Message ID | 1548227352-14910-2-git-send-email-masonccyang@mxic.com.tw (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | spi: Add Renesas R-Car Gen3 RPC-IF SPI driver | expand |
Hello! On 01/23/2019 10:09 AM, Mason Yang wrote: > Add a driver for Renesas R-Car Gen3 RPC-IF SPI controller. > > Signed-off-by: Mason Yang <masonccyang@mxic.com.tw> > Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> [...] > diff --git a/drivers/spi/spi-renesas-rpc.c b/drivers/spi/spi-renesas-rpc.c > new file mode 100644 > index 0000000..0127f25 > --- /dev/null > +++ b/drivers/spi/spi-renesas-rpc.c > @@ -0,0 +1,805 @@ > +// SPDX-License-Identifier: GPL-2.0 > +// > +// Copyright (C) 2018 ~ 2019 Renesas Solutions Corp. > +// Copyright (C) 2018 Macronix International Co., Ltd. > +// > +// R-Car Gen3 RPC-IF SPI/QSPI/Octa driver Do we really support Octa? > +// > +// Authors: > +// Mason Yang <masonccyang@mxic.com.tw> > +// [...] > +#include <asm/unaligned.h> [...] > + } else if (rx_buf) { > + // > + // RPC-IF spoils the data for the commands without an address > + // phase (like RDID) in the manual mode, so we'll have to work > + // around this issue by using the external address space read > + // mode instead. > + // > + if (!(smenr & RPC_SMENR_ADE(0xf))) { > + regmap_update_bits(rpc->regmap, RPC_CMNCR, > + RPC_CMNCR_MD, 0); > + regmap_write(rpc->regmap, RPC_DRCR, > + RPC_DRCR_RBURST(32) | RPC_DRCR_RBE); > + regmap_write(rpc->regmap, RPC_DREAR, RPC_DREAR_EAC(1)); > + regmap_write(rpc->regmap, RPC_DRCMR, rpc->cmd); > + regmap_write(rpc->regmap, RPC_DRDMCR, rpc->dummy); > + regmap_write(rpc->regmap, RPC_DROPR, 0); > + regmap_write(rpc->regmap, RPC_DRENR, smenr); > + memcpy_fromio(rx_buf, rpc->dirmap, rpc->xferlen); Just noticed: if rpc->dirmap is NULL (which is now allowed), we have a kernel oops here. :-( [...] > +static int rpc_spi_probe(struct platform_device *pdev) > +{ > + struct spi_controller *ctlr; > + struct resource *res; > + struct rpc_spi *rpc; > + struct device_node *rpc_if; Don't think that's a good name. Why not call it 'flash'? > + int ret; > + > + rpc_if = of_get_next_child(pdev->dev.of_node, NULL); > + if (!rpc_if) { > + dev_warn(&pdev->dev, "no spi-nor device node found\n"); How about "no flash node found"? > + return -ENODEV; > + } > + > + ret = of_device_is_compatible(rpc_if, "jedec,spi-nor"); > + if (!ret) { > + dev_warn(&pdev->dev, "no compatible spi-nor device found\n"); "no SPI-NOR device found"? [...] > + ret = spi_register_controller(ctlr); > + if (ret) { > + dev_err(&pdev->dev, "spi_register_controller failed\n"); > + goto err_put_ctlr; > + } > + return 0; > + > +err_put_ctlr: > + spi_controller_put(ctlr); > + pm_runtime_disable(&pdev->dev); > + > + return ret; > +} [...] > +static const struct of_device_id rpc_spi_of_ids[] = { > + { .compatible = "renesas,r8a77995-rpc", }, Why (if it has no differences with the generic gen3 value below)? Please remove. > + { .compatible = "renesas,rcar-gen3-rpc", }, > + { /* sentinel */ } > +}; > +MODULE_DEVICE_TABLE(of, rpc_spi_of_ids); [...] MBR, Sergei
On 1/23/19 8:09 AM, Mason Yang wrote: > Add a driver for Renesas R-Car Gen3 RPC-IF SPI controller. > > Signed-off-by: Mason Yang <masonccyang@mxic.com.tw> > Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> [...] > +#define RPC_CMNCR 0x0000 // R/W Is there any reason for using those horrible C++ comments ? [...] > +module_platform_driver(rpc_spi_driver); RPC is not a SPI controller, it's a SPI and HF controller. Also, how difficult will it be to add the HF support ? > +MODULE_AUTHOR("Mason Yang <masonccyang@mxic.com.tw>"); > +MODULE_DESCRIPTION("Renesas R-Car Gen3 RPC-IF SPI controller driver"); > +MODULE_LICENSE("GPL v2"); >
On 1/24/19 3:23 AM, masonccyang@mxic.com.tw wrote: > Hi Marek, Hi, >> "Marek Vasut" <marek.vasut@gmail.com> >> 2019/01/24 上午 09:54 >> >> >> > +#define RPC_CMNCR 0x0000 // R/W >> >> Is there any reason for using those horrible C++ comments ? > > By Mark's comments for the SPDX header needs to be C++ style and > I patch the whole RPC driver comments using C++ style otherwise it looks > messy. I think the C++ comments should only be applied to the SPDX identifier, maybe the header, but not the entire file. >> [...] >> >> > +module_platform_driver(rpc_spi_driver); >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> Also, how difficult will it be to add the HF support ? > > One of my customers needs RPC SPI driver for our company's > Octal-Flash,MX25UW51245G. > We don't have HF product and hope you could understanding. I am worried that when we need to add RPC HF support (which is what all boards but the D3 Draak use), we will have to rewrite the entire driver and/or convert it to MFD and that would be a tremendous undertaking. I'd prefer to have the driver ready for the HF addition before it's accepted upstream.
Hi Mason, On Thu, Jan 24, 2019 at 3:07 AM <masonccyang@mxic.com.tw> wrote: > > "Sergei Shtylyov" <sergei.shtylyov@cogentembedded.com> > > 2019/01/24 上午 02:04 > > > +static const struct of_device_id rpc_spi_of_ids[] = { > > > + { .compatible = "renesas,r8a77995-rpc", }, > > > > Why (if it has no differences with the generic gen3 value below)? > > Please remove. > > By Geert's comments in dts, > > rpc: rpc@ee200000 { > compatible = "renesas,r8a7795-rpc", "renesas,rcar-gen3-rpc"; > reg = <0 0xee200000 0 0x200>, <0 0x08000000 0 0x4000000>, > <0 0xee208000 0 0x100>; > > I modified it to "renesas,r8a77995-rpc" because this is Soc I developed and validated. There's a difference between compatible values in DT and in drivers: - DT should list all compatible values that are applicable, from most-specific to least-specific, - Drivers should match against the least-specific compatible value that is sufficient to get the job done. In this case, that's "renesas,rcar-gen3-rpc". The driver will probably have to be extended to match against "renesas,r8a77970-rpc" when R-Car V3M support is added, as RPC on V3M has an extra register that must be set, and thus cannot be declared compatible with "renesas,rcar-gen3-rpc". I hope this explanation makes it clearer. Gr{oetje,eeting}s, Geert
On 01/24/2019 05:06 AM, masonccyang@mxic.com.tw wrote: >> > +// SPDX-License-Identifier: GPL-2.0 >> > +// >> > +// Copyright (C) 2018 ~ 2019 Renesas Solutions Corp. >> > +// Copyright (C) 2018 Macronix International Co., Ltd. >> > +// >> > +// R-Car Gen3 RPC-IF SPI/QSPI/Octa driver >> >> Do we really support Octa? > > yes and it needs to patch spi-nor layer and RPC driver for sure. It? >> > +// >> > +// Authors: >> > +// Mason Yang <masonccyang@mxic.com.tw> >> > +// >> [...] >> > +#include <asm/unaligned.h> >> [...] >> > + } else if (rx_buf) { >> > + // >> > + // RPC-IF spoils the data for the commands without an address >> > + // phase (like RDID) in the manual mode, so we'll have to work >> > + // around this issue by using the external address space read >> > + // mode instead. >> > + // >> > + if (!(smenr & RPC_SMENR_ADE(0xf))) { >> > + regmap_update_bits(rpc->regmap, RPC_CMNCR, >> > + RPC_CMNCR_MD, 0); >> > + regmap_write(rpc->regmap, RPC_DRCR, >> > + RPC_DRCR_RBURST(32) | RPC_DRCR_RBE); >> > + regmap_write(rpc->regmap, RPC_DREAR, RPC_DREAR_EAC(1)); >> > + regmap_write(rpc->regmap, RPC_DRCMR, rpc->cmd); >> > + regmap_write(rpc->regmap, RPC_DRDMCR, rpc->dummy); >> > + regmap_write(rpc->regmap, RPC_DROPR, 0); >> > + regmap_write(rpc->regmap, RPC_DRENR, smenr); >> > + memcpy_fromio(rx_buf, rpc->dirmap, rpc->xferlen); >> >> Just noticed: if rpc->dirmap is NULL (which is now allowed), we >> have a kernel oops >> here. :-( > > Do you patch your spi-mem.c ? What do yuo mean? The only extra patch I have to apply now to renesas.git repo's devel branch is "mtd: devices: m25p80: Use the spi-mem dirmap API"... > both dirmap_create() in your rpc driver and spi-mem layer should handle it well. I just saw that you don't fail the driver's probe() method if devm_ioremap_resource() fails for the "dirmap" and "wbuf" resources... >> [...] >> > +static const struct of_device_id rpc_spi_of_ids[] = { >> > + { .compatible = "renesas,r8a77995-rpc", }, >> >> Why (if it has no differences with the generic gen3 value below)? >> Please remove. > > By Geert's comments in dts, > > rpc: rpc@ee200000 { > compatible = "renesas,r8a7795-rpc", "renesas,rcar-gen3-rpc"; > reg = <0 0xee200000 0 0x200>, <0 0x08000000 0 0x4000000>, > <0 0xee208000 0 0x100>; > > I modified it to "renesas,r8a77995-rpc" because this is Soc I developed and validated. Hopefully this matter is cleared up by Geert... :-) > thanks, > Mason MBR, Sergei
On 01/24/2019 05:06 AM, masonccyang@mxic.com.tw wrote: >> > +// SPDX-License-Identifier: GPL-2.0 >> > +// >> > +// Copyright (C) 2018 ~ 2019 Renesas Solutions Corp. >> > +// Copyright (C) 2018 Macronix International Co., Ltd. >> > +// >> > +// R-Car Gen3 RPC-IF SPI/QSPI/Octa driver >> >> Do we really support Octa? > > yes and it needs to patch spi-nor layer and RPC driver for sure. It? >> > +// >> > +// Authors: >> > +// Mason Yang <masonccyang@mxic.com.tw> >> > +// >> [...] >> > +#include <asm/unaligned.h> >> [...] >> > + } else if (rx_buf) { >> > + // >> > + // RPC-IF spoils the data for the commands without an address >> > + // phase (like RDID) in the manual mode, so we'll have to work >> > + // around this issue by using the external address space read >> > + // mode instead. >> > + // >> > + if (!(smenr & RPC_SMENR_ADE(0xf))) { >> > + regmap_update_bits(rpc->regmap, RPC_CMNCR, >> > + RPC_CMNCR_MD, 0); >> > + regmap_write(rpc->regmap, RPC_DRCR, >> > + RPC_DRCR_RBURST(32) | RPC_DRCR_RBE); >> > + regmap_write(rpc->regmap, RPC_DREAR, RPC_DREAR_EAC(1)); >> > + regmap_write(rpc->regmap, RPC_DRCMR, rpc->cmd); >> > + regmap_write(rpc->regmap, RPC_DRDMCR, rpc->dummy); >> > + regmap_write(rpc->regmap, RPC_DROPR, 0); >> > + regmap_write(rpc->regmap, RPC_DRENR, smenr); >> > + memcpy_fromio(rx_buf, rpc->dirmap, rpc->xferlen); >> >> Just noticed: if rpc->dirmap is NULL (which is now allowed), we >> have a kernel oops >> here. :-( > > Do you patch your spi-mem.c ? What do you mean? The only extra patch I have to apply now to renesas.git repo's devel branch is "mtd: devices: m25p80: Use the spi-mem dirmap API"... > both dirmap_create() in your rpc driver and spi-mem layer should handle it well. I just saw that you don't fail the driver's probe() method if devm_ioremap_resource() fails for the "dirmap" and "wbuf" resources... >> [...] >> > +static const struct of_device_id rpc_spi_of_ids[] = { >> > + { .compatible = "renesas,r8a77995-rpc", }, >> >> Why (if it has no differences with the generic gen3 value below)? >> Please remove. > > By Geert's comments in dts, > > rpc: rpc@ee200000 { > compatible = "renesas,r8a7795-rpc", "renesas,rcar-gen3-rpc"; > reg = <0 0xee200000 0 0x200>, <0 0x08000000 0 0x4000000>, > <0 0xee208000 0 0x100>; > > I modified it to "renesas,r8a77995-rpc" because this is Soc I developed and validated. Hopefully this matter is cleared up by Geert... :-) > thanks, > Mason MBR, Sergei
On 01/24/2019 05:23 AM, masonccyang@mxic.com.tw wrote: >> [...] >> >> > +module_platform_driver(rpc_spi_driver); >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> Also, how difficult will it be to add the HF support ? > > One of my customers needs RPC SPI driver for our company's Octal-Flash,MX25UW51245G. > We don't have HF product and hope you could understanding. I'm looking into the HF support (we have a driver internally). > > thanks & best regards, > Mason MBR, Sergei
On 1/24/19 7:28 AM, masonccyang@mxic.com.tw wrote: > Hi Marek, Hi, >> "Marek Vasut" <marek.vasut@gmail.com> >> 2019/01/24 上午 11:14 >> >> >> >> > +module_platform_driver(rpc_spi_driver); >> >> >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> >> >> Also, how difficult will it be to add the HF support ? >> > >> > One of my customers needs RPC SPI driver for our company's >> > Octal-Flash,MX25UW51245G. >> > We don't have HF product and hope you could understanding. >> >> I am worried that when we need to add RPC HF support (which is what all >> boards but the D3 Draak use), we will have to rewrite the entire driver >> and/or convert it to MFD and that would be a tremendous undertaking. I'd >> prefer to have the driver ready for the HF addition before it's accepted >> upstream. >> > > I think maybe your concerned would be happened only if HF driver goes with > spi-mem layer. > > A comment for HF from Daniel Fishman. FYR. > > https://www.quora.com/What-is-a-hyper-flash-memory-and-how-is-it-different-from-normal-flash-memory I have a decent idea what HF and SPI NOR are, since I wrote the RPC driver for both HF and SPI mode for U-Boot (as I mentioned earlier). The HF in Linux would use the CFI NOR part of MTD framework. My concern is that when we need to add HF support into this driver, this driver will have to be basically rewritten, since the architecture won't allow for that. I'd like to avoid that, since the majority of Gen3 boards, expect for the D3 Draak, use RPC in HF mode.
On 1/28/19 2:38 AM, masonccyang@mxic.com.tw wrote: > Hi Marek, Hi, >> "Marek Vasut" <marek.vasut@gmail.com> >> >> >> > +module_platform_driver(rpc_spi_driver); >> >> >> >> >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> >> >> >> >> Also, how difficult will it be to add the HF support ? >> >> > >> >> > One of my customers needs RPC SPI driver for our company's >> >> > Octal-Flash,MX25UW51245G. >> >> > We don't have HF product and hope you could understanding. >> >> >> >> I am worried that when we need to add RPC HF support (which is what all >> >> boards but the D3 Draak use), we will have to rewrite the entire driver >> >> and/or convert it to MFD and that would be a tremendous > undertaking. I'd >> >> prefer to have the driver ready for the HF addition before it's > accepted >> >> upstream. >> >> >> > >> > I think maybe your concerned would be happened only if HF driver > goes with >> > spi-mem layer. >> > >> > A comment for HF from Daniel Fishman. FYR. >> > >> > https://www.quora.com/What-is-a-hyper-flash-memory-and-how-is-it- >> different-from-normal-flash-memory >> >> I have a decent idea what HF and SPI NOR are, since I wrote the RPC >> driver for both HF and SPI mode for U-Boot (as I mentioned earlier). >> >> The HF in Linux would use the CFI NOR part of MTD framework. My concern >> is that when we need to add HF support into this driver, this driver >> will have to be basically rewritten, since the architecture won't allow >> for that. I'd like to avoid that, since the majority of Gen3 boards, >> expect for the D3 Draak, use RPC in HF mode. > > FYI~ > > MX25UW51245g(64MByte Octa) S26KL512S(64MByte HF) > 8 IO 8 IO > 200MHz DDR@1.8v 166MHz DDR@1.8v > > support Read-while-write Not support > good for OTA,etc > powerful application What does that mean ?
On 1/29/19 3:26 AM, masonccyang@mxic.com.tw wrote: > Hi Marek, Hi, >> >> "Marek Vasut" <marek.vasut@gmail.com> >> >> >> >> > +module_platform_driver(rpc_spi_driver); >> >> >> >> >> >> >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> >> >> >> >> >> >> Also, how difficult will it be to add the HF support ? >> >> >> > >> >> >> > One of my customers needs RPC SPI driver for our company's >> >> >> > Octal-Flash,MX25UW51245G. >> >> >> > We don't have HF product and hope you could understanding. >> >> >> >> >> >> I am worried that when we need to add RPC HF support (which is > what all >> >> >> boards but the D3 Draak use), we will have to rewrite the entire > driver >> >> >> and/or convert it to MFD and that would be a tremendous >> > undertaking. I'd >> >> >> prefer to have the driver ready for the HF addition before it's >> > accepted >> >> >> upstream. >> >> >> >> >> > >> >> > I think maybe your concerned would be happened only if HF driver >> > goes with >> >> > spi-mem layer. >> >> > >> >> > A comment for HF from Daniel Fishman. FYR. >> >> > >> >> > https://www.quora.com/What-is-a-hyper-flash-memory-and-how-is-it- >> >> different-from-normal-flash-memory >> >> >> >> I have a decent idea what HF and SPI NOR are, since I wrote the RPC >> >> driver for both HF and SPI mode for U-Boot (as I mentioned earlier). >> >> >> >> The HF in Linux would use the CFI NOR part of MTD framework. My concern >> >> is that when we need to add HF support into this driver, this driver >> >> will have to be basically rewritten, since the architecture won't allow >> >> for that. I'd like to avoid that, since the majority of Gen3 boards, >> >> expect for the D3 Draak, use RPC in HF mode. >> > >> > FYI~ >> > >> > MX25UW51245g(64MByte Octa) S26KL512S(64MByte HF) >> > 8 IO 8 IO >> > 200MHz DDR@1.8v 166MHz DDR@1.8v >> > >> > support Read-while-write Not support >> > good for OTA,etc >> > powerful application >> >> What does that mean ? > > I have no idea why would you say "since the majority of Gen3 boards use > RPC in HF mode" ? Well, the H3/M3W/M3N S-X(S) and the H3/M3 ULCB and E3 Ebisu all boot from HF. Only the D3 Draak uses QSPI NOR. > So far as I know that HF is provided by Cypress only and > any mass production product use the component which is provided by only > one provider > will be a big risk. > > Compare to HF, there are more provider of SPI/Octa could support the > mass production product > as their second provider. > > In addition, from the technical points of view, mx25uw51245g is more > powerful than HF and > good for complicate user application, i.e., OTA and so on. Did you consider protocol overhead too ? I don't think you can compare them just by raw numbers of pins and bus frequency. Note that over-the-air update (if that's what you mean by OTA) is completely separate from the underlying storage device. > I think customer have more choice for their flash memory component. Note that none of this is really relevant to my concerns above regarding HF support. This driver should be implemented as MFD driver and the SPI part should use the MFD core part , just like the future HF part.
On 1/30/19 3:22 AM, masonccyang@mxic.com.tw wrote: > Hi Marek, Hi, >> "Marek Vasut" <marek.vasut@gmail.com> >> 2019/01/29 下午 12:45 >> >> To >> >> masonccyang@mxic.com.tw, >> >> cc >> >> bbrezillon@kernel.org, broonie@kernel.org, "Geert Uytterhoeven" >> <geert+renesas@glider.be>, "Simon Horman" <horms@verge.net.au>, >> juliensu@mxic.com.tw, linux-kernel@vger.kernel.org, linux-renesas- >> soc@vger.kernel.org, linux-spi@vger.kernel.org, >> sergei.shtylyov@cogentembedded.com, zhengxunli@mxic.com.tw >> >> Subject >> >> Re: [PATCH v7 1/2] spi: Add Renesas R-Car Gen3 RPC-IF SPI controller > driver >> >> On 1/29/19 3:26 AM, masonccyang@mxic.com.tw wrote: >> > Hi Marek, >> >> Hi, >> >> >> >> "Marek Vasut" <marek.vasut@gmail.com> >> >> >> >> >> > +module_platform_driver(rpc_spi_driver); >> >> >> >> >> >> >> >> >> >> RPC is not a SPI controller, it's a SPI and HF