ntb.lists.linux.dev archive mirror
 help / color / mirror / Atom feed
* Problem with 8G dma memory allocation
@ 2023-10-02 22:15 am jom
  0 siblings, 0 replies; only message in thread
From: am jom @ 2023-10-02 22:15 UTC (permalink / raw)
  To: ntb

I have a Supermicro x10drs-2u board, Intel Xeon E5, 2600v3, with PLX
NTB switch. I have installed Centos 7 and mainline linux kernel 6.5.3
with added PLX driver from TrueNAS. There 2 main boards with PLX ntb
devices already attached in hardware through a bus named M2

CMA and DMA are active and I can reserve like 8G, CMA memory at 4G.
There are other network devices using CMA and they are working. I know
that ntb_hw_plx.ko by Truenas needs 8G memory space

When i load ntb_transport.ko module, the related kernel thread crashes
by the following error message:

Unable to alloc MW buff of size 8,589,934,592

The problem is with dma_alloc_coherent() which fails and has errors
and it seems that it can not find any CMA reserved memory as the crash
dump reports (by dmesg) invalid operation in _alloc_pages() at last
and the "cma" term is not seen in the crash dump .

for CMA alignment i chose 2, and 2 areas and 8G size. what size and
alignment should i set by the kernel config variables?

Is there any kernel i can use to drive PLX ntb bridge device? Kernel
detects several PLX devices regarding this switch like PEX8717,
PEX8732 and 87d0 and ntb_hw_plx.ko is being attached to 8732 network
device

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-10-02 22:15 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-10-02 22:15 Problem with 8G dma memory allocation am jom

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).