for sysfs split
firmware_loader: Move definitions from sysfs_upload.h to sysfs.h

Saravana Kannan (2):
driver core: Add "*" wildcard support to driver_async_probe cmdline param driver core: Extend deferred probe timeout on driver registration

Schspa Shi (1):
driver: base: fix UAF when driver_attach failed

SeongJae Park (1):
scripts/get_abi: Fix wrong script file name in the help message

Sergey Shtylyov (1):
platform: finally disallow IRQ0 in platform_get_irq() and its ilk

Takashi Iwai (5):
firmware: Add the support for ZSTD-compressed firmware files
selftests: firmware: Use smaller dictionary for XZ compression
selftests: firmware: Fix the request_firmware_into_buf() test for XZ format
selftests: firmware: Simplify test patterns
selftests: firmware: Add ZSTD compressed file tests

Thiébaud Weksteen (1):
firmware_loader: use kernel credentials when reading firmware

Won Chung (3):
driver core: Add sysfs support for physical location of a device driver core: location: Free struct acpi_pld_info *pld driver core: location: Add "back" as a possible output for panel

Zhang Wensheng (1):
driver core: fix deadlock in __device_attach

 Documentation/ABI/testing/sysfs-class-firmware     |  77 ++++
 .../ABI/testing/sysfs-devices-physical_location    |  42 ++
 Documentation/admin-guide/kernel-parameters.txt    |  11 +-
 Documentation/driver-api/firmware/fw_upload.rst    | 126 ++++++
 Documentation/driver-api/firmware/index.rst        |   1 +
 MAINTAINERS                                        |   1 +
 drivers/amba/bus.c                                 |  28 +-
 drivers/base/Makefile                              |   1 +
 drivers/base/arch_topology.c                       |   5 +
 drivers/base/base.h                                |   1 +
 drivers/base/bus.c                                 |   4 +-
 drivers/base/core.c                                |  15 +
 drivers/base/dd.c                                  |  44 ++-
 drivers/base/driver.c                              |  70 ++++
 drivers/base/firmware_loader/Kconfig               |  42 +-
 drivers/base/firmware_loader/Makefile              |   2 +
 drivers/base/firmware_loader/fallback.c            | 430 ---------------------
 drivers/base/firmware_loader/fallback.h            |  46 +--
 drivers/base/firmware_loader/firmware.h            |  16 +
 drivers/base/firmware_loader/main.c                |  94 ++++-
 drivers/base/firmware_loader/sysfs.c               | 422 ++++++++++++++++++++
 drivers/base/firmware_loader/sysfs.h               | 117 ++++++
 drivers/base/firmware_loader/sysfs_upload.c        | 397 +++++++++++++++++++
 drivers/base/firmware_loader/sysfs_upload.h        |  41 ++
 drivers/base/memory.c                              |   5 +-
 drivers/base/node.c                                |   1 +
 drivers/base/physical_location.c                   | 143 +++++++
 drivers/base/physical_location.h                   |  16 +
 drivers/base/platform.c                            |  34 +-
 drivers/bus/fsl-mc/fsl-mc-bus.c                    |  25 +-
 drivers/clk/imx/clk-scu.c                          |   7 +-
 drivers/firmware/edd.c                             |   3 +-
 drivers/hv/vmbus_drv.c                             |  28 +-
 drivers/pci/pci-sysfs.c                            |  28 +-
 drivers/rpmsg/rpmsg_core.c                         |  42 +-
 drivers/rpmsg/rpmsg_internal.h                     |   5 +-
 drivers/rpmsg/rpmsg_ns.c                           |   4 +-
 drivers/s390/cio/cio.h                             |   6 +-
 drivers/s390/cio/css.c                             |  28 +-
 drivers/slimbus/qcom-ngd-ctrl.c                    |  13 +-
 drivers/spi/spi.c                                  |  26 +-
 drivers/vdpa/vdpa.c                                |  29 +-
 fs/kernfs/dir.c                                    |  31 +-
 fs/kernfs/file.c                                   |  47 ++-
 include/linux/amba/bus.h                           |   6 +-
 include/linux/device.h                             |  73 ++++
 include/linux/device/driver.h                      |   2 +
 include/linux/export.h                             |   7 +-
 include/linux/firmware.h                           |  82 ++++
 include/linux/fsl/mc.h                             |   6 +-
 include/linux/hyperv.h                             |   6 +-
 include/linux/pci.h                                |   6 +-
 include/linux/platform_device.h                    |   6 +-
 include/linux/rpmsg.h                              |  14 +-
 include/linux/spi/spi.h                            |   2 +
 include/linux/topology.h                           |   7 -
 include/linux/vdpa.h                               |   4 +-
 include/trace/events/thermal_pressure.h            |  29 ++
 lib/test_firmware.c                                | 381 ++++++++++++++++++
 scripts/get_abi.pl                                 |   4 +-
 tools/testing/selftests/firmware/Makefile          |   2 +-
 tools/testing/selftests/firmware/config            |   1 +
 tools/testing/selftests/firmware/fw_filesystem.sh  | 170 ++++----
 tools/testing/selftests/firmware/fw_lib.sh         |  19 +-
 tools/testing/selftests/firmware/fw_run_tests.sh   |   4 +
 tools/testing/selftests/firmware/fw_upload.sh      | 214 ++++++++++
 66 files changed, 2736 insertions(+), 863 deletions(-)
 create mode 100644 Documentation/ABI/testing/sysfs-class-firmware
 create mode 100644 Documentation/ABI/testing/sysfs-devices-physical_location
 create mode 100644 Documentation/driver-api/firmware/fw_upload.rst
 create mode 100644 drivers/base/firmware_loader/sysfs.c
 create mode 100644 drivers/base/firmware_loader/sysfs.h
 create mode 100644 drivers/base/firmware_loader/sysfs_upload.c
 create mode 100644 drivers/base/firmware_loader/sysfs_upload.h
 create mode 100644 drivers/base/physical_location.c
 create mode 100644 drivers/base/physical_location.h
 create mode 100644 include/trace/events/thermal_pressure.h
 create mode 100755 tools/testing/selftests/firmware/fw_upload.sh
author maintainer
#1 lore raw

