From: Marc Zyngier <marc.zyngier@arm.com>
To: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Cc: "linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>,
Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
Bjorn Helgaas <helgaas@google.com>,
Trent Piepho <tpiepho@impinj.com>,
Jingoo Han <jingoohan1@gmail.com>,
"faiz_abbas@ti.com" <faiz_abbas@ti.com>,
Joao Pinto <Joao.Pinto@synopsys.com>, Vignesh R <vigneshr@ti.com>
Subject: Re: [PATCH 0/3] PCI: designware: Fixing MSI handling flow
Date: Wed, 14 Nov 2018 09:54:39 +0000 [thread overview]
Message-ID: <86a7mcdlwg.wl-marc.zyngier@arm.com> (raw)
In-Reply-To: <e73891a5-d798-ee9c-607c-9648d07f6505@synopsys.com>
On Tue, 13 Nov 2018 23:16:33 +0000,
Gustavo Pimentel <gustavo.pimentel@synopsys.com> wrote:
>
> On 13/11/2018 22:57, Marc Zyngier wrote:
> > It recently came to light that the Designware PCIe driver is rather
> > broken in the way it handles MSI[1]:
> >
> > - It masks interrupt by disabling them, meaning that MSIs generated
> > during the masked window are simply lost. Oops.
> >
> > - Acking of the currently pending MSI is done outside of the interrupt
> > flow, getting moved around randomly and ultimately breaking the
> > driver. Not great.
> >
> > This series attempts to address this by switching to using the MASK
> > register for masking interrupts (!), and move the ack into the
> > appropriate callback, giving it a fixed place in the MSI handling
> > flow.
> >
> > Note that this is only compile-tested on my arm64 laptop, as I'm
> > travelling and do not have the required HW to test it anyway. I'd
> > welcome both review and testing by the interested parties (dwc
> > maintainer and users affected by existing bugs).
>
> As we spoke on the conference, as soon as I get back and I've the necessary
> conditions I'll test the discussed modifications on my HW.
I just realised (at 1am!) that the first patch is awfully buggy:
- ENABLE and MASK have opposite polarities
- There is nothing initialising the ENABLE and MASK registers
I've stashed the following fix-up on top of the existing series:
diff --git a/drivers/pci/controller/dwc/pcie-designware-host.c b/drivers/pci/controller/dwc/pcie-designware-host.c
index f06e67c60593..0fa9e8fdce66 100644
--- a/drivers/pci/controller/dwc/pcie-designware-host.c
+++ b/drivers/pci/controller/dwc/pcie-designware-host.c
@@ -166,7 +166,7 @@ static void dw_pci_bottom_mask(struct irq_data *data)
pp->irq_status[ctrl] &= ~(1 << bit);
dw_pcie_wr_own_conf(pp, PCIE_MSI_INTR0_MASK + res, 4,
- pp->irq_status[ctrl]);
+ ~pp->irq_status[ctrl]);
}
raw_spin_unlock_irqrestore(&pp->lock, flags);
@@ -189,7 +189,7 @@ static void dw_pci_bottom_unmask(struct irq_data *data)
pp->irq_status[ctrl] |= 1 << bit;
dw_pcie_wr_own_conf(pp, PCIE_MSI_INTR0_MASK + res, 4,
- pp->irq_status[ctrl]);
+ ~pp->irq_status[ctrl]);
}
raw_spin_unlock_irqrestore(&pp->lock, flags);
@@ -664,10 +664,15 @@ void dw_pcie_setup_rc(struct pcie_port *pp)
num_ctrls = pp->num_vectors / MAX_MSI_IRQS_PER_CTRL;
/* Initialize IRQ Status array */
- for (ctrl = 0; ctrl < num_ctrls; ctrl++)
- dw_pcie_rd_own_conf(pp, PCIE_MSI_INTR0_ENABLE +
+ for (ctrl = 0; ctrl < num_ctrls; ctrl++) {
+ dw_pcie_wr_own_conf(pp, PCIE_MSI_INTR0_MASK +
(ctrl * MSI_REG_CTRL_BLOCK_SIZE),
- 4, &pp->irq_status[ctrl]);
+ 4, ~0);
+ dw_pcie_wr_own_conf(pp, PCIE_MSI_INTR0_ENABLE +
+ (ctrl * MSI_REG_CTRL_BLOCK_SIZE),
+ 4, ~0);
+ pp->irq_status[ctrl] = 0;
+ }
/* Setup RC BARs */
dw_pcie_writel_dbi(pci, PCI_BASE_ADDRESS_0, 0x00000004);
Please let me know if this helps.
M.