controller. >> >> >> >> >> >> >> >> >> >> Also, how difficult will it be to add the HF support ? >> >> >> >> > >> >> >> >> > One of my customers needs RPC SPI driver for our company's >> >> >> >> > Octal-Flash,MX25UW51245G. >> >> >> >> > We don't have HF product and hope you could understanding. >> >> >> >> >> >> >> >> I am worried that when we need to add RPC HF support (which is >> > what all >> >> >> >> boards but the D3 Draak use), we will have to rewrite the entire >> > driver >> >> >> >> and/or convert it to MFD and that would be a tremendous >> >> > undertaking. I'd >> >> >> >> prefer to have the driver ready for the HF addition before it's >> >> > accepted >> >> >> >> upstream. >> >> >> >> >> >> >> > >> >> >> > I think maybe your concerned would be happened only if HF driver >> >> > goes with >> >> >> > spi-mem layer. >> >> >> > >> >> >> > A comment for HF from Daniel Fishman. FYR. >> >> >> > >> >> >> > https://www.quora.com/What-is-a-hyper-flash-memory-and-how-is-it- >> >> >> different-from-normal-flash-memory >> >> >> >> >> >> I have a decent idea what HF and SPI NOR are, since I wrote the RPC >> >> >> driver for both HF and SPI mode for U-Boot (as I mentioned earlier). >> >> >> >> >> >> The HF in Linux would use the CFI NOR part of MTD framework. My > concern >> >> >> is that when we need to add HF support into this driver, this driver >> >> >> will have to be basically rewritten, since the architecture > won't allow >> >> >> for that. I'd like to avoid that, since the majority of Gen3 boards, >> >> >> expect for the D3 Draak, use RPC in HF mode. >> >> > >> >> > FYI~ >> >> > >> >> > MX25UW51245g(64MByte Octa) S26KL512S(64MByte HF) >> >> > 8 IO 8 IO >> >> > 200MHz DDR@1.8v 166MHz DDR@1.8v >> >> > >> >> > support Read-while-write Not support >> >> > good for OTA,etc >> >> > powerful application >> >> >> >> What does that mean ? >> > >> > I have no idea why would you say "since the majority of Gen3 boards use >> > RPC in HF mode" ? >> >> Well, the H3/M3W/M3N S-X(S) and the H3/M3 ULCB and E3 Ebisu all boot >> from HF. Only the D3 Draak uses QSPI NOR. > > It's understandable because mx25uw51245g is a new product and it has been > adopted by Renesas’ Automotive Instrument Cluster RH850/D1M1A MCU. The aforementioned boards are not going away however. There's too many users to ignore those. > We also have patched R-Car's BL for booting from Octa-Flash as bellow log: > > NOTICE: BL2: R-Car D3 Initial Program Loader(CA53) Rev.0.5.1 > NOTICE: BL2: PRR is R-Car D3 Ver1.0 > NOTICE: BL2: Boot device is MXIC_OctaFlash > NOTICE: BL2: LCM state is CM > NOTICE: BL2: DDR3L-1866(rev.0.02) > NOTICE: BL2: QoS is default setting(rev.0.07) > NOTICE: BL2: v1.3(release): > NOTICE: BL2: Built : 09:56:31, Sep 26 2018 > NOTICE: BL2: Normal boot > NOTICE: BL2: dst=0xe63111f0 src=0x8180000 len=512(0x200) > NOTICE: BL2: dst=0x43f00000 src=0x8180400 len=6144(0x1800) > NOTICE: BL2: dst=0x44000000 src=0x81c0000 len=65536(0x10000) > NOTICE: BL2: dst=0x50000000 src=0x8640000 len=1048576(0x100000) This looks like a very old ATF version. Anyway, please submit those patches upstream, they should be generic. >> > So far as I know that HF is provided by Cypress only and >> > any mass production product use the component which is provided by only >> > one provider >> > will be a big risk. >> > >> > Compare to HF, there are more provider of SPI/Octa could support the >> > mass production product >> > as their second provider. >> > >> > In addition, from the technical points of view, mx25uw51245g is more >> > powerful than HF and >> > good for complicate user application, i.e., OTA and so on. >> >> Did you consider protocol overhead too ? I don't think you can compare >> them just by raw numbers of pins and bus frequency. >> >> Note that over-the-air update (if that's what you mean by OTA) is >> completely separate from the underlying storage device. > > It's key feature of mx25uw51245g supports Read-while-Write capability > that allows read access from one memory bank while writing to another > memory bank. Note that this sales pitch is rather off-topic. >> > I think customer have more choice for their flash memory component. >> Note that none of this is really relevant to my concerns above regarding >> HF support. This driver should be implemented as MFD driver and the SPI >> part should use the MFD core part , just like the future HF part. >> > > Even though they are both internal NOR memory chip, is it a trend of Kernel > for the different HW controller and protocol ? I'm not sure I understand the question, but the HF behaves like a CFI NOR, that's why it should be operated via the CFI part of the MTD framework. The SPI NOR behaves, well, like a SPI NOR and so should be operated via the SPI NOR part of the MTD framework. However, a lot of the code operating the RPC is common for the HF and SPI NOR modes, which is why that part should be a MFD driver or some common core driver. > What about others,i.e., raw-NAND and spi-NAND ? I don't think the RPC supports those, does it ? > oh, a little bit leaving the topic ! > anyway, nice to talk to you about this R-Car Gen3 RPC driver. [...]
On Wed, 30 Jan 2019 08:15:25 +0100 Marek Vasut <marek.vasut@gmail.com> wrote: > >> > So far as I know that HF is provided by Cypress only and > >> > any mass production product use the component which is provided by only > >> > one provider > >> > will be a big risk. > >> > > >> > Compare to HF, there are more provider of SPI/Octa could support the > >> > mass production product > >> > as their second provider. > >> > > >> > In addition, from the technical points of view, mx25uw51245g is more > >> > powerful than HF and > >> > good for complicate user application, i.e., OTA and so on. > >> > >> Did you consider protocol overhead too ? I don't think you can compare > >> them just by raw numbers of pins and bus frequency. > >> > >> Note that over-the-air update (if that's what you mean by OTA) is > >> completely separate from the underlying storage device. > > > > It's key feature of mx25uw51245g supports Read-while-Write capability > > that allows read access from one memory bank while writing to another > > memory bank. > > Note that this sales pitch is rather off-topic. > Fully agree with Marek on that point. The discussion is about supporting both HF and SPI-MEM modes (or, at least have a plan to support HF at some point) not figuring out which option is better.
diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 9f89cb1..6ad1782 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -544,6 +544,12 @@ config SPI_RSPI help SPI driver for Renesas RSPI and QSPI blocks. +config SPI_RENESAS_RPC + tristate "Renesas R-Car Gen3 RPC-IF controller" + depends on ARCH_RENESAS || COMPILE_TEST + help + SPI driver for Renesas R-Car Gen3 RPC-IF. + config SPI_QCOM_QSPI tristate "QTI QSPI controller" depends on ARCH_QCOM diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile index f296270..9150732 100644 --- a/drivers/spi/Makefile +++ b/drivers/spi/Makefile @@ -84,6 +84,7 @@ obj-$(CONFIG_SPI_QUP) += spi-qup.o obj-$(CONFIG_SPI_ROCKCHIP) += spi-rockchip.o obj-$(CONFIG_SPI_RB4XX) += spi-rb4xx.o obj-$(CONFIG_SPI_RSPI) += spi-rspi.o +obj-$(CONFIG_SPI_RENESAS_RPC) += spi-renesas-rpc.o obj-$(CONFIG_SPI_S3C24XX) += spi-s3c24xx-hw.o spi-s3c24xx-hw-y := spi-s3c24xx.o spi-s3c24xx-hw-$(CONFIG_SPI_S3C24XX_FIQ) += spi-s3c24xx-fiq.o diff --git a/drivers/spi/spi-renesas-rpc.c b/drivers/spi/spi-renesas-rpc.c new file mode 100644 index 0000000..0127f25 --- /dev/null +++ b/drivers/spi/spi-renesas-rpc.c @@ -0,0 +1,805 @@ +// SPDX-License-Identifier: GPL-2.0 +// +// Copyright (C) 2018 ~ 2019 Renesas Solutions Corp. +// Copyright (C) 2018 Macronix International Co., Ltd. +// +// R-Car Gen3 RPC-IF SPI/QSPI/Octa driver +// +// Authors: +// Mason Yang <masonccyang@mxic.com.tw> +// + +#include <linux/clk.h> +#include <linux/io.h> +#include <linux/log2.h> +#include <linux/iopoll.h> +#include <linux/module.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/of.h> +#include <linux/regmap.h> +#include <linux/reset.h> +#include <linux/spi/spi.h> +#include <linux/spi/spi-mem.h> + +#include <asm/unaligned.h> + +#define RPC_CMNCR 0x0000 // R/W +#define RPC_CMNCR_MD BIT(31) +#define RPC_CMNCR_SFDE BIT(24) // undocumented bit but must be set +#define RPC_CMNCR_MOIIO3(val) (((val) & 0x3) << 22) +#define RPC_CMNCR_MOIIO2(val) (((val) & 0x3) << 20) +#define RPC_CMNCR_MOIIO1(val) (((val) & 0x3) << 18) +#define RPC_CMNCR_MOIIO0(val) (((val) & 0x3) << 16) +#define RPC_CMNCR_MOIIO_HIZ (RPC_CMNCR_MOIIO0(3) | RPC_CMNCR_MOIIO1(3) | \ + RPC_CMNCR_MOIIO2(3) | RPC_CMNCR_MOIIO3(3)) +#define RPC_CMNCR_IO3FV(val) (((val) & 0x3) << 14) // undocumented +#define RPC_CMNCR_IO2FV(val) (((val) & 0x3) << 12) // undocumented +#define RPC_CMNCR_IO0FV(val) (((val) & 0x3) << 8) +#define RPC_CMNCR_IOFV_HIZ (RPC_CMNCR_IO0FV(3) | RPC_CMNCR_IO2FV(3) | \ + RPC_CMNCR_IO3FV(3)) +#define RPC_CMNCR_BSZ(val) (((val) & 0x3) << 0) + +#define RPC_SSLDR 0x0004 // R/W +#define RPC_SSLDR_SPNDL(d) (((d) & 0x7) << 16) +#define RPC_SSLDR_SLNDL(d) (((d) & 0x7) << 8) +#define RPC_SSLDR_SCKDL(d) (((d) & 0x7) << 0) + +#define RPC_DRCR 0x000C // R/W +#define RPC_DRCR_SSLN BIT(24) +#define RPC_DRCR_RBURST(v) ((((v) - 1) & 0x1F) << 16) +#define RPC_DRCR_RCF BIT(9) +#define RPC_DRCR_RBE BIT(8) +#define RPC_DRCR_SSLE BIT(0) + +#define RPC_DRCMR 0x0010 // R/W +#define RPC_DRCMR_CMD(c) (((c) & 0xFF) << 16) +#define RPC_DRCMR_OCMD(c) (((c) & 0xFF) << 0) + +#define RPC_DREAR 0x0014 // R/W +#define RPC_DREAR_EAC(c) (((c) & 0x7) << 0) + +#define RPC_DROPR 0x0018 // R/W + +#define RPC_DRENR 0x001C // R/W +#define RPC_DRENR_CDB(o) (u32)((((o) & 0x3) << 30)) +#define RPC_DRENR_OCDB(o) (((o) & 0x3) << 28) +#define RPC_DRENR_ADB(o) (((o) & 0x3) << 24) +#define RPC_DRENR_OPDB(o) (((o) & 0x3) << 20) +#define RPC_DRENR_DRDB(o) (((o) & 0x3) << 16) +#define RPC_DRENR_DME BIT(15) +#define RPC_DRENR_CDE BIT(14) +#define RPC_DRENR_OCDE BIT(12) +#define RPC_DRENR_ADE(v) (((v) & 0xF) << 8) +#define RPC_DRENR_OPDE(v) (((v) & 0xF) << 4) + +#define RPC_SMCR 0x0020 // R/W +#define RPC_SMCR_SSLKP BIT(8) +#define RPC_SMCR_SPIRE BIT(2) +#define RPC_SMCR_SPIWE BIT(1) +#define RPC_SMCR_SPIE BIT(0) + +#define RPC_SMCMR 0x0024 // R/W +#define RPC_SMCMR_CMD(c) (((c) & 0xFF) << 16) +#define RPC_SMCMR_OCMD(c) (((c) & 0xFF) << 0) + +#define RPC_SMADR 0x0028 // R/W +#define RPC_SMOPR 0x002C // R/W +#define RPC_SMOPR_OPD3(o) (((o) & 0xFF) << 24) +#define RPC_SMOPR_OPD2(o) (((o) & 0xFF) << 16) +#define RPC_SMOPR_OPD1(o) (((o) & 0xFF) << 8) +#define RPC_SMOPR_OPD0(o) (((o) & 0xFF) << 0) + +#define RPC_SMENR 0x0030 // R/W +#define RPC_SMENR_CDB(o) (((o) & 0x3) << 30) +#define RPC_SMENR_OCDB(o) (((o) & 0x3) << 28) +#define RPC_SMENR_ADB(o) (((o) & 0x3) << 24) +#define RPC_SMENR_OPDB(o) (((o) & 0x3) << 20) +#define RPC_SMENR_SPIDB(o) (((o) & 0x3) << 16) +#define RPC_SMENR_DME BIT(15) +#define RPC_SMENR_CDE BIT(14) +#define RPC_SMENR_OCDE BIT(12) +#define RPC_SMENR_ADE(v) (((v) & 0xF) << 8) +#define RPC_SMENR_OPDE(v) (((v) & 0xF) << 4) +#define RPC_SMENR_SPIDE(v) (((v) & 0xF) << 0) + +#define RPC_SMRDR0 0x0038 // R +#define RPC_SMRDR1 0x003C // R +#define RPC_SMWDR0 0x0040 // W +#define RPC_SMWDR1 0x0044 // W + +#define RPC_CMNSR 0x0048 // R +#define RPC_CMNSR_SSLF BIT(1) +#define RPC_CMNSR_TEND BIT(0) + +#define RPC_DRDMCR 0x0058 // R/W +#define RPC_DRDRENR 0x005C // R/W + +#define RPC_SMDMCR 0x0060 // R/W +#define RPC_SMDMCR_DMCYC(v) ((((v) - 1) & 0x1F) << 0) + +#define RPC_SMDRENR 0x0064 // R/W +#define RPC_SMDRENR_HYPE (0x7 << 12) +#define RPC_SMDRENR_ADDRE BIT(8) +#define RPC_SMDRENR_OPDRE BIT(4) +#define