PATCH 1/2 driver core: Fix wait_for_device_probe() & deferred_probe_timeout interaction

From: Saravana Kannan <[email protected]>

Mounting NFS rootfs was timing out when deferred_probe_timeout was non-zero [1]. This was because ip_auto_config() initcall times out waiting for the network interfaces to show up when deferred_probe_timeout was non-zero. While ip_auto_config() calls wait_for_device_probe() to make sure any currently running deferred probe work or asynchronous probe finishes, that wasn't sufficient to account for devices being deferred until deferred_probe_timeout.

Commit 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits until the deferred_probe_timeout fires") tried to fix that by making sure wait_for_device_probe() waits for deferred_probe_timeout to expire before returning.

However, if wait_for_device_probe() is called from the kernel_init() context:

- Before deferred_probe_initcall() [2], it causes the boot process to hang due to a deadlock.

- After deferred_probe_initcall() [3], it blocks kernel_init() from continuing till deferred_probe_timeout expires and beats the point of deferred_probe_timeout that's trying to wait for userspace to load modules.

Neither of this is good. So revert the changes to
wait_for_device_probe().

[1] - https://lore.kernel.org/lkml/TYAPR01MB45443DF63B9EF29054F7C41FD8C60@TYAPR01MB4544.jpnprd01.prod.outlook.com/
[2] - https://lore.kernel.org/lkml/[email protected]/
[3] - https://lore.kernel.org/lkml/[email protected]/

Fixes 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits until the deferred_probe_timeout fires")
Cc John Stultz <[email protected]>
Cc "David S. Miller" <[email protected]>
Cc Alexey Kuznetsov <[email protected]>
Cc Hideaki YOSHIFUJI <[email protected]>
Cc Jakub Kicinski <[email protected]>
Cc Rafael J. Wysocki <[email protected]>
Cc Rob Herring <[email protected]>
Cc Geert Uytterhoeven <[email protected]>
Cc Yoshihiro Shimoda <[email protected]>
Cc Robin Murphy <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Sudeep Holla <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Naresh Kamboju <[email protected]>
Cc Basil Eljuse <[email protected]>
Cc Ferry Toth <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Anders Roxell <[email protected]>
---
 drivers/base/dd.c | 5 -----
 1 file changed, 5 deletions(-)
drivers/base/dd.c +0 -5
index 2fc8507f59ee..91f63cd33b12 100644
@@ -263,7 +263,6 @@ int driver_deferred_probe_timeout;
 #endif
 