--
Jazz is not dead, it just smell funny.
next prev parent reply other threads:[~2018-11-14 9:54 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-13 22:57 [PATCH 0/3] PCI: designware: Fixing MSI handling flow Marc Zyngier
2018-11-13 22:57 ` [PATCH 1/3] PCI: designware: Use interrupt masking instead of disabling Marc Zyngier
2018-12-03 18:02 ` [1/3] " Niklas Cassel
2018-12-04 9:41 ` [PATCH 1/3] " Gustavo Pimentel
2018-11-13 22:57 ` [PATCH 2/3] PCI: designware: Take lock when ACKing an interrupt Marc Zyngier
2018-11-14 19:08 ` Trent Piepho
2018-12-03 18:02 ` [2/3] " Niklas Cassel
2018-12-04 9:41 ` [PATCH 2/3] " Gustavo Pimentel
2018-11-13 22:57 ` [PATCH 3/3] PCI: designware: Move interrupt acking into the proper callback Marc Zyngier
2018-11-14 19:01 ` Trent Piepho
2018-12-03 18:02 ` [3/3] " Niklas Cassel
2018-12-04 9:41 ` [PATCH 3/3] " Gustavo Pimentel
2018-12-04 10:20 ` Kishon Vijay Abraham I
2018-12-04 13:45 ` Marc Zyngier
2018-12-07 8:12 ` Kishon Vijay Abraham I
2018-12-07 9:45 ` Marc Zyngier
2018-12-07 10:13 ` Kishon Vijay Abraham I
2018-12-11 12:35 ` Lorenzo Pieralisi
2018-12-12 5:54 ` Kishon Vijay Abraham I
2018-11-13 23:16 ` [PATCH 0/3] PCI: designware: Fixing MSI handling flow Gustavo Pimentel
2018-11-14 9:54 ` Marc Zyngier [this message]
2018-11-14 19:19 ` Trent Piepho
2018-11-14 22:01 ` Marc Zyngier
2018-11-14 22:25 ` Trent Piepho
2018-11-14 22:44 ` Marc Zyngier
2018-11-14 23:23 ` Trent Piepho
2018-11-19 20:37 ` Trent Piepho
2018-11-22 12:03 ` Gustavo Pimentel
2018-11-22 16:07 ` Gustavo Pimentel
2018-11-22 16:26 ` Lorenzo Pieralisi
2018-11-22 16:38 ` Marc Zyngier
2018-11-22 17:40 ` Gustavo Pimentel
2018-11-26 16:06 ` Trent Piepho
2018-11-27 7:51 ` Marc Zyngier
2018-11-27 17:23 ` Trent Piepho
2018-11-22 17:49 ` Gustavo Pimentel
2018-11-26 15:52 ` Trent Piepho
2018-11-27 7:50 ` Marc Zyngier
2018-11-27 18:12 ` Trent Piepho
2018-12-07 16:16 ` Gustavo Pimentel
2018-11-14 18:28 ` Trent Piepho
2018-11-14 22:07 ` Marc Zyngier
2018-11-14 22:50 ` Trent Piepho
2018-11-15 15:22 ` Gustavo Pimentel
2018-11-15 18:37 ` Trent Piepho
2018-11-15 19:29 ` Marc Zyngier
2018-11-19 20:14 ` Trent Piepho
2018-11-21 17:24 ` Stanimir Varbanov
2018-12-01 23:50 ` Niklas Cassel
2018-12-02 11:28 ` Stanimir Varbanov
2018-12-03 10:42 ` Lorenzo Pieralisi
2018-12-03 13:09 ` Niklas Cassel
2018-12-03 17:42 ` Lorenzo Pieralisi
2018-12-03 20:31 ` Trent Piepho
2018-12-10 16:17 ` Lorenzo Pieralisi
2018-12-10 16:30 ` Marc Zyngier
2018-12-10 18:15 ` Trent Piepho
2018-12-10 18:31 ` Marc Zyngier
2018-12-10 20:34 ` Trent Piepho
2018-12-12 9:10 ` Gustavo Pimentel
2018-12-12 8:55 ` Gustavo Pimentel
2018-12-11 11:43 ` Lorenzo Pieralisi
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=86a7mcdlwg.wl-marc.zyngier@arm.com \
--to=marc.zyngier@arm.com \
--cc=Joao.Pinto@synopsys.com \
--cc=faiz_abbas@ti.com \
--cc=gustavo.pimentel@synopsys.com \
--cc=helgaas@google.com \
--cc=jingoohan1@gmail.com \
--cc=linux-pci@vger.kernel.org \
--cc=lorenzo.pieralisi@arm.com \
--cc=tpiepho@impinj.com \
--cc=vigneshr@ti.com \
/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).