From: Peter Maydell <peter.maydell@linaro.org>
To: qemu-devel@nongnu.org
Subject: [Qemu-devel] [PULL 11/30] ARM i.MX timers: fix reset handling
Date: Mon, 27 Feb 2017 18:04:40 +0000 [thread overview]
Message-ID: <1488218699-31035-12-git-send-email-peter.maydell@linaro.org> (raw)
In-Reply-To: <1488218699-31035-1-git-send-email-peter.maydell@linaro.org>
From: Kurban Mallachiev <mallachiev@ispras.ru>
The i.MX timer device can be reset by writing to the SWR bit
of the CR register. This has to behave differently from hard
(power-on) reset because it does not reset all of the bits
in the CR register.
We were incorrectly implementing soft reset and hard reset
the same way, and in addition had a logic error which meant
that we were clearing the bits that soft-reset is supposed
to preserve and not touching the bits that soft-reset clears.
This was not correct behaviour for either kind of reset.
Separate out the soft reset and hard reset code paths, and
correct the handling of reset of the CR register so that it
is correct in both cases.
Signed-off-by: Kurban Mallachiev <mallachiev@ispras.ru>
[PMM: rephrased commit message, spacing on operators;
use bool rather than int for is_soft_reset]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
hw/timer/imx_gpt.c | 33 +++++++++++++++++++++++++--------
1 file changed, 25 insertions(+), 8 deletions(-)
diff --git a/hw/timer/imx_gpt.c b/hw/timer/imx_gpt.c
index 010ccbf..4b9b54b 100644
--- a/hw/timer/imx_gpt.c
+++ b/hw/timer/imx_gpt.c
@@ -296,18 +296,23 @@ static uint64_t imx_gpt_read(void *opaque, hwaddr offset, unsigned size)
return reg_value;
}
-static void imx_gpt_reset(DeviceState *dev)
-{
- IMXGPTState *s = IMX_GPT(dev);
+static void imx_gpt_reset_common(IMXGPTState *s, bool is_soft_reset)
+{
/* stop timer */
ptimer_stop(s->timer);
- /*
- * Soft reset doesn't touch some bits; hard reset clears them
+ /* Soft reset and hard reset differ only in their handling of the CR
+ * register -- soft reset preserves the values of some bits there.
*/
- s->cr &= ~(GPT_CR_EN|GPT_CR_ENMOD|GPT_CR_STOPEN|GPT_CR_DOZEN|
- GPT_CR_WAITEN|GPT_CR_DBGEN);
+ if (is_soft_reset) {
+ /* Clear all CR bits except those that are preserved by soft reset. */
+ s->cr &= GPT_CR_EN | GPT_CR_ENMOD | GPT_CR_STOPEN | GPT_CR_DOZEN |
+ GPT_CR_WAITEN | GPT_CR_DBGEN |
+ (GPT_CR_CLKSRC_MASK << GPT_CR_CLKSRC_SHIFT);
+ } else {
+ s->cr = 0;
+ }
s->sr = 0;
s->pr = 0;
s->ir = 0;
@@ -333,6 +338,18 @@ static void imx_gpt_reset(DeviceState *dev)
}
}
+static void imx_gpt_soft_reset(DeviceState *dev)
+{
+ IMXGPTState *s = IMX_GPT(dev);
+ imx_gpt_reset_common(s, true);
+}
+
+static void imx_gpt_reset(DeviceState *dev)
+{
+ IMXGPTState *s = IMX_GPT(dev);
+ imx_gpt_reset_common(s, false);
+}
+
static void imx_gpt_write(void *opaque, hwaddr offset, uint64_t value,
unsigned size)
{
@@ -348,7 +365,7 @@ static void imx_gpt_write(void *opaque, hwaddr offset, uint64_t value,
s->cr = value & ~0x7c14;
if (s->cr & GPT_CR_SWR) { /* force reset */
/* handle the reset */
- imx_gpt_reset(DEVICE(s));
+ imx_gpt_soft_reset(DEVICE(s));
} else {
/* set our freq, as the source might have changed */
imx_gpt_set_freq(s);
--
2.7.4
next prev parent reply other threads:[~2017-02-27 18:05 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-02-27 18:04 [Qemu-devel] [PULL 00/30] target-arm queue Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 01/30] target-arm: Implement BCM2835 hardware RNG Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 02/30] bcm2835_rng: Use qcrypto_random_bytes() rather than rand() Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 03/30] sd: sdhci: mask transfer mode register value Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 04/30] sd: sdhci: check transfer mode register in multi block transfer Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 05/30] sd: sdhci: conditionally invoke " Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 06/30] sd: sdhci: Remove block count enable check in single block transfers Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 07/30] hw/arm/virt: fix cpu object reference leak Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 08/30] Add missing fp_access_check() to aarch64 crypto instructions Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 09/30] cputlb: Don't assume do_unassigned_access() never returns Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 10/30] hw/arm/virt: Add a user option to disallow ITS instantiation Peter Maydell
2017-02-27 18:04 ` Peter Maydell [this message]
2017-02-27 18:04 ` [Qemu-devel] [PULL 12/30] armv7m: Rename nvic_state to NVICState Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 13/30] armv7m: Implement reading and writing of PRIGROUP Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 14/30] armv7m: Rewrite NVIC to not use any GIC code Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 15/30] armv7m: Fix condition check for taking exceptions Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 16/30] arm: gic: Remove references to NVIC Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 17/30] armv7m: Escalate exceptions to HardFault if necessary Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 18/30] armv7m: Remove unused armv7m_nvic_acknowledge_irq() return value Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 19/30] armv7m: Simpler and faster exception start Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 20/30] armv7m: VECTCLRACTIVE and VECTRESET are UNPREDICTABLE Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 21/30] armv7m: Extract "exception taken" code into functions Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 22/30] armv7m: Check exception return consistency Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 23/30] armv7m: Raise correct kind of UsageFault for attempts to execute ARM code Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 24/30] armv7m: Allow SHCSR writes to change pending and active bits Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 25/30] bcm2835_sdhost: add bcm2835 sdhost controller Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 26/30] hw/sd: add card-reparenting function Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 27/30] bcm2835_gpio: add bcm2835 gpio controller Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 28/30] bcm2835: add sdhost and gpio controllers Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 29/30] hw/arm/exynos: Fix Linux kernel division by zero for PLLs Peter Maydell
2017-02-27 18:04 ` [Qemu-devel] [PULL 30/30] hw/arm/exynos: Fix proper mapping of CPUs by providing real cluster ID Peter Maydell
2017-02-27 19:14 ` [Qemu-devel] [PULL 00/30] target-arm queue no-reply
2017-02-28 12:07 ` Peter Maydell
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=1488218699-31035-12-git-send-email-peter.maydell@linaro.org \
--to=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.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).