From: "Pali Rohár" <pali@kernel.org>
To: u-boot@lists.denx.de
Subject: [PATCH] arm: a37xx: pci: Fix processing PIO transfers
Date: Thu, 22 Apr 2021 16:23:04 +0200 [thread overview]
Message-ID: <20210422142304.31065-1-pali@kernel.org> (raw)
Trying to clear PIO_START register when it is non-zero (which indicates
that previous PIO transfer has not finished yet) causes an External
Abort with SError 0xbf000002.
This bug is currently worked around in TF-A by handling External Aborts
in EL3 and ignoring this particular SError.
This workaround was also discussed at:
https://git.trustedfirmware.org/TF-A/trusted-firmware-a.git/commit/?id=3c7dcdac5c50
https://lore.kernel.org/linux-pci/20190316161243.29517-1-repk at triplefau.lt/
https://lore.kernel.org/linux-pci/971be151d24312cc533989a64bd454b4 at www.loen.fr/
https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/1541
Implement a proper fix to prevent this External Abort. As it is not
possible to cancel a pending PIO transfer, simply do not start a new one
if previous has not finished yet. In this case return an error to the
caller.
In most cases this SError happens when there is no PCIe card connected
or when PCIe link is down. The reason is that in these cases a PIO
transfer takes about 1.44 seconds. For this reason we also increase the
wait timeout in pcie_advk_wait_pio() to 1.5 seconds.
If PIO read transfer for PCI_VENDOR_ID register times out, or if it
isn't possible to read it yet because previous transfer is not finished,
return Completion Retry Status value instead of failing, to give the
caller a chance to send a new read request.
Signed-off-by: Pali Roh?r <pali@kernel.org>
Reviewed-by: Marek Beh?n <marek.behun@nic.cz>
---
drivers/pci/pci-aardvark.c | 42 +++++++++++++++++++++++++-------------
1 file changed, 28 insertions(+), 14 deletions(-)
diff --git a/drivers/pci/pci-aardvark.c b/drivers/pci/pci-aardvark.c
index 3b9309f52c..c43d4f309b 100644
--- a/drivers/pci/pci-aardvark.c
+++ b/drivers/pci/pci-aardvark.c
@@ -132,8 +132,9 @@
PCIE_CONF_FUNC(PCI_FUNC(devfn)) | PCIE_CONF_REG(where))
/* PCIe Retries & Timeout definitions */
-#define MAX_RETRIES 10
-#define PIO_WAIT_TIMEOUT 100
+#define PIO_MAX_RETRIES 1500
+#define PIO_WAIT_TIMEOUT 1000
+#define LINK_MAX_RETRIES 10
#define LINK_WAIT_TIMEOUT 100000
#define CFG_RD_UR_VAL 0xFFFFFFFF
@@ -192,7 +193,7 @@ static int pcie_advk_addr_valid(pci_dev_t bdf, int first_busno)
*
* @pcie: The PCI device to access
*
- * Wait up to 1 micro second for PIO access to be accomplished.
+ * Wait up to 1.5 seconds for PIO access to be accomplished.
*
* Return 1 (true) if PIO access is accomplished.
* Return 0 (false) if PIO access is timed out.
@@ -202,7 +203,7 @@ static int pcie_advk_wait_pio(struct pcie_advk *pcie)
uint start, isr;
uint count;
- for (count = 0; count < MAX_RETRIES; count++) {
+ for (count = 0; count < PIO_MAX_RETRIES; count++) {
start = advk_readl(pcie, PIO_START);
isr = advk_readl(pcie, PIO_ISR);
if (!start && isr)
@@ -214,7 +215,7 @@ static int pcie_advk_wait_pio(struct pcie_advk *pcie)
udelay(PIO_WAIT_TIMEOUT);
}
- dev_err(pcie->dev, "config read/write timed out\n");
+ dev_err(pcie->dev, "PIO read/write transfer time out\n");
return 0;
}
@@ -323,9 +324,14 @@ static int pcie_advk_read_config(const struct udevice *bus, pci_dev_t bdf,
return 0;
}
- /* Start PIO */
- advk_writel(pcie, 0, PIO_START);
- advk_writel(pcie, 1, PIO_ISR);
+ if (advk_readl(pcie, PIO_START)) {
+ dev_err(pcie->dev,
+ "Previous PIO read/write transfer is still running\n");
+ if (offset != PCI_VENDOR_ID)
+ return -EINVAL;
+ *valuep = CFG_RD_CRS_VAL;
+ return 0;
+ }
/* Program the control register */
reg = advk_readl(pcie, PIO_CTRL);
@@ -342,10 +348,15 @@ static int pcie_advk_read_config(const struct udevice *bus, pci_dev_t bdf,
advk_writel(pcie, 0, PIO_ADDR_MS);
/* Start the transfer */
+ advk_writel(pcie, 1, PIO_ISR);
advk_writel(pcie, 1, PIO_START);
- if (!pcie_advk_wait_pio(pcie))
- return -EINVAL;
+ if (!pcie_advk_wait_pio(pcie)) {
+ if (offset != PCI_VENDOR_ID)
+ return -EINVAL;
+ *valuep = CFG_RD_CRS_VAL;
+ return 0;
+ }
/* Check PIO status and get the read result */
ret = pcie_advk_check_pio_status(pcie, true, ®);
@@ -420,9 +431,11 @@ static int pcie_advk_write_config(struct udevice *bus, pci_dev_t bdf,
return 0;
}
- /* Start PIO */
- advk_writel(pcie, 0, PIO_START);
- advk_writel(pcie, 1, PIO_ISR);
+ if (advk_readl(pcie, PIO_START)) {
+ dev_err(pcie->dev,
+ "Previous PIO read/write transfer is still running\n");
+ return -EINVAL;
+ }
/* Program the control register */
reg = advk_readl(pcie, PIO_CTRL);
@@ -450,6 +463,7 @@ static int pcie_advk_write_config(struct udevice *bus, pci_dev_t bdf,
dev_dbg(pcie->dev, "\tPIO req. - strb = 0x%02x\n", reg);
/* Start the transfer */
+ advk_writel(pcie, 1, PIO_ISR);
advk_writel(pcie, 1, PIO_START);
if (!pcie_advk_wait_pio(pcie)) {
@@ -494,7 +508,7 @@ static int pcie_advk_wait_for_link(struct pcie_advk *pcie)
int retries;
/* check if the link is up or not */
- for (retries = 0; retries < MAX_RETRIES; retries++) {
+ for (retries = 0; retries < LINK_MAX_RETRIES; retries++) {
if (pcie_advk_link_up(pcie)) {
printf("PCIE-%d: Link up\n", pcie->first_busno);
return 0;
--
2.26.3
next reply other threads:[~2021-04-22 14:23 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-22 14:23 Pali Rohár [this message]
2021-04-26 9:22 ` [PATCH] arm: a37xx: pci: Fix processing PIO transfers Stefan Roese
2021-04-29 6:46 ` Stefan Roese
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=20210422142304.31065-1-pali@kernel.org \
--to=pali@kernel.org \
--cc=u-boot@lists.denx.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