Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

[Christoph Hellwig]

FYI, Linus' latest rant shows up in bugzilla, but never made it to me,
just as the other replies from Artem and Hans.  This just shows how
broken bugzilla is as a reporting tool.  Please be a little more calm,
I've always taken reported regressions series and as a first priority,
but it really does not help if information is hidden away.  Adding
Artem to the Cc list and drop bugzilla to make this work a bit better.

Artem, can you test the patch below?  This fixes the broken dma_mask
handling in platform_device_register_full that could override a
perfectly valid mask with 0.  If this doesn't work, can you throw
in a dump_stack() into the working kernel build to see where
platform_device_register_full and setup_pdev_dma_masks get called
for your system?

diff --git a/drivers/base/platform.c b/drivers/base/platform.c
index 7fa654f1288b..03035661eb6b 100644
--- a/drivers/base/platform.c
+++ b/drivers/base/platform.c
@@ -662,19 +662,6 @@ struct platform_device *platform_device_register_full(
 	pdev->dev.of_node_reused = pdevinfo->of_node_reused;
 
 	if (pdevinfo->dma_mask) {
-		/*
-		 * This memory isn't freed when the device is put,
-		 * I don't have a nice idea for that though.  Conceptually
-		 * dma_mask in struct device should not be a pointer.
-		 * See http://thread.gmane.org/gmane.linux.kernel.pci/9081
-		 */
-		pdev->dev.dma_mask =
-			kmalloc(sizeof(*pdev->dev.dma_mask), GFP_KERNEL);
-		if (!pdev->dev.dma_mask)
-			goto err;
-
-		kmemleak_ignore(pdev->dev.dma_mask);
-
 		*pdev->dev.dma_mask = pdevinfo->dma_mask;
 		pdev->dev.coherent_dma_mask = pdevinfo->dma_mask;
 	}

_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

[Linus Torvalds]

On Wed, Mar 11, 2020 at 9:24 AM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> So it will have a different commit ID, updated message, and be a mix
> of my patch and Christoph's.

I ended up pushing it out before starting on the pull requests, so
it's out there now.

Artem, it would be good to have confirmation that my (modified) tip of
tree now works for you. I don't actually doubt it does, but a final
confirmation would be appreciated.

            Linus
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

[Artem S. Tashkinov]

On 3/11/20 4:48 PM, Linus Torvalds wrote:
> On Wed, Mar 11, 2020 at 9:24 AM Linus Torvalds
> <torvalds@linux-foundation.org> wrote:
>>
>> So it will have a different commit ID, updated message, and be a mix
>> of my patch and Christoph's.
>
> I ended up pushing it out before starting on the pull requests, so
> it's out there now.
>
> Artem, it would be good to have confirmation that my (modified) tip of
> tree now works for you. I don't actually doubt it does, but a final
> confirmation would be appreciated.
>

Should I test kernel 5.6-rc5 with this patch applied or wait for you to
commit it?
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

[Artem S. Tashkinov]

On 3/11/20 4:21 PM, Linus Torvalds wrote:
> On Wed, Mar 11, 2020 at 9:02 AM Artem S. Tashkinov <aros@gmx.com> wrote:
>>
>> With this patch the system works (I haven't created an initrd, so it
>> doesn't completely boot and panics on not being able to mount root fs
>> but that's expected).
>
> Perfect.
>
> I ended up applying my earlier cleanup patch with just the added
> removal of the kfree(), which was the actual trigger of the bug.
>
> It's commit e423fb6929d4 ("driver code: clarify and fix platform
> device DMA mask allocation") in my tree. I've not pushed it out yet (I
> have a few pending pull requests), but it should be out shortly.


I've been able to compile and run
e3a36eb6dfaeea8175c05d5915dcf0b939be6dab successfully. I won't claim
this patch doesn't break something for other people :-)

Best regards,
Artem
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [Bug 206175] Fedora >= 5.4 kernels instantly freeze on boot without producing any display output

[Linus Torvalds]

On Wed, Mar 11, 2020 at 10:21 AM Artem S. Tashkinov <aros@gmx.com> wrote:
>
> I've been able to compile and run
> e3a36eb6dfaeea8175c05d5915dcf0b939be6dab successfully. I won't claim
> this patch doesn't break something for other people :-)

Thanks, that's all I was looking for.

If it breaks something for somebody else, I have a solid plan for that
too: I'll just sit in a corner and cry.

               Linus
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu