From: Tiezhu Yang <yangtiezhu@loongson.cn>
To: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
Huacai Chen <chenhc@lemote.com>,
Bjorn Helgaas <bhelgaas@google.com>, Jens Axboe <axboe@kernel.dk>,
linux-mips <linux-mips@vger.kernel.org>,
linux-pci <linux-pci@vger.kernel.org>,
linux-ide <linux-ide@vger.kernel.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
Xuefeng Li <lixuefeng@loongson.cn>
Subject: Re: 回复:[PATCH 4/6] MIPS: Loongson: Add DMA support for 7A1000
Date: Tue, 10 Mar 2020 10:40:18 +0800 [thread overview]
Message-ID: <a2f747e6-500e-0a4a-9e63-397b4aa0191f@loongson.cn> (raw)
In-Reply-To: <170be7768e3.c920b6d62296.161774605460069850@flygoat.com>
On 03/09/2020 04:44 PM, Jiaxun Yang wrote:
>
> ---- 在 星期一, 2020-03-09 16:23:24 Tiezhu Yang <yangtiezhu@loongson.cn> 撰写 ----
> > Implement __phys_to_dma() and __dma_to_phys() according to the
> > node id offset in 7A1000 DMA route config register.
>
> That design shocked me a lot. And It is known that some firmware didn't configure
> HT Recieve window correctly to make it work. So probably for mainline kernel,
> just set DMA_MASK to limit LS7A DMA address to Node0 would be a better Option?
Hi Jiaxun,
Let me rethink it and find a proper way.
>
> > Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> > ---
> > arch/mips/include/asm/mach-loongson64/boot_param.h | 1 +
> > arch/mips/loongson64/dma.c | 49 +++++++++++++++++++++-
> > arch/mips/loongson64/init.c | 13 ++++++
> > 3 files changed, 61 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/mips/include/asm/mach-loongson64/boot_param.h b/arch/mips/include/asm/mach-loongson64/boot_param.h
> > index 225a563..60e7a7e 100644
> > --- a/arch/mips/include/asm/mach-loongson64/boot_param.h
> > +++ b/arch/mips/include/asm/mach-loongson64/boot_param.h
> > @@ -218,5 +218,6 @@ struct loongson_system_configuration {
> > extern struct efi_memory_map_loongson *loongson_memmap;
> > extern struct loongson_system_configuration loongson_sysconf;
> > extern struct board_devices *eboard;
> > +extern u32 node_id_offset;
> >
> > #endif
> > diff --git a/arch/mips/loongson64/dma.c b/arch/mips/loongson64/dma.c
> > index 5e86635..997c257 100644
> > --- a/arch/mips/loongson64/dma.c
> > +++ b/arch/mips/loongson64/dma.c
> > @@ -2,24 +2,69 @@
> > #include <linux/dma-direct.h>
> > #include <linux/init.h>
> > #include <linux/swiotlb.h>
> > +#include <boot_param.h>
> >
> > -dma_addr_t __phys_to_dma(struct device *dev, phys_addr_t paddr)
> > +struct loongson_dma_ops {
> > + dma_addr_t (*phys_to_dma)(struct device *dev, phys_addr_t paddr);
> > + phys_addr_t (*dma_to_phys)(struct device *dev, dma_addr_t daddr);
> > +};
> > +
> > +struct loongson_dma_ops loongson_dma;
> > +
> > +dma_addr_t __rs780e_phys_to_dma(struct device *dev, phys_addr_t paddr)
> > {
> > /* We extract 2bit node id (bit 44~47, only bit 44~45 used now) from
> > * Loongson-3's 48bit address space and embed it into 40bit */
> > long nid = (paddr >> 44) & 0x3;
> > +
> > return ((nid << 44) ^ paddr) | (nid << 37);
> > }
> >
> > -phys_addr_t __dma_to_phys(struct device *dev, dma_addr_t daddr)
> > +dma_addr_t __ls7a_phys_to_dma(struct device *dev, phys_addr_t paddr)
> > +{
> > + long nid = (paddr >> 44) & 0x3;
> > +
> > + return ((nid << 44) ^ paddr) | (nid << (36 + node_id_offset));
> > +}
> > +
> > +
> > +dma_addr_t __phys_to_dma(struct device *dev, phys_addr_t paddr)
> > +{
> > + return loongson_dma.phys_to_dma(dev, paddr);
> > +}
> > +
> > +phys_addr_t __rs780e_dma_to_phys(struct device *dev, dma_addr_t daddr)
> > {
> > /* We extract 2bit node id (bit 44~47, only bit 44~45 used now) from
> > * Loongson-3's 48bit address space and embed it into 40bit */
> > long nid = (daddr >> 37) & 0x3;
> > +
> > return ((nid << 37) ^ daddr) | (nid << 44);
> > }
> >
> > +phys_addr_t __ls7a_dma_to_phys(struct device *dev, dma_addr_t daddr)
> > +{
> > + long nid = (daddr >> (36 + node_id_offset)) & 0x3;
> > +
> > + return ((nid << (36 + node_id_offset)) ^ daddr) | (nid << 44);
> > +}
> > +
> > +phys_addr_t __dma_to_phys(struct device *dev, dma_addr_t daddr)
> > +{
> > + return loongson_dma.dma_to_phys(dev, daddr);
> > +}
> > +
> > void __init plat_swiotlb_setup(void)
> > {
> > swiotlb_init(1);
> > +
> > + if (strstr(eboard->name, "780E")) {
> > + loongson_dma.phys_to_dma = __rs780e_phys_to_dma;
> > + loongson_dma.dma_to_phys = __rs780e_dma_to_phys;
> > + }
> > +
> > + if (strstr(eboard->name, "7A1000")) {
> > + loongson_dma.phys_to_dma = __ls7a_phys_to_dma;
> > + loongson_dma.dma_to_phys = __ls7a_dma_to_phys;
> > + }
> > }
> > diff --git a/arch/mips/loongson64/init.c b/arch/mips/loongson64/init.c
> > index 5ac1a0f..dd8463d 100644
> > --- a/arch/mips/loongson64/init.c
> > +++ b/arch/mips/loongson64/init.c
> > @@ -12,6 +12,11 @@
> > #include <asm/fw/fw.h>
> >
> > #include <loongson.h>
> > +#include <boot_param.h>
> > +
> > +#define NODE_ID_OFFSET_ADDR 0x90000E001001041CULL
> > +
> > +u32 node_id_offset;
> >
> > static void __init mips_nmi_setup(void)
> > {
> > @@ -23,6 +28,11 @@ static void __init mips_nmi_setup(void)
> > flush_icache_range((unsigned long)base, (unsigned long)base + 0x80);
> > }
> >
> > +static void ls7a_early_config(void)
> > +{
> > + node_id_offset = (*(u32 *)NODE_ID_OFFSET_ADDR >> 8) & 0x1F;
>
> Please avoid raw pointer. Use readl/writel instead.
OK, I will do it.
Thanks,
Tiezhu Yang
>
> > +}
> > +
> > void __init prom_init(void)
> > {
> > fw_init_cmdline();
> > @@ -32,6 +42,9 @@ void __init prom_init(void)
> > set_io_port_base((unsigned long)
> > ioremap(LOONGSON_PCIIO_BASE, LOONGSON_PCIIO_SIZE));
> >
> > + if (strstr(eboard->name, "7A1000"))
> > + ls7a_early_config();
> > +
> > prom_init_numa_memory();
> >
> > /* Hardcode to CPU UART 0 */
> > --
> > 2.1.0
> >
> >
next prev parent reply other threads:[~2020-03-10 2:40 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-09 8:23 [PATCH 0/6] Add basic support for Loongson 7A1000 bridge chip Tiezhu Yang
2020-03-09 8:23 ` [PATCH 1/6] PCI: Add Loongson vendor ID and 7A1000 device IDs Tiezhu Yang
2020-03-09 8:50 ` 回复:[PATCH " Jiaxun Yang
2020-03-10 2:31 ` Tiezhu Yang
2020-03-10 2:37 ` Jiaxun Yang
2020-03-10 9:28 ` Tiezhu Yang
2020-03-09 8:23 ` [PATCH 2/6] AHCI: Add support for Loongson 7A1000 SATA controller Tiezhu Yang
2020-03-09 8:23 ` [PATCH 3/6] MIPS: Loongson: Use firmware arguments to get board name Tiezhu Yang
2020-03-09 8:36 ` 回复:[PATCH " Jiaxun Yang
2020-03-10 2:38 ` Tiezhu Yang
2020-03-09 8:23 ` [PATCH 4/6] MIPS: Loongson: Add DMA support for 7A1000 Tiezhu Yang
2020-03-09 8:44 ` 回复:[PATCH " Jiaxun Yang
2020-03-10 2:40 ` Tiezhu Yang [this message]
2020-03-13 8:24 ` [PATCH " Christoph Hellwig
2020-03-13 9:03 ` Tiezhu Yang
2020-03-09 8:23 ` [PATCH 5/6] MIPS: Loongson: Add PCI " Tiezhu Yang
2020-03-09 8:47 ` 回复:[PATCH " Jiaxun Yang
2020-03-10 3:03 ` Tiezhu Yang
2020-03-09 8:23 ` [PATCH 6/6] MIPS: Loongson: Add support for 7A1000 interrupt controller Tiezhu Yang
2020-03-09 9:07 ` 回复:[PATCH " Jiaxun Yang
2020-03-10 4:28 ` Tiezhu Yang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a2f747e6-500e-0a4a-9e63-397b4aa0191f@loongson.cn \
--to=yangtiezhu@loongson.cn \
--cc=axboe@kernel.dk \
--cc=bhelgaas@google.com \
--cc=chenhc@lemote.com \
--cc=jiaxun.yang@flygoat.com \
--cc=linux-ide@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mips@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lixuefeng@loongson.cn \
--cc=tsbogend@alpha.franken.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).