Re: [PATCH v3 1/2] mtd: spi-nor: add support to non-uniform SFDP SPI NOR flash memories

[Boris Brezillon <boris.brezillon@xxxxxxxxxxx>]

On Tue, 23 Oct 2018 08:59:22 +0000
Yogesh Narayan Gaur <yogeshnarayan.gaur@xxxxxxx> wrote:

> Hi, 
> 
> > -----Original Message-----
> > From: Boris Brezillon [mailto:boris.brezillon@xxxxxxxxxxx]
> > Sent: Tuesday, October 23, 2018 2:18 PM
> > To: Yogesh Narayan Gaur <yogeshnarayan.gaur@xxxxxxx>
> > Cc: Cristian.Birsan@xxxxxxxxxxxxx; Tudor Ambarus
> > <tudor.ambarus@xxxxxxxxxxxxx>; richard@xxxxxx; Mark Brown
> > <broonie@xxxxxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx;
> > nicolas.ferre@xxxxxxxxxxxxx; marek.vasut@xxxxxxxxx;
> > cyrille.pitchen@xxxxxxxxxxxxx; linux-mtd@xxxxxxxxxxxxxxxxxxx; Cyrille Pitchen
> > <cyrille.pitchen@xxxxxxxxxx>; computersforpeace@xxxxxxxxx;
> > dwmw2@xxxxxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> > Subject: Re: [PATCH v3 1/2] mtd: spi-nor: add support to non-uniform SFDP SPI
> > NOR flash memories
> > 
> > On Tue, 23 Oct 2018 08:18:35 +0000
> > Yogesh Narayan Gaur <yogeshnarayan.gaur@xxxxxxx> wrote:
> >   
> > >
> > > I have added the prints in m25p80_read() and in flexspi controller prepare_lut  
> > and read_rxfifo() func.  
> > > In these have added prints for data variable of struct op and data which being  
> > read by the controller from the flash.  
> > >
> > > [    2.091467] smpt[0]=[addr_width:00000003, read_dumy:00000008,  
> > read_opcode:00000065, data_mask:00000008]  
> > > [    2.099113] m25p80_read, nor[op:00000065 addr_width:00000003,  
> > dummy:00000008, len:00000001  
> > > [    2.107367] m25p80_read, cmd[opcode:65 bwidth:1] aadr[val:4, nbytes:3,  
> > bwidth:1]  
> > > [    2.114753] m25p80_read, dummy[nbytes:1 bwidth:1] data[bwidth:1,  
> > nbytes:1]  
> > > [    2.121706] nxp_fspi_prepare_lut cmd[opcode:65 bwidth:1] aadr[val:4,  
> > nbytes:3, bwidth:1]  
> > > [    2.129786] dummy[nbytes:1 bwidth:1] data[dir:0 bwidth:1, nbytes:1]
> > > [    2.136132] nxp-fspi 20c0000.flexspi: CMD[65] lutval[0:8180465  
> > 1:24003008      2:0     3:0]  
> > > [    2.144223] nxp_fspi_read_rxfifo, ReadData op.buf[0x00]
> > > [    2.151004] smpt_read[1] addr[00000004], data_byte[00000000]  
> > err:00000000  
> > >
> > >
> > > [    2.157782] smpt[2]=[addr_width:00000003, read_dumy:00000008,  
> > read_opcode:00000065, data_mask:00000004]  
> > > [    2.165429] m25p80_read, nor[op:00000065 addr_width:00000003,  
> > dummy:00000008, len:00000001  
> > > [    2.173683] m25p80_read, cmd[opcode:65 bwidth:1] aadr[val:2, nbytes:3,  
> > bwidth:1]  
> > > [    2.181068] m25p80_read, dummy[nbytes:1 bwidth:1] data[bwidth:1,  
> > nbytes:1]  
> > > [    2.188021] nxp_fspi_prepare_lut cmd[opcode:65 bwidth:1] aadr[val:2,  
> > nbytes:3, bwidth:1]  
> > > [    2.196101] dummy[nbytes:1 bwidth:1] data[dir:0 bwidth:1, nbytes:1]
> > > [    2.202447] nxp-fspi 20c0000.flexspi: CMD[65] lutval[0:8180465  
> > 1:24003008      2:0     3:0]  
> > > [    2.210539] nxp_fspi_read_rxfifo, ReadData op.buf[0x02]
> > > [    2.217319] smpt_read[3] addr[00000002], data_byte[00000002]  
> > err:00000000  
> > >
> > >
> > > [    2.224098] smpt[4]=[addr_width:00000003, read_dumy:00000008,  
> > read_opcode:00000065, data_mask:00000002]  
> > > [    2.231744] m25p80_read, nor[op:00000065 addr_width:00000003,  
> > dummy:00000008, len:00000001  