 EXPORT_SYMBOL_GPL(driver_deferred_probe_timeout);
-static DECLARE_WAIT_QUEUE_HEAD(probe_timeout_waitqueue);
 
 static int __init deferred_probe_timeout_setup(char *str)
 {
@@ -318,7 +317,6 @@ static void deferred_probe_timeout_work_func(struct work_struct *work)
 	list_for_each_entry(p, &deferred_probe_pending_list, deferred_probe)
 		dev_info(p->device, "deferred probe pending\n");
 	mutex_unlock(&deferred_probe_mutex);
-	wake_up_all(&probe_timeout_waitqueue);
 }
 static DECLARE_DELAYED_WORK(deferred_probe_timeout_work, deferred_probe_timeout_work_func);
 
@@ -736,9 +734,6 @@ int driver_probe_done(void)
  */
 void wait_for_device_probe(void)
 {
-	/* wait for probe timeout */
-	wait_event(probe_timeout_waitqueue, !driver_deferred_probe_timeout);
-
 	/* wait for the deferred probe workqueue to finish */
 	flush_work(&deferred_probe_work);
 
Signature · 1 line
2.36.1
author maintainer
#2 lore raw

PATCH 2/2 driver core: Set default deferred_probe_timeout back to 0.

From: Saravana Kannan <[email protected]>

Since we had to effectively reverted
commit 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits until the deferred_probe_timeout fires") in an earlier patch, a non-zero deferred_probe_timeout will break NFS rootfs mounting [1] again. So, set the default back to zero until we can fix that.

[1] - https://lore.kernel.org/lkml/TYAPR01MB45443DF63B9EF29054F7C41FD8C60@TYAPR01MB4544.jpnprd01.prod.outlook.com/

---
 drivers/base/dd.c | 5 -----
 1 file changed, 5 deletions(-)
drivers/base/dd.c +0 -5
index 91f63cd33b12..251b5ba1b84a 100644
@@ -256,12 +256,7 @@ static int deferred_devs_show(struct seq_file *s, void *data)
 }
 DEFINE_SHOW_ATTRIBUTE(deferred_devs);
 
-#ifdef CONFIG_MODULES
-int driver_deferred_probe_timeout = 10;
-#else
 int driver_deferred_probe_timeout;
-#endif
-
 EXPORT_SYMBOL_GPL(driver_deferred_probe_timeout);
 
 static int __init deferred_probe_timeout_setup(char *str)
Signature · 1 line
2.36.1
maintainer Reviewed-by on v1 1/2
#3 lore raw

Re: [PATCH 1/2] driver core: Fix wait_for_device_probe() & deferred_probe_timeout interaction

In reply to Greg Kroah-Hartman
Quoted 53 lines · Greg Kroah-Hartman

From: Saravana Kannan <[email protected]>

Mounting NFS rootfs was timing out when deferred_probe_timeout was non-zero [1]. This was because ip_auto_config() initcall times out waiting for the network interfaces to show up when deferred_probe_timeout was non-zero. While ip_auto_config() calls wait_for_device_probe() to make sure any currently running deferred probe work or asynchronous probe finishes, that wasn't sufficient to account for devices being deferred until deferred_probe_timeout.

Commit 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits until the deferred_probe_timeout fires") tried to fix that by making sure wait_for_device_probe() waits for deferred_probe_timeout to expire before returning.

However, if wait_for_device_probe() is called from the kernel_init() context:

- Before deferred_probe_initcall() [2], it causes the boot process to hang due to a deadlock.

- After deferred_probe_initcall() [3], it blocks kernel_init() from continuing till deferred_probe_timeout expires and beats the point of deferred_probe_timeout that's trying to wait for userspace to load modules.

Neither of this is good. So revert the changes to
wait_for_device_probe().

[1] - https://lore.kernel.org/lkml/TYAPR01MB45443DF63B9EF29054F7C41FD8C60@TYAPR01MB4544.jpnprd01.prod.outlook.com/
[2] - https://lore.kernel.org/lkml/[email protected]/
[3] - https://lore.kernel.org/lkml/[email protected]/

Fixes 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits until the deferred_probe_timeout fires")
Cc John Stultz <[email protected]>
Cc "David S. Miller" <[email protected]>
Cc Alexey Kuznetsov <[email protected]>
Cc Hideaki YOSHIFUJI <[email protected]>
Cc Jakub Kicinski <[email protected]>
Cc Rafael J. Wysocki <[email protected]>
Cc Rob Herring <[email protected]>
Cc Geert Uytterhoeven <[email protected]>
Cc Yoshihiro Shimoda <[email protected]>
Cc Robin Murphy <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Sudeep Holla <[email protected]>
Cc Andy Shevchenko <[email protected]>
Cc Naresh Kamboju <[email protected]>
Cc Basil Eljuse <[email protected]>
Cc Ferry Toth <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Anders Roxell <[email protected]>
Quoted 8 lines
Quoted 37 lines
---
 drivers/base/dd.c | 5 -----
 1 file changed, 5 deletions(-)
drivers/base/dd.c +0 -6
index 2fc8507f59ee..91f63cd33b12 100644
@@ -263,7 +263,6 @@ int driver_deferred_probe_timeout;
 #endif
 