RPC_SMDRENR_SPIDRE BIT(0) + +#define RPC_PHYCNT 0x007C // R/W +#define RPC_PHYCNT_CAL BIT(31) +#define PRC_PHYCNT_OCTA_AA BIT(22) +#define PRC_PHYCNT_OCTA_SA BIT(23) +#define PRC_PHYCNT_EXDS BIT(21) +#define RPC_PHYCNT_OCT BIT(20) +#define RPC_PHYCNT_STRTIM(v) (((v) & 0x7) << 15) +#define RPC_PHYCNT_WBUF2 BIT(4) +#define RPC_PHYCNT_WBUF BIT(2) +#define RPC_PHYCNT_PHYMEM(v) (((v) & 0x3) << 0) + +#define RPC_PHYOFFSET1 0x0080 // R/W +#define RPC_PHYOFFSET1_DDRTMG(v) (((v) & 0x3) << 28) +#define RPC_PHYOFFSET2 0x0084 // R/W +#define RPC_PHYOFFSET2_OCTTMG(v) (((v) & 0x7) << 8) + +#define RPC_WBUF_SIZE 256 // Write Buffer size + +struct rpc_spi { + struct clk *clk_rpc; + void __iomem *base; + void __iomem *dirmap; + void __iomem *wbuf; + struct regmap *regmap; + u32 cur_speed_hz; + u32 cmd; + u32 addr; + u32 dummy; + u32 smcr; + u32 smenr; + u32 xferlen; + u32 totalxferlen; + enum spi_mem_data_dir xfer_dir; + struct reset_control *rstc; +}; + +static int rpc_spi_set_freq(struct rpc_spi *rpc, unsigned long freq) +{ + int ret; + + if (rpc->cur_speed_hz == freq) + return 0; + + ret = clk_set_rate(rpc->clk_rpc, freq); + if (ret) + return ret; + + rpc->cur_speed_hz = freq; + return ret; +} + +static void rpc_spi_hw_init(struct rpc_spi *rpc) +{ + // + // NOTE: The 0x260 are undocumented bits, but they must be set. + // RPC_PHYCNT_STRTIM is strobe timing adjustment bit, + // 0x0 : the delay is biggest, + // 0x1 : the delay is 2nd biggest, + // On H3 ES1.x, the value should be 0, while on others, + // the value should be 6. + // + regmap_write(rpc->regmap, RPC_PHYCNT, RPC_PHYCNT_CAL | + RPC_PHYCNT_STRTIM(6) | 0x260); + + // + // NOTE: The 0x1511144 are undocumented bits, but they must be set + // for RPC_PHYOFFSET1. + // The 0x31 are undocumented bits, but they must be set + // for RPC_PHYOFFSET2. + // + regmap_write(rpc->regmap, RPC_PHYOFFSET1, RPC_PHYOFFSET1_DDRTMG(3) | + 0x1511144); + regmap_write(rpc->regmap, RPC_PHYOFFSET2, 0x31 | + RPC_PHYOFFSET2_OCTTMG(4)); + regmap_write(rpc->regmap, RPC_SSLDR, RPC_SSLDR_SPNDL(7) | + RPC_SSLDR_SLNDL(7) | RPC_SSLDR_SCKDL(7)); + regmap_write(rpc->regmap, RPC_CMNCR, RPC_CMNCR_MD | RPC_CMNCR_SFDE | + RPC_CMNCR_MOIIO_HIZ | RPC_CMNCR_IOFV_HIZ | + RPC_CMNCR_BSZ(0)); +} + +static int wait_msg_xfer_end(struct rpc_spi *rpc) +{ + u32 sts; + + return regmap_read_poll_timeout(rpc->regmap, RPC_CMNSR, sts, + sts & RPC_CMNSR_TEND, 0, USEC_PER_SEC); +} + +static u8 rpc_bits_set(u32 nbytes) +{ + nbytes = clamp(nbytes, 1U, 4U); + + return GENMASK(3, 4 - nbytes); +} + +static int rpc_spi_io_xfer(struct rpc_spi *rpc, + const void *tx_buf, void *rx_buf) +{ + u32 smenr, smcr, data, pos = 0; + int ret; + + regmap_update_bits(rpc->regmap, RPC_CMNCR, RPC_CMNCR_MD, RPC_CMNCR_MD); + regmap_write(rpc->regmap, RPC_SMDRENR, 0); + regmap_write(rpc->regmap, RPC_SMCMR, rpc->cmd); + regmap_write(rpc->regmap, RPC_SMDMCR, rpc->dummy); + regmap_write(rpc->regmap, RPC_SMADR, rpc->addr); + smenr = rpc->smenr; + + if (tx_buf) { + while (pos < rpc->xferlen) { + u32 nbytes = rpc->xferlen - pos; + + regmap_write(rpc->regmap, RPC_SMWDR0, + get_unaligned((u32 *)(tx_buf + pos))); + + smcr = rpc->smcr | RPC_SMCR_SPIE; + + if (nbytes > 4) { + nbytes = 4; + smcr |= RPC_SMCR_SSLKP; + } + + regmap_write(rpc->regmap, RPC_SMENR, smenr); + regmap_write(rpc->regmap, RPC_SMCR, smcr); + ret = wait_msg_xfer_end(rpc); + if (ret) + goto err_out; + + pos += nbytes; + smenr = rpc->smenr & ~RPC_SMENR_CDE & + ~RPC_SMENR_ADE(0xf); + } + } else if (rx_buf) { + // + // RPC-IF spoils the data for the commands without an address + // phase (like RDID) in the manual mode, so we'll have to work + // around this issue by using the external address space read + // mode instead. + // + if (!(smenr & RPC_SMENR_ADE(0xf))) { + regmap_update_bits(rpc->regmap, RPC_CMNCR, + RPC_CMNCR_MD, 0); + regmap_write(rpc->regmap, RPC_DRCR, + RPC_DRCR_RBURST(32) | RPC_DRCR_RBE); + regmap_write(rpc->regmap, RPC_DREAR, RPC_DREAR_EAC(1)); + regmap_write(rpc->regmap, RPC_DRCMR, rpc->cmd); + regmap_write(rpc->regmap, RPC_DRDMCR, rpc->dummy); + regmap_write(rpc->regmap, RPC_DROPR, 0); + regmap_write(rpc->regmap, RPC_DRENR, smenr); + memcpy_fromio(rx_buf, rpc->dirmap, rpc->xferlen); + regmap_write(rpc->regmap, RPC_DRCR, RPC_DRCR_RCF); + } else { + while (pos < rpc->xferlen) { + u32 nbytes = rpc->xferlen - pos; + + if (nbytes > 4) + nbytes = 4; + + regmap_write(rpc->regmap, RPC_SMENR, smenr); + regmap_write(rpc->regmap, RPC_SMCR, rpc->smcr | + RPC_SMCR_SPIE); + ret = wait_msg_xfer_end(rpc); + if (ret) + goto err_out; + + regmap_read(rpc->regmap, RPC_SMRDR0, &data); + memcpy(rx_buf + pos, &data, nbytes); + pos += nbytes; + + regmap_write(rpc->regmap, RPC_SMADR, + rpc->addr + pos); + } + } + } else { + regmap_write(rpc->regmap, RPC_SMENR, rpc->smenr); + regmap_write(rpc->regmap, RPC_SMCR, rpc->smcr | RPC_SMCR_SPIE); + ret = wait_msg_xfer_end(rpc); + if (ret) + goto err_out; + } + + return 0; + +err_out: + return reset_control_reset(rpc->rstc); +} + +static void rpc_spi_mem_set_prep_op_cfg(struct spi_device *spi, + const struct spi_mem_op *op, + u64 *offs, size_t *len) +{ + struct rpc_spi *rpc = spi_controller_get_devdata(spi->controller); + + rpc->cmd = RPC_SMCMR_CMD(op->cmd.opcode); + rpc->smenr = RPC_SMENR_CDE | + RPC_SMENR_CDB(ilog2(op->cmd.buswidth)); + rpc->totalxferlen = 1; + rpc->xfer_dir = SPI_MEM_NO_DATA; + rpc->xferlen = 0; + rpc->addr = 0; + + if (op->addr.nbytes) { + rpc->smenr |= RPC_SMENR_ADB(ilog2(op->addr.buswidth)); + if (op->addr.nbytes == 4) + rpc->smenr |= RPC_SMENR_ADE(0xf); + else + rpc->smenr |= RPC_SMENR_ADE(0x7); + + if (offs && len) + rpc->addr = *offs; + else + rpc->addr = op->addr.val; + rpc->totalxferlen += op->addr.nbytes; + } + + if (op->dummy.nbytes) { + rpc->smenr |= RPC_SMENR_DME; + rpc->dummy = RPC_SMDMCR_DMCYC(op->dummy.nbytes); + rpc->totalxferlen += op->dummy.nbytes; + } + + if (op->data.nbytes || (offs && len)) { + switch (op->data.dir) { + case SPI_MEM_DATA_IN: + rpc->smcr = RPC_SMCR_SPIRE; + rpc->xfer_dir = SPI_MEM_DATA_IN; + break; + case SPI_MEM_DATA_OUT: + rpc->smcr = RPC_SMCR_SPIWE; + rpc->xfer_dir = SPI_MEM_DATA_OUT; + break; + default: + break; + } + + if (offs && len) { + rpc->smenr |= RPC_SMENR_SPIDE(rpc_bits_set(*len)) | + RPC_SMENR_SPIDB(ilog2(op->data.buswidth)); + rpc->xferlen = *len; + rpc->totalxferlen += *len; + } else { + rpc->smenr |= + RPC_SMENR_SPIDE(rpc_bits_set(op->data.nbytes)) | + RPC_SMENR_SPIDB(ilog2(op->data.buswidth)); + rpc->xferlen = op->data.nbytes; + rpc->totalxferlen += op->data.nbytes; + } + } +} + +static bool rpc_spi_mem_supports_op(struct spi_mem *mem, + const struct spi_mem_op *op) +{ + if (op->data.buswidth > 4 || op->addr.buswidth > 4 || + op->dummy.buswidth > 4 || op->cmd.buswidth > 4 || + op->addr.nbytes > 4) + return false; + + return true; +} + +static ssize_t rpc_spi_mem_dirmap_read(struct spi_mem_dirmap_desc *desc, + u64 offs, size_t len, void *buf) +{ + struct rpc_spi *rpc = + spi_controller_get_devdata(desc->mem->spi->controller); + int ret; + + if (offs + desc->info.offset + len > U32_MAX) + return -EINVAL; + + if (len > 0x4000000) + len = 0x4000000; + + ret = rpc_spi_set_freq(rpc, desc->mem->spi->max_speed_hz); + if (ret) + return ret; + + rpc_spi_mem_set_prep_op_cfg(desc->mem->spi, + &desc->info.op_tmpl, &offs, &len); + + regmap_update_bits(rpc->regmap, RPC_CMNCR, RPC_CMNCR_MD, 0); + regmap_write(rpc->regmap, RPC_DRCR, RPC_DRCR_RBURST(32) | + RPC_DRCR_RBE); + + regmap_write(rpc->regmap, RPC_DRCMR, rpc->cmd); + regmap_write(rpc->regmap, RPC_DREAR, RPC_DREAR_EAC(1)); + regmap_write(rpc->regmap, RPC_DROPR, 0); + regmap_write(rpc->regmap, RPC_DRENR, rpc->smenr); + regmap_write(rpc->regmap, RPC_DRDMCR, rpc->dummy); + regmap_write(rpc->regmap, RPC_DRDRENR, 0); + + memcpy_fromio(buf, rpc->dirmap + desc->info.offset + offs, len); + + return len; +} + +static ssize_t rpc_spi_mem_dirmap_write(struct spi_mem_dirmap_desc *desc, + u64 offs, size_t len, const void *buf) +{ + struct rpc_spi *rpc = + spi_controller_get_devdata(desc->mem->spi->controller); + int ret; + + if (offs + desc->info.offset + len > U32_MAX) + return -EINVAL; + + if (len > RPC_WBUF_SIZE) + len = RPC_WBUF_SIZE; + + ret = rpc_spi_set_freq(rpc, desc->mem->spi->max_speed_hz); + if (ret) + return ret; + + rpc_spi_mem_set_prep_op_cfg(desc->mem->spi, + &desc->info.op_tmpl, &offs, &len); + + regmap_update_bits(rpc->regmap, RPC_CMNCR, RPC_CMNCR_MD, RPC_CMNCR_MD); + + regmap_write(rpc->regmap, RPC_SMDRENR, 0); + + regmap_update_bits(rpc->regmap, RPC_PHYCNT, RPC_PHYCNT_STRTIM(7) | + RPC_PHYCNT_WBUF2 | RPC_PHYCNT_WBUF, + RPC_PHYCNT_WBUF2 | RPC_PHYCNT_WBUF); + + memcpy_toio(rpc->wbuf, buf, len); + + regmap_write(rpc->regmap, RPC_SMCMR, rpc->cmd); + regmap_write(rpc->regmap, RPC_SMADR, offs); + regmap_write(rpc->regmap, RPC_SMENR, rpc->smenr); + regmap_write(rpc->regmap, RPC_SMCR, rpc->smcr | RPC_SMCR_SPIE); + ret = wait_msg_xfer_end(rpc); + if (ret) + goto err_out; + + regmap_write(rpc->regmap, RPC_DRCR, RPC_DRCR_RCF); + + regmap_update_bits(rpc->regmap, RPC_PHYCNT, RPC_PHYCNT_STRTIM(7) | + RPC_PHYCNT_WBUF2 | RPC_PHYCNT_WBUF, + RPC_PHYCNT_STRTIM(6)); + + return len; + +err_out: + return reset_control_reset(rpc->rstc); +} + +static int rpc_spi_mem_dirmap_create(struct spi_mem_dirmap_desc *desc) +{ + struct rpc_spi *rpc = + spi_controller_get_devdata(desc->mem->spi->controller); + + if (desc->info.offset + desc->info.length > U32_MAX) + return -ENOTSUPP; + + if (!rpc_spi_mem_supports_op(desc->mem, &desc->info.op_tmpl)) + return -ENOTSUPP; + + if (!rpc->dirmap && + desc->info.op_tmpl.data.dir == SPI_MEM_DATA_IN) + return -ENOTSUPP; + + if (!rpc->wbuf && + desc->info.op_tmpl.data.dir == SPI_MEM_DATA_OUT) + return -ENOTSUPP; + + return 0; +} + +static int rpc_spi_mem_exec_op(struct spi_mem *mem, + const struct spi_mem_op *op) +{ + struct rpc_spi *rpc = spi_controller_get_devdata(mem->spi->controller); + int ret; + + ret = rpc_spi_set_freq(rpc, mem->spi->max_speed_hz); + if (ret) + return ret; + + rpc_spi_mem_set_prep_op_cfg(mem->spi, op, NULL, NULL); + + ret = rpc_spi_io_xfer(rpc, + op->data.dir == SPI_MEM_DATA_OUT ? + op->data.buf.out : NULL, + op->data.dir == SPI_MEM_DATA_IN ? + op->data.buf.in : NULL); + + return ret; +} + +static const struct spi_controller_mem_ops rpc_spi_mem_ops = { + .supports_op = rpc_spi_mem_supports_op, + .exec_op = rpc_spi_mem_exec_op, + .dirmap_create = rpc_spi_mem_dirmap_create, + .dirmap_read = rpc_spi_mem_dirmap_read, + .dirmap_write = rpc_spi_mem_dirmap_write, +}; + +static void rpc_spi_transfer_setup(struct rpc_spi *rpc, + struct spi_message *msg) +{ + struct spi_transfer *t, xfer[4] = { }; + u32 i, xfercnt, xferpos = 0; + + rpc->totalxferlen = 0; + rpc->xfer_dir = SPI_MEM_NO_DATA; + + list_for_each_entry(t, &msg->transfers, transfer_list) { + if (t->tx_buf) { + xfer[xferpos].tx_buf = t->tx_buf; + xfer[xferpos].tx_nbits = t->tx_nbits; + } + + if (t->rx_buf) { + xfer[xferpos].rx_buf = t->rx_buf; + xfer[xferpos].rx_nbits = t->rx_nbits; + } + + if (t->len) { + xfer[xferpos++].len = t->len; + rpc->totalxferlen += t->len; + } + + if (list_is_last(&t->transfer_list, &msg->transfers)) { + if (xferpos > 1) { + if (t->rx_buf) { + rpc->xfer_dir = SPI_MEM_DATA_IN; + rpc->smcr = RPC_SMCR_SPIRE; + } else if (t->tx_buf) { + rpc->xfer_dir = SPI_MEM_DATA_OUT; + rpc->smcr = RPC_SMCR_SPIWE; + } + } + } + } + + xfercnt = xferpos; + rpc->xferlen = xfer[--xferpos].len; + rpc->cmd = RPC_SMCMR_CMD(((u8 *)xfer[0].tx_buf)[0]); + rpc->smenr = RPC_SMENR_CDE | + RPC_SMENR_CDB(ilog2((unsigned int)xfer[0].tx_nbits)); + rpc->addr = 0; + + if (xfercnt > 2 && xfer[1].len && xfer[1].tx_buf) { + rpc->smenr |= + RPC_SMENR_ADB(ilog2((unsigned int)xfer[1].tx_nbits)); + + for (i = 0; i < xfer[1].len; i++) + rpc->addr |= ((u8 *)xfer[1].tx_buf)[i] << + (8 * (xfer[1].len - i - 1)); + + if (xfer[1].len == 4) + rpc->smenr |= RPC_SMENR_ADE(0xf); + else + rpc->smenr |= RPC_SMENR_ADE(0x7); + } + + if (xfercnt > 3 && xfer[2].len && xfer[2].tx_buf) { + rpc->smenr |= RPC_SMENR_DME; + rpc->dummy = RPC_SMDMCR_DMCYC(xfer[2].len); + } + + for (i = xfercnt - 1; i < xfercnt && xfercnt > 1; i++) { + if (xfer[i].rx_buf) { + rpc->smenr |= + RPC_SMENR_SPIDE(rpc_bits_set(xfer[i].len)) | + RPC_SMENR_SPIDB(ilog2 + ((unsigned int)xfer[i].rx_nbits)); + } else if (xfer[i].tx_buf) { + rpc->smenr |= + RPC_SMENR_SPIDE(rpc_bits_set(xfer[i].len)) | + RPC_SMENR_SPIDB(ilog2 + ((unsigned int)xfer[i].tx_nbits)); + } + } +} + +static inline int rpc_spi_xfer_message(struct rpc_spi *rpc, + struct spi_transfer *data_xfer) +{ + int ret; + + ret = rpc_spi_set_freq(rpc, data_xfer->speed_hz); + if (ret) + return ret; + + ret = rpc_spi_io_xfer(rpc, + rpc->xfer_dir == SPI_MEM_DATA_OUT ? + data_xfer->tx_buf : NULL, + rpc->xfer_dir == SPI_MEM_DATA_IN ? + data_xfer->rx_buf : NULL); + + return ret; +} + +static int rpc_spi_transfer_one_message(struct spi_controller *ctlr, + struct spi_message *msg) +{ + struct rpc_spi *rpc = spi_controller_get_devdata(ctlr); + struct spi_transfer *data_xfer; + int ret; + + rpc_spi_transfer_setup(rpc, msg); + + data_xfer = list_last_entry(&msg->transfers, struct spi_transfer, + transfer_list); + + ret = rpc_spi_xfer_message(rpc, data_xfer); + if (ret) + goto out; + + msg->status = 0; + msg->actual_length = rpc->totalxferlen; +out: + spi_finalize_current_message(ctlr); + return 0; +} + +static const struct regmap_range rpc_spi_volatile_ranges[] = { + regmap_reg_range(RPC_SMRDR0, RPC_SMRDR0), + regmap_reg_range(RPC_SMWDR0, RPC_SMWDR0), + regmap_reg_range(RPC_CMNSR, RPC_CMNSR), +}; + +static const struct regmap_access_table rpc_spi_volatile_table = { + .yes_ranges = rpc_spi_volatile_ranges, + .n_yes_ranges = ARRAY_SIZE(rpc_spi_volatile_ranges), +}; + +static const struct regmap_config rpc_spi_regmap_config = { + .reg_bits = 32, + .val_bits = 32, + .reg_stride = 4, + .fast_io = true, + .max_register = RPC_PHYOFFSET2, + .volatile_table = &rpc_spi_volatile_table, +}; + +static int rpc_spi_probe(struct platform_device *pdev) +{ + struct spi_controller *ctlr; + struct resource *res; + struct rpc_spi *rpc; + struct device_node *rpc_if; + int ret; + + rpc_if = of_get_next_child(pdev->dev.of_node, NULL); + if (!rpc_if) { + dev_warn(&pdev->dev, "no spi-nor device node found\n"); + return -ENODEV; + } + + ret = of_device_is_compatible(rpc_if, "jedec,spi-nor"); + if (!ret) { + dev_warn(&pdev->dev, "no compatible spi-nor device found\n"); + return -ENODEV; + } + + ctlr = spi_alloc_master(&pdev->dev, sizeof(*rpc)); + if (!ctlr) + return -ENOMEM; + + platform_set_drvdata(pdev, ctlr); + + rpc = spi_controller_get_devdata(ctlr); + + ctlr->dev.of_node = pdev->dev.of_node; + + rpc->clk_rpc = devm_clk_get(&pdev->dev, "rpc"); + if (IS_ERR(rpc->clk_rpc)) + return PTR_ERR(rpc->clk_rpc); + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "regs"); + rpc->base = devm_ioremap_resource(&pdev->dev, res); + if (IS_ERR(rpc->base)) + return PTR_ERR(rpc->base); + + rpc->regmap = devm_regmap_init_mmio(&pdev->dev, rpc->base, + &rpc_spi_regmap_config); + if (IS_ERR(rpc->regmap)) { + dev_err(&pdev->dev, + "failed to init regmap for rpc-spi, error %ld\n", + PTR_ERR(rpc->regmap)); + return PTR_ERR(rpc->regmap); + } + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dirmap"); + rpc->dirmap = devm_ioremap_resource(&pdev->dev, res); + if (IS_ERR(rpc->dirmap)) + rpc->dirmap = NULL; + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "wbuf"); + rpc->wbuf = devm_ioremap_resource(&pdev->dev, res); + if (IS_ERR(rpc->wbuf)) + rpc->wbuf = NULL; + + rpc->rstc = devm_reset_control_get_exclusive(&pdev->dev, NULL); + if (IS_ERR(rpc->rstc)) + return PTR_ERR(rpc->rstc); + + pm_runtime_enable(&pdev->dev); + ctlr->auto_runtime_pm = true; + + ctlr->num_chipselect = 1; + ctlr->mem_ops = &rpc_spi_mem_ops; + ctlr->transfer_one_message = rpc_spi_transfer_one_message; + + ctlr->bits_per_word_mask = SPI_BPW_MASK(8); + ctlr->mode_bits = SPI_CPOL | SPI_CPHA | SPI_TX_QUAD | SPI_RX_QUAD; + + rpc_spi_hw_init(rpc); + + ret = spi_register_controller(ctlr); + if (ret) { + dev_err(&pdev->dev, "spi_register_controller failed\n"); + goto err_put_ctlr; + } + return 0; + +err_put_ctlr: + spi_controller_put(ctlr); + pm_runtime_disable(&pdev->dev); + + return ret; +} + +static int rpc_spi_remove(struct platform_device *pdev) +{ + struct spi_controller *ctlr = platform_get_drvdata(pdev); + + pm_runtime_disable(&pdev->dev); + spi_unregister_controller(ctlr); + + return 0; +} + +static const struct of_device_id rpc_spi_of_ids[] = { + { .compatible = "renesas,r8a77995-rpc", }, + { .compatible = "renesas,rcar-gen3-rpc", }, + { /* sentinel */ } +}; +MODULE_DEVICE_TABLE(of, rpc_spi_of_ids); + +#ifdef CONFIG_PM_SLEEP +static int rpc_spi_suspend(struct device *dev) +{ + struct spi_controller *ctlr = dev_get_drvdata(dev); + + return spi_controller_suspend(ctlr); +} + +static int rpc_spi_resume(struct device *dev) +{ + struct spi_controller *ctlr = dev_get_drvdata(dev); + + return spi_controller_resume(ctlr); +} + +static SIMPLE_DEV_PM_OPS(rpc_spi_pm_ops, rpc_spi_suspend, rpc_spi_resume); +#define DEV_PM_OPS (&rpc_spi_pm_ops) +#else +#define DEV_PM_OPS NULL +#endif + +static struct platform_driver rpc_spi_driver = { + .probe = rpc_spi_probe, + .remove = rpc_spi_remove, + .driver = { + .name = "rpc-spi", + .of_match_table = rpc_spi_of_ids, + .pm = DEV_PM_OPS, + }, +}; +module_platform_driver(rpc_spi_driver); + +MODULE_AUTHOR("Mason Yang <masonccyang@mxic.com.tw>"); +MODULE_DESCRIPTION("Renesas R-Car Gen3 RPC-IF SPI controller driver"); +MODULE_LICENSE("GPL v2");