> > > [    2.239998] m25p80_read, cmd[opcode:65 bwidth:1] aadr[val:4, nbytes:3,  
> > bwidth:1]  
> > > [    2.247383] m25p80_read, dummy[nbytes:1 bwidth:1] data[bwidth:1,  
> > nbytes:1]  
> > > [    2.254336] nxp_fspi_prepare_lut cmd[opcode:65 bwidth:1] aadr[val:4,  
> > nbytes:3, bwidth:1]  
> > > [    2.262416] dummy[nbytes:1 bwidth:1] data[dir:0 bwidth:1, nbytes:1]
> > > [    2.268762] nxp-fspi 20c0000.flexspi: CMD[65] lutval[0:8180465  
> > 1:24003008      2:0     3:0]  
> > > [    2.276854] nxp_fspi_read_rxfifo, ReadData op.buf[0x00]
> > > [    2.283634] smpt_read[5] addr[00000004], data_byte[00000000]  
> > err:00000000  
> > >
> > >
> > > [    2.290412] spi_nor_get_map_in_use:2915 map_id=0 smpt_len:16 i=:6
> > > [    2.296496] End [addr_width:00000003, read_dumy:00000008,  
> > read_opcode:00000065] ReturnVal:00000000  
> > > [    2.305444] spi_nor_parse_smpt:3065
> > > [    2.308924] m25p80 spi0.0: failed to parse SMPT (err = -22)
> > >
> > >  
> > > >
> > > > Next thing you can do is read the CR2NV reg (using the RDAR command)
> > > > and check the RL (Read Latency) and AL (Address Length) values.  
> > >
> > > Please let me know how to read CR2NV register.  
> > 
> > Actually, RDAR is already what you use to read the map_id, and we need to use
> > it to read the register that contains the number of dummy cycles and the number
> > of address bytes to use for RDAR operations. Looks like we have a chicken and
> > egg situation here :-).
> > 
> > Let's try something else:
> > 
> > 1/ create an u8 array of 16 entries named data_bytes
> > 
> > for each loop iteration (the first for loop):
> > 2/ set ->addr_width to 3 and ->read_dummy to 0 3/ call spi_nor_read_raw(nor,
> > addr, ARRAY_SIZE(data_bytes), data_bytes) 4/ dump the data_bytes buf 5/ set -  
> > >addr_width to 4 6/ call spi_nor_read_raw(nor, addr, ARRAY_SIZE(data_bytes),  
> > data_bytes) 7/ dump the data_bytes buf
> > 
> > If the SPI driver is working correctly, we should be able to figure out the right
> > value for ->addr_width and ->read_dummy.
> >   
> 
> Ok, I would try with the suggested change.
> But till the point we not able to figure out actual failure reason can we have below changes pushed-in

We're not there yet. The change just reached Linus tree, and we have
plenty of time to fix the issue before 4.20 is out. So no, I don't want
to rush this patch in until we figure out what's really going on.

> 
> @@ -3139,6 +3177,17 @@ static int spi_nor_parse_sfdp(struct spi_nor *nor,
>                 switch (SFDP_PARAM_HEADER_ID(param_header)) {
>                 case SFDP_SECTOR_MAP_ID:
>                         err = spi_nor_parse_smpt(nor, param_header);
> +                       if (err) {
> +                               dev_warn(dev,
> +                               "failed to parse SMPT (err = %d)\n",
> +                               err);
> +                               /*
> +                                * SMPT parsing is optional, let's not drop
> +                                * all information we extracted so far just
> +                                * because it failed.
> +                                */
> +                               err = 0;
> +                       }
> 
> --
> Regards
> Yogesh Gaur.
>  
> > Thanks,
> > 
> > Boris  


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/