From: Alan Cox <alan@linux.intel.com>
To: jeff@garzik.org.com, linux-kernel@vger.kernel.org,
linux-ide@vger.kernel.org
Subject: [RFC 4/4] libata: Make sil680 do its own exec_command posting
Date: Wed, 17 Feb 2010 13:15:00 +0000 [thread overview]
Message-ID: <20100217131321.16338.68370.stgit@localhost.localdomain> (raw)
In-Reply-To: <20100217130847.16338.55586.stgit@localhost.localdomain>
Use our own mmio area to avoid PCI posting. This avoids the rather slow
paranoid implementation in the default handler.
(Note - this is pure paranoia, brute force testing says the SIL680 is doing
the work itself somewhere as does the fact its worked without it for years
in the old IDE code)
We cannot make this logic generic as
- There are only a couple of potential SFF style users
- The other one can come up in PIO only modes with no bmdma
Signed-off-by: Alan Cox <alan@linux.intel.com>
---
drivers/ata/pata_sil680.c | 30 ++++++++++++++++++++++++++----
1 files changed, 26 insertions(+), 4 deletions(-)
diff --git a/drivers/ata/pata_sil680.c b/drivers/ata/pata_sil680.c
index a2ace48..065339d 100644
--- a/drivers/ata/pata_sil680.c
+++ b/drivers/ata/pata_sil680.c
@@ -190,15 +190,37 @@ static void sil680_set_dmamode(struct ata_port *ap, struct ata_device *adev)
pci_write_config_word(pdev, ua, ultra);
}
+/**
+ * sil680_sff_exec_command - issue ATA command to host controller
+ * @ap: port to which command is being issued
+ * @tf: ATA taskfile register set
+ *
+ * Issues ATA command, with proper synchronization with interrupt
+ * handler / other threads. Use our MMIO space for PCI posting to avoid
+ * a hideously slow cycle all the way to the device.
+ *
+ * LOCKING:
+ * spin_lock_irqsave(host lock)
+ */
+void sil680_sff_exec_command(struct ata_port *ap,
+ const struct ata_taskfile *tf)
+{
+ DPRINTK("ata%u: cmd 0x%X\n", ap->print_id, tf->command);
+ iowrite8(tf->command, ap->ioaddr.command_addr);
+ ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_CMD);
+}
+
static struct scsi_host_template sil680_sht = {
ATA_BMDMA_SHT(DRV_NAME),
};
+
static struct ata_port_operations sil680_port_ops = {
- .inherits = &ata_bmdma32_port_ops,
- .cable_detect = sil680_cable_detect,
- .set_piomode = sil680_set_piomode,
- .set_dmamode = sil680_set_dmamode,
+ .inherits = &ata_bmdma32_port_ops,
+ .sff_exec_command = sil680_sff_exec_command,
+ .cable_detect = sil680_cable_detect,
+ .set_piomode = sil680_set_piomode,
+ .set_dmamode = sil680_set_dmamode,
};
/**
next prev parent reply other threads:[~2010-02-17 13:38 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-02-17 13:10 [RFC 1/4] libata: cache device select Alan Cox
2010-02-17 13:11 ` [RFC 2/4] libata: Remove excess delay in the tf_load path Alan Cox
2010-02-17 13:13 ` [RFC 3/4] libata: Remove excess command issue delays Alan Cox
2010-02-17 14:10 ` Sergei Shtylyov
2010-02-17 15:34 ` Alan Cox
2010-02-17 13:15 ` Alan Cox [this message]
2010-02-18 5:13 ` [RFC 1/4] libata: cache device select Mark Lord
2010-02-18 10:16 ` Alan Cox
2010-03-01 20:15 ` Jeff Garzik
2010-03-02 17:28 ` Alan Cox
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=20100217131321.16338.68370.stgit@localhost.localdomain \
--to=alan@linux.intel.com \
--cc=jeff@garzik.org.com \
--cc=linux-ide@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
/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).