[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <3d185e02-3484-5226-76bf-e3ee98af8ed9@amd.com>
Date: Thu, 16 Feb 2023 00:04:39 -0600
From: Mario Limonciello <mario.limonciello@....com>
To: Mika Westerberg <mika.westerberg@...ux.intel.com>
Cc: Andreas Noever <andreas.noever@...il.com>,
Michael Jamet <michael.jamet@...el.com>,
Yehezkel Bernat <YehezkelShB@...il.com>, Sanju.Mehta@....com,
stable@...r.kernel.org, linux-usb@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/2] thunderbolt: Adjust how NVM reading works
On 2/15/23 23:49, Mika Westerberg wrote:
> Hi Mario,
>
> On Wed, Feb 15, 2023 at 11:25:19AM -0600, Mario Limonciello wrote:
>> Some TBT3 devices have a hard time reliably responding to bit banging
>> requests correctly when connected to AMD USB4 hosts running Linux.
>>
>> These problems are not reported in any other CM supported on AMD platforms,
>> and comparing the Windows and Pre-OS implementations the Linux CM is the
>> only one that utilizes bit banging to access the DROM.
>> Other CM implementations access the DROM directly from the NVM instead of
>> bit banging.
>>
>> Adjust the flow to use this method to fetch the NVM when the downstream
>> device is Thunderbolt 3 and only use bit banging to access TBT 2 or TBT 1
>> devices. As the flow is modified, also remove the retry sequence that was
>> introduced from commit f022ff7bf377 ("thunderbolt: Retry DROM read once
>> if parsing fails") as it will not be necessary if the NVM is fetched this
>> way.
>>
>> Cc: stable@...r.kernel.org
>> Fixes: f022ff7bf377 ("thunderbolt: Retry DROM read once if parsing fails")
> I don't think it fixes a regression of above commit and I don't think
> this is stable material because it is quite a big change. I would rather
> let it sit in -rcX for a while to make sure no user visible changes are
> accidentally introduced. Is this OK for you?
No worry on dropping these two tags. I'll do that for a v3.
The original change in v1 was more stable material than this is now.
While adopting your feedback below I'll keep in mind patch ordering
for refactoring as I do want to get the core of the solution back to stable
eventually when we're collectively confident on it.
> Did you check that the UUID of these (and other possible) devices stay
> the same before and after the patch?
Yeah I did across a few hotplug cycles and suspend cycles.
>
>> Signed-off-by: Mario Limonciello <mario.limonciello@....com>
>> ---
>> v1->v2:
>> * Update commit message to indicate which CMs are tested
>> * Adjust flow to only fetch DROM from NVM on TBT3 and bit bang on TBT1/2
>> ---
>> drivers/thunderbolt/eeprom.c | 145 +++++++++++++++++++----------------
>> 1 file changed, 80 insertions(+), 65 deletions(-)
>>
>> diff --git a/drivers/thunderbolt/eeprom.c b/drivers/thunderbolt/eeprom.c
>> index c90d22f56d4e..d1be72b6afdb 100644
>> --- a/drivers/thunderbolt/eeprom.c
>> +++ b/drivers/thunderbolt/eeprom.c
>> @@ -416,7 +416,7 @@ static int tb_drom_parse_entries(struct tb_switch *sw, size_t header_size)
>> if (pos + 1 == drom_size || pos + entry->len > drom_size
>> || !entry->len) {
>> tb_sw_warn(sw, "DROM buffer overrun\n");
>> - return -EILSEQ;
>> + return -EIO;
>> }
>>
>> switch (entry->type) {
>> @@ -544,7 +544,37 @@ static int tb_drom_read_n(struct tb_switch *sw, u16 offset, u8 *val,
>> return tb_eeprom_read_n(sw, offset, val, count);
>> }
>>
>> -static int tb_drom_parse(struct tb_switch *sw)
>> +static int tb_drom_bit_bang(struct tb_switch *sw, u16 *size)
>> +{
>> + int res;
>> +
>> + res = tb_drom_read_n(sw, 14, (u8 *) size, 2);
>> + if (res)
>> + return res;
>> + *size &= 0x3ff;
>> + *size += TB_DROM_DATA_START;
>> + tb_sw_dbg(sw, "reading drom (length: %#x)\n", *size);
>> + if (*size < sizeof(struct tb_drom_header)) {
>> + tb_sw_warn(sw, "drom too small, aborting\n");
>> + return -EIO;
>> + }
>> +
>> + sw->drom = kzalloc(*size, GFP_KERNEL);
>> + if (!sw->drom)
>> + return -ENOMEM;
>> +
>> + res = tb_drom_read_n(sw, 0, sw->drom, *size);
>> + if (res)
>> + goto err;
>> +
>> + return 0;
>> +err:
>> + kfree(sw->drom);
>> + sw->drom = NULL;
>> + return res;
>> +}
> Can you split the refactoring part into a separate patch?
Sure.
>> +
>> +static int tb_drom_parse_v1(struct tb_switch *sw)
>> {
>> const struct tb_drom_header *header =
>> (const struct tb_drom_header *)sw->drom;
>> @@ -555,7 +585,7 @@ static int tb_drom_parse(struct tb_switch *sw)
>> tb_sw_warn(sw,
>> "DROM UID CRC8 mismatch (expected: %#x, got: %#x)\n",
>> header->uid_crc8, crc);
>> - return -EILSEQ;
>> + return -EIO;
>> }
>> if (!sw->uid)
>> sw->uid = header->uid;
>> @@ -589,6 +619,43 @@ static int usb4_drom_parse(struct tb_switch *sw)
>> return tb_drom_parse_entries(sw, USB4_DROM_HEADER_SIZE);
>> }
>>
>> +static int tb_drom_parse(struct tb_switch *sw, u16 *size)
>> +{
>> + struct tb_drom_header *header = (void *) sw->drom;
>> + int res;
>> +
>> + if (header->data_len + TB_DROM_DATA_START != *size) {
>> + tb_sw_warn(sw, "drom size mismatch\n");
>> + goto err;
>> + }
>> +
>> + tb_sw_dbg(sw, "DROM version: %d\n", header->device_rom_revision);
>> +
>> + switch (header->device_rom_revision) {
>> + case 3:
>> + res = usb4_drom_parse(sw);
>> + break;
>> + default:
>> + tb_sw_warn(sw, "DROM device_rom_revision %#x unknown\n",
>> + header->device_rom_revision);
>> + fallthrough;
>> + case 1:
>> + res = tb_drom_parse_v1(sw);
>> + break;
>> + }
>> +
>> + if (res) {
>> + tb_sw_warn(sw, "parsing DROM failed\n");
>> + goto err;
>> + }
>> +
>> + return 0;
>> +err:
>> + kfree(sw->drom);
>> + sw->drom = NULL;
>> + return -EIO;
>> +}
>> +
>> /**
>> * tb_drom_read() - Copy DROM to sw->drom and parse it
>> * @sw: Router whose DROM to read and parse
>> @@ -602,8 +669,7 @@ static int usb4_drom_parse(struct tb_switch *sw)
>> int tb_drom_read(struct tb_switch *sw)
>> {
>> u16 size;
>> - struct tb_drom_header *header;
>> - int res, retries = 1;
>> + int res;
>>
>> if (sw->drom)
>> return 0;
>> @@ -614,11 +680,11 @@ int tb_drom_read(struct tb_switch *sw)
>> * in a device property. Use it if available.
>> */
>> if (tb_drom_copy_efi(sw, &size) == 0)
>> - goto parse;
>> + return tb_drom_parse(sw, &size);
>>
>> /* Non-Apple hardware has the DROM as part of NVM */
>> if (tb_drom_copy_nvm(sw, &size) == 0)
>> - goto parse;
>> + return tb_drom_parse(sw, &size);
>>
>> /*
>> * USB4 hosts may support reading DROM through router
>> @@ -627,7 +693,7 @@ int tb_drom_read(struct tb_switch *sw)
>> if (tb_switch_is_usb4(sw)) {
>> usb4_switch_read_uid(sw, &sw->uid);
>> if (!usb4_copy_host_drom(sw, &size))
>> - goto parse;
>> + return tb_drom_parse(sw, &size);
>> } else {
>> /*
>> * The root switch contains only a dummy drom
>> @@ -640,64 +706,13 @@ int tb_drom_read(struct tb_switch *sw)
>> return 0;
>> }
>>
>> - res = tb_drom_read_n(sw, 14, (u8 *) &size, 2);
>> + /* TBT3 devices have the DROM as part of NVM */
>> + if (sw->generation < 3)
> This is true for TBT2 devices too. I think you want to check for the
> sw->cap_lc here instead. If it is set the device has LC and therefore we
> can use the LC UUID registers to figure out the UUID in later stages.
> Otherwise we need to read it through bitbanging.
OK thanks, will adjust.
>> + res = tb_drom_bit_bang(sw, &size);
>> + else
>> + res = tb_drom_copy_nvm(sw, &size);
>> if (res)
>> return res;
>> - size &= 0x3ff;
>> - size += TB_DROM_DATA_START;
>> - tb_sw_dbg(sw, "reading drom (length: %#x)\n", size);
>> - if (size < sizeof(*header)) {
>> - tb_sw_warn(sw, "drom too small, aborting\n");
>> - return -EIO;
>> - }
>> -
>> - sw->drom = kzalloc(size, GFP_KERNEL);
>> - if (!sw->drom)
>> - return -ENOMEM;
>> -read:
>> - res = tb_drom_read_n(sw, 0, sw->drom, size);
>> - if (res)
>> - goto err;
>> -
>> -parse:
>> - header = (void *) sw->drom;
>> -
>> - if (header->data_len + TB_DROM_DATA_START != size) {
>> - tb_sw_warn(sw, "drom size mismatch\n");
>> - if (retries--) {
>> - msleep(100);
>> - goto read;
>> - }
>> - goto err;
>> - }
>>
>> - tb_sw_dbg(sw, "DROM version: %d\n", header->device_rom_revision);
>> -
>> - switch (header->device_rom_revision) {
>> - case 3:
>> - res = usb4_drom_parse(sw);
>> - break;
>> - default:
>> - tb_sw_warn(sw, "DROM device_rom_revision %#x unknown\n",
>> - header->device_rom_revision);
>> - fallthrough;
>> - case 1:
>> - res = tb_drom_parse(sw);
>> - break;
>> - }
>> -
>> - /* If the DROM parsing fails, wait a moment and retry once */
>> - if (res == -EILSEQ && retries--) {
>> - tb_sw_warn(sw, "parsing DROM failed\n");
>> - msleep(100);
>> - goto read;
>> - }
>> -
>> - if (!res)
>> - return 0;
>> -
>> -err:
>> - kfree(sw->drom);
>> - sw->drom = NULL;
>> - return -EIO;
>> + return tb_drom_parse(sw, &size);
>> }
>> --
>> 2.34.1
Powered by blists - more mailing lists