amdgpu with 8+ cards for GPU mining?

Submitted by Christian König on Feb. 19, 2018, 1:21 p.m.


Message ID
State New
Headers show
Series "amdgpu with 8+ cards for GPU mining?" ( rev: 1 ) in AMD X.Org drivers

Not browsing as part of any series.

Commit Message

Christian König Feb. 19, 2018, 1:21 p.m.
Hi Joseph,

as a band aid you can try the attached patch. It should at least fix the 
crash at hand and allow amdgpu to continue with the boot process.


Am 19.02.2018 um 14:13 schrieb Christian König:
> Hi Joseph,
> and here is the root cause of the problem:
>> 0b:00.0 VGA compatible controller: Advanced Micro Devices, Inc. 
>> [AMD/ATI] Ellesmere [Radeon RX 470/480/570/580] (rev ef) (prog-if 00 
>> [VGA controller])
>>     Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Device 0b31
>>     Control: I/O- Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- 
>> ParErr- Stepping- SERR- FastB2B- DisINTx-
>>     Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- 
>> <TAbort- <MAbort- >SERR- <PERR- INTx-
>>     Interrupt: pin A routed to IRQ 11
>>     Region 0: Memory at <ignored> (64-bit, prefetchable) [disabled]
>>     Region 2: Memory at b0000000 (64-bit, prefetchable) [disabled] 
>> [size=2M]
> The BIOS is not able to assign resources to one of the VGA adapters 
> when there are more than eight installed.
> You could try with pci=realloc, but I doubt that there is much we can 
> do in the operating system when the BIOS messed things up like that.
> What we should do is to prevent amdgpu from crashing so badly, e.g. 
> allow to cleanly continue with the working hardware even when one of 
> the devices doesn't work.
>> when I load in amdgpu, everything froze, so I don't have the log.
> You can work around that using netconsole, see 
> Documentation/networking/netconsole.txt.
> Going to try to fix that by just using the screen shot you send 
> earlier, but it would be better if I can get a full log.
> Regards,
> Christian.
> Am 19.02.2018 um 12:55 schrieb Joseph Wang:
>> Here is the lspci without amdgpu loaded.  when I load in amdgpu, 
>> everything froze, so I don't have the log.
>> _______________________________________________
>> amd-gfx mailing list

Patch hide | download patch | download mbox

From 9a90837362d8620d247d943b0e8ed93250a3ad3c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <>
Date: Mon, 19 Feb 2018 14:17:18 +0100
Subject: [PATCH] PCI: stop crashing in pci_release_resource
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Is it entirely possible that the BIOS wasn't able to assign resources to
a device. In this case don't crash in pci_release_resource() when we try
to resize the resource.

Signed-off-by: Christian König <>
 drivers/pci/setup-res.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c
index e815111f3f81..fd72c87a9b72 100644
--- a/drivers/pci/setup-res.c
+++ b/drivers/pci/setup-res.c
@@ -401,6 +401,9 @@  void pci_release_resource(struct pci_dev *dev, int resno)
 	struct resource *res = dev->resource + resno;
+	if (!res->parent)
+		return;
 	dev_info(&dev->dev, "BAR %d: releasing %pR\n", resno, res);
 	res->end = resource_size(res) - 1;