 EXPORT_SYMBOL_GPL(driver_deferred_probe_timeout);
-static DECLARE_WAIT_QUEUE_HEAD(probe_timeout_waitqueue);
 
 static int __init deferred_probe_timeout_setup(char *str)
 {
@@ -318,7 +317,6 @@ static void deferred_probe_timeout_work_func(struct work_struct *work)
        list_for_each_entry(p, &deferred_probe_pending_list, deferred_probe)
                dev_info(p->device, "deferred probe pending\n");
        mutex_unlock(&deferred_probe_mutex);
-       wake_up_all(&probe_timeout_waitqueue);
 }
 static DECLARE_DELAYED_WORK(deferred_probe_timeout_work, deferred_probe_timeout_work_func);
 
@@ -736,9 +734,6 @@ int driver_probe_done(void)
  */
 void wait_for_device_probe(void)
 {
-       /* wait for probe timeout */
-       wait_event(probe_timeout_waitqueue, !driver_deferred_probe_timeout);
-
        /* wait for the deferred probe workqueue to finish */
        flush_work(&deferred_probe_work);
 
--

2.36.1

merge bot Merged
#4 lore raw
In reply to Greg KH
The pull request you sent on Fri, 3 Jun 2022 13:28:39 +0200:

The pull request you sent on Fri, 3 Jun 2022 13:28:39 +0200:

Quoted 1 line

git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git tags/driver-core-5.19-rc1

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/500a434fc593f1fdb274c0e6fe09a0b9c0711a4b

Thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
#5 lore raw
In reply to Greg KH
Quoted 4 lines · Greg KH

It's that last change that I'm the most worried about. It has been reported to cause boot problems for a number of systems, and I have a tested patch series that resolves this issue.

Ok, pulled and the two patches applied on top.

Linus

author maintainer
#6 lore raw
Quoted 7 lines · Linus Torvalds
Quoted 4 lines · Greg KH

It's that last change that I'm the most worried about. It has been reported to cause boot problems for a number of systems, and I have a tested patch series that resolves this issue.

Ok, pulled and the two patches applied on top.

Thanks!

#7 lore raw
In reply to Greg KH

Augh.

This was very badly done, and I'm not talking about the deferred probe timeout things that caused problems for people.

Quoted 5 lines · Greg KH

- firmware_loader reorganization and additions including the ability to have XZ compressed firmware images and the ability for userspace to initiate the firmware load when it needs to, instead of being always initiated by the kernel.

This is actively misleading.

We *always* supported XZ compressed firmware images, and it was enabled by CONFIG_FW_LOADER_COMPRESS,

What's new is the option to use ZSTD compression.

However, the Kconfig file addition for this was done as badly as the above explanation was, and the FW_LOADER_COMPRESS_XZ option was added with a help message and a default value that both are complete garbage.

So when you do "make oldconfig", you would be expected to say 'N', and in the process you lose the existing XZ compression.

Only when the resulting kernel doesn't boot, and you spent half an hour trying to bisect things, and you start looking closer, do you notice that "ooh, the config changed in bad ways".

Yeah, I'm a bit grumpy. This was *really* annoying.

The commit that does this breakage is literally called "firmware: Add the support for ZSTD-compressed firmware files", and only when looking closer do you notice that IT REMOVES SUPPORT FOR XZ COMPRESSION BY DEFAULT.

Because even when keeping the FW_LOADER_COMPRESS option enabled, the XZ compression is just gone, gone, gone, unless you realize that it was implicitly enabled before, and now needs that default disable of FW_LOADER_COMPRESS_XZ to be enabled.

I've said this before, and I'll say it here again (and I bet I'll have to say it in the future too): the kernel config is probably the most annoying part of building a kernel for anybody.

And it damn well does NOT HELP when people then actively break things, and ask actively bad and misleading questions. In this case, for example, it's not just that the XZ option is now misleading by default, it's also that the whole thing has been set up so that you can say "enable compressed images", but then HAVE NO ACTUAL COMPRESSION METHOD!

Grr. This was *REALLY* badly done.

Linus

#8 lore raw
Quoted 48 lines · Linus Torvalds

Augh.

This was very badly done, and I'm not talking about the deferred probe timeout things that caused problems for people.

Quoted 5 lines · Greg KH

- firmware_loader reorganization and additions including the ability to have XZ compressed firmware images and the ability for userspace to initiate the firmware load when it needs to, instead of being always initiated by the kernel.

This is actively misleading.

We *always* supported XZ compressed firmware images, and it was enabled by CONFIG_FW_LOADER_COMPRESS,

What's new is the option to use ZSTD compression.

However, the Kconfig file addition for this was done as badly as the above explanation was, and the FW_LOADER_COMPRESS_XZ option was added with a help message and a default value that both are complete garbage.

So when you do "make oldconfig", you would be expected to say 'N', and in the process you lose the existing XZ compression.

Only when the resulting kernel doesn't boot, and you spent half an hour trying to bisect things, and you start looking closer, do you notice that "ooh, the config changed in bad ways".

Yeah, I'm a bit grumpy. This was *really* annoying.

The commit that does this breakage is literally called "firmware: Add the support for ZSTD-compressed firmware files", and only when looking closer do you notice that IT REMOVES SUPPORT FOR XZ COMPRESSION BY DEFAULT.

Because even when keeping the FW_LOADER_COMPRESS option enabled, the XZ compression is just gone, gone, gone, unless you realize that it was implicitly enabled before, and now needs that default disable of FW_LOADER_COMPRESS_XZ to be enabled.

I've said this before, and I'll say it here again (and I bet I'll have to say it in the future too): the kernel config is probably the most annoying part of building a kernel for anybody.

Quoted 8 lines

And it damn well does NOT HELP when people then actively break things, and ask actively bad and misleading questions. In this case, for example, it's not just that the XZ option is now misleading by default, it's also that the whole thing has been set up so that you can say "enable compressed images", but then HAVE NO ACTUAL COMPRESSION METHOD!

Grr. This was *REALLY* badly done.

Mea culpa, I hesitated to add "default y", but it should have been added at least for the case with the config takeover case. Now I see that you've already added "default y" to CONFIG_FW_LOADER_COMPRESS_XZ. Thanks for taking care of this!

Takashi

author maintainer
#9 lore raw
Quoted 57 lines · Takashi Iwai
Quoted 55 lines · Linus Torvalds

Augh.

This was very badly done, and I'm not talking about the deferred probe timeout things that caused problems for people.

Quoted 5 lines · Greg KH

- firmware_loader reorganization and additions including the ability to have XZ compressed firmware images and the ability for userspace to initiate the firmware load when it needs to, instead of being always initiated by the kernel.

This is actively misleading.

We *always* supported XZ compressed firmware images, and it was enabled by CONFIG_FW_LOADER_COMPRESS,

What's new is the option to use ZSTD compression.

However, the Kconfig file addition for this was done as badly as the above explanation was, and the FW_LOADER_COMPRESS_XZ option was added with a help message and a default value that both are complete garbage.

So when you do "make oldconfig", you would be expected to say 'N', and in the process you lose the existing XZ compression.

Only when the resulting kernel doesn't boot, and you spent half an hour trying to bisect things, and you start looking closer, do you notice that "ooh, the config changed in bad ways".

Yeah, I'm a bit grumpy. This was *really* annoying.

The commit that does this breakage is literally called "firmware: Add the support for ZSTD-compressed firmware files", and only when looking closer do you notice that IT REMOVES SUPPORT FOR XZ COMPRESSION BY DEFAULT.

Because even when keeping the FW_LOADER_COMPRESS option enabled, the XZ compression is just gone, gone, gone, unless you realize that it was implicitly enabled before, and now needs that default disable of FW_LOADER_COMPRESS_XZ to be enabled.

I've said this before, and I'll say it here again (and I bet I'll have to say it in the future too): the kernel config is probably the most annoying part of building a kernel for anybody.

And it damn well does NOT HELP when people then actively break things, and ask actively bad and misleading questions. In this case, for example, it's not just that the XZ option is now misleading by default, it's also that the whole thing has been set up so that you can say "enable compressed images", but then HAVE NO ACTUAL COMPRESSION METHOD!

Quoted 6 lines
Quoted 1 line

Grr. This was *REALLY* badly done.

Mea culpa, I hesitated to add "default y", but it should have been added at least for the case with the config takeover case. Now I see that you've already added "default y" to CONFIG_FW_LOADER_COMPRESS_XZ. Thanks for taking care of this!

Sorry about that, I missed it too as I just selected that option to ensure it built and worked properly on my systems :(

greg